Robust for abnormal response from LLMs. (#4747)

### What problem does this PR solve?


### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -15,16 +15,16 @@ body:
     value: "Please provide the following information to help us understand the issue."
 - type: input
   attributes:
-    label: Branch name
-    description: Enter the name of the branch where you encountered the issue.
-    placeholder: e.g., main
+    label: RAGFlow workspace code commit ID
+    description: Enter the commit ID associated with the issue.
+    placeholder: e.g., 26d3480e
   validations:
     required: true
 - type: input
   attributes:
-    label: Commit ID
-    description: Enter the commit ID associated with the issue.
-    placeholder: e.g., c3b2a1
+    label: RAGFlow image version
+    description: Enter the image version(shown in RAGFlow UI, `System` page) associated with the issue.
+    placeholder: e.g., 26d3480e(v0.13.0~174)
   validations:
     required: true
 - type: textarea

.github/workflows/release.yml

@@ -0,0 +1,124 @@
+name: release
+
+on:
+  schedule:
+    - cron: '0 13 * * *'  # This schedule runs every 13:00:00Z(21:00:00+08:00)
+  # The "create tags" trigger is specifically focused on the creation of new tags, while the "push tags" trigger is activated when tags are pushed, including both new tag creations and updates to existing tags.
+  create:
+    tags:
+      - "v*.*.*"                  # normal release
+      - "nightly"                 # the only one mutable tag
+
+# https://docs.github.com/en/actions/using-jobs/using-concurrency
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  release:
+    runs-on: [ "self-hosted", "overseas" ]
+    steps:
+      - name: Ensure workspace ownership
+        run: echo "chown -R $USER $GITHUB_WORKSPACE" && sudo chown -R $USER $GITHUB_WORKSPACE
+
+      # https://github.com/actions/checkout/blob/v3/README.md
+      - name: Check out code
+        uses: actions/checkout@v4
+        with:
+          token: ${{ secrets.MY_GITHUB_TOKEN }}  # Use the secret as an environment variable
+          fetch-depth: 0
+          fetch-tags: true
+
+      - name: Prepare release body
+        run: |
+          if [[ $GITHUB_EVENT_NAME == 'create' ]]; then
+            RELEASE_TAG=${GITHUB_REF#refs/tags/}
+            if [[ $RELEASE_TAG == 'nightly' ]]; then
+              PRERELEASE=true
+            else
+              PRERELEASE=false
+            fi
+            echo "Workflow triggered by create tag: $RELEASE_TAG"
+          else
+            RELEASE_TAG=nightly
+            PRERELEASE=true
+            echo "Workflow triggered by schedule"
+          fi
+          echo "RELEASE_TAG=$RELEASE_TAG" >> $GITHUB_ENV
+          echo "PRERELEASE=$PRERELEASE" >> $GITHUB_ENV
+          RELEASE_DATETIME=$(date --rfc-3339=seconds)
+          echo Release $RELEASE_TAG created from $GITHUB_SHA at $RELEASE_DATETIME > release_body.md
+
+      - name: Move the existing mutable tag
+        # https://github.com/softprops/action-gh-release/issues/171
+        run: |
+          git fetch --tags
+          if [[ $GITHUB_EVENT_NAME == 'schedule' ]]; then
+            # Determine if a given tag exists and matches a specific Git commit.
+            # actions/checkout@v4 fetch-tags doesn't work when triggered by schedule
+            if [ "$(git rev-parse -q --verify "refs/tags/$RELEASE_TAG")" = "$GITHUB_SHA" ]; then
+              echo "mutable tag $RELEASE_TAG exists and matches $GITHUB_SHA"
+            else
+              git tag -f $RELEASE_TAG $GITHUB_SHA
+              git push -f origin $RELEASE_TAG:refs/tags/$RELEASE_TAG
+              echo "created/moved mutable tag $RELEASE_TAG to $GITHUB_SHA"
+            fi
+          fi
+
+      - name: Create or overwrite a release
+        # https://github.com/actions/upload-release-asset has been replaced by https://github.com/softprops/action-gh-release
+        uses: softprops/action-gh-release@v2
+        with:
+          token: ${{ secrets.MY_GITHUB_TOKEN }}  # Use the secret as an environment variable
+          prerelease: ${{ env.PRERELEASE }}
+          tag_name: ${{ env.RELEASE_TAG }}
+          # The body field does not support environment variable substitution directly.
+          body_path: release_body.md
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      # https://github.com/marketplace/actions/docker-login
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: infiniflow
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      # https://github.com/marketplace/actions/build-and-push-docker-images
+      - name: Build and push full image
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          push: true
+          tags: infiniflow/ragflow:${{ env.RELEASE_TAG }}
+          file: Dockerfile
+          platforms: linux/amd64
+
+      # https://github.com/marketplace/actions/build-and-push-docker-images
+      - name: Build and push slim image
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          push: true
+          tags: infiniflow/ragflow:${{ env.RELEASE_TAG }}-slim
+          file: Dockerfile
+          build-args: LIGHTEN=1
+          platforms: linux/amd64
+
+      - name: Build ragflow-sdk
+        if: startsWith(github.ref, 'refs/tags/v')
+        run: |
+          cd sdk/python && \
+          uv build
+
+      - name: Publish package distributions to PyPI
+        if: startsWith(github.ref, 'refs/tags/v')
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          packages-dir: sdk/python/dist/
+          password: ${{ secrets.PYPI_API_TOKEN }}
+          verbose: true

.github/workflows/tests.yml

@@ -42,35 +42,46 @@ jobs:
       - name: Ensure workspace ownership
         run: echo "chown -R $USER $GITHUB_WORKSPACE" && sudo chown -R $USER $GITHUB_WORKSPACE
 
+      # https://github.com/actions/checkout/issues/1781
       - name: Check out code
         uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          fetch-tags: true
 
-      - name: Build ragflow:dev-slim
+      # https://github.com/astral-sh/ruff-action
+      - name: Static check with Ruff
+        uses: astral-sh/ruff-action@v2
+        with:
+          version: ">=0.8.2"
+          args: "check --ignore E402"
+
+      - name: Build ragflow:nightly-slim
         run: |
           RUNNER_WORKSPACE_PREFIX=${RUNNER_WORKSPACE_PREFIX:-$HOME}
-          cp -r ${RUNNER_WORKSPACE_PREFIX}/huggingface.co ${RUNNER_WORKSPACE_PREFIX}/nltk_data ${RUNNER_WORKSPACE_PREFIX}/libssl*.deb .
-          sudo docker pull ubuntu:24.04
-          sudo docker build -f Dockerfile.slim -t infiniflow/ragflow:dev-slim .
+          sudo docker pull ubuntu:22.04
+          sudo docker build --progress=plain --build-arg LIGHTEN=1 --build-arg NEED_MIRROR=1 -f Dockerfile -t infiniflow/ragflow:nightly-slim .
 
-      - name: Build ragflow:dev
+      - name: Build ragflow:nightly
         run: |
-          sudo docker build -f Dockerfile -t infiniflow/ragflow:dev .
+          sudo docker build --progress=plain --build-arg NEED_MIRROR=1 -f Dockerfile -t infiniflow/ragflow:nightly .
 
-      - name: Start ragflow:dev-slim
+      - name: Start ragflow:nightly-slim
         run: |
+          echo "RAGFLOW_IMAGE=infiniflow/ragflow:nightly-slim" >> docker/.env
           sudo docker compose -f docker/docker-compose.yml up -d
 
-      - name: Stop ragflow:dev-slim
+      - name: Stop ragflow:nightly-slim
         if: always()  # always run this step even if previous steps failed
         run: |
           sudo docker compose -f docker/docker-compose.yml down -v
 
-      - name: Start ragflow:dev
+      - name: Start ragflow:nightly
         run: |
-          echo "RAGFLOW_IMAGE=infiniflow/ragflow:dev" >> docker/.env
+          echo "RAGFLOW_IMAGE=infiniflow/ragflow:nightly" >> docker/.env
           sudo docker compose -f docker/docker-compose.yml up -d
 
-      - name: Run tests
+      - name: Run sdk tests against Elasticsearch
         run: |
           export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
           export HOST_ADDRESS=http://host.docker.internal:9380
@@ -78,9 +89,49 @@ jobs:
             echo "Waiting for service to be available..."
             sleep 5
           done
-          cd sdk/python && poetry install && source .venv/bin/activate && cd test && pytest t_dataset.py t_chat.py t_session.py
+          cd sdk/python && uv sync --python 3.10 --frozen && uv pip install . && source .venv/bin/activate && cd test/test_sdk_api && pytest -s --tb=short get_email.py t_dataset.py t_chat.py t_session.py t_document.py t_chunk.py
 
-      - name: Stop ragflow:dev
+      - name: Run frontend api tests against Elasticsearch
+        run: |
+          export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
+          export HOST_ADDRESS=http://host.docker.internal:9380
+          until sudo docker exec ragflow-server curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
+            echo "Waiting for service to be available..."
+            sleep 5
+          done
+          cd sdk/python && uv sync --python 3.10 --frozen && uv pip install . && source .venv/bin/activate && cd test/test_frontend_api && pytest -s --tb=short get_email.py test_dataset.py 
+          
+
+      - name: Stop ragflow:nightly
         if: always()  # always run this step even if previous steps failed
         run: |
           sudo docker compose -f docker/docker-compose.yml down -v
+
+      - name: Start ragflow:nightly
+        run: |
+          sudo DOC_ENGINE=infinity docker compose -f docker/docker-compose.yml up -d
+
+      - name: Run sdk tests against Infinity
+        run: |
+          export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
+          export HOST_ADDRESS=http://host.docker.internal:9380
+          until sudo docker exec ragflow-server curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
+            echo "Waiting for service to be available..."
+            sleep 5
+          done
+          cd sdk/python && uv sync --python 3.10 --frozen && uv pip install . && source .venv/bin/activate && cd test/test_sdk_api && pytest -s --tb=short get_email.py t_dataset.py t_chat.py t_session.py t_document.py t_chunk.py
+
+      - name: Run frontend api tests against Infinity
+        run: |
+          export http_proxy=""; export https_proxy=""; export no_proxy=""; export HTTP_PROXY=""; export HTTPS_PROXY=""; export NO_PROXY=""
+          export HOST_ADDRESS=http://host.docker.internal:9380
+          until sudo docker exec ragflow-server curl -s --connect-timeout 5 ${HOST_ADDRESS} > /dev/null; do
+            echo "Waiting for service to be available..."
+            sleep 5
+          done
+          cd sdk/python && uv sync --python 3.10 --frozen && uv pip install . && source .venv/bin/activate && cd test/test_frontend_api && pytest -s --tb=short get_email.py test_dataset.py
+
+      - name: Stop ragflow:nightly
+        if: always()  # always run this step even if previous steps failed
+        run: |
+          sudo DOC_ENGINE=infinity docker compose -f docker/docker-compose.yml down -v

.gitignore

@@ -36,3 +36,5 @@ sdk/python/ragflow.egg-info/
 sdk/python/build/
 sdk/python/dist/
 sdk/python/ragflow_sdk.egg-info/
+huggingface.co/
+nltk_data/

CONTRIBUTING.md

@@ -1,6 +1,6 @@
 # Contribution guidelines
 
-This document offers guidlines and major considerations for submitting your contributions to RAGFlow.
+This document offers guidelines and major considerations for submitting your contributions to RAGFlow.
 
 - To report a bug, file a [GitHub issue](https://github.com/infiniflow/ragflow/issues/new/choose) with us.
 - For further questions, you can explore existing discussions or initiate a new one in [Discussions](https://github.com/orgs/infiniflow/discussions).

Dockerfile

@@ -1,41 +1,141 @@
 # base stage
-FROM ubuntu:24.04 AS base
+FROM ubuntu:22.04 AS base
 USER root
+SHELL ["/bin/bash", "-c"]
 
-ARG ARCH=amd64
-ENV LIGHTEN=0
+ARG NEED_MIRROR=0
+ARG LIGHTEN=0
+ENV LIGHTEN=${LIGHTEN}
 
 WORKDIR /ragflow
 
-RUN rm -f /etc/apt/apt.conf.d/docker-clean \
-    && echo 'Binary::apt::APT::Keep-Downloaded-Packages "true";' > /etc/apt/apt.conf.d/keep-cache
+# Copy models downloaded via download_deps.py
+RUN mkdir -p /ragflow/rag/res/deepdoc /root/.ragflow
+RUN --mount=type=bind,from=infiniflow/ragflow_deps:latest,source=/huggingface.co,target=/huggingface.co \
+    cp /huggingface.co/InfiniFlow/huqie/huqie.txt.trie /ragflow/rag/res/ && \
+    tar --exclude='.*' -cf - \
+        /huggingface.co/InfiniFlow/text_concat_xgb_v1.0 \
+        /huggingface.co/InfiniFlow/deepdoc \
+        | tar -xf - --strip-components=3 -C /ragflow/rag/res/deepdoc 
+RUN --mount=type=bind,from=infiniflow/ragflow_deps:latest,source=/huggingface.co,target=/huggingface.co \
+    if [ "$LIGHTEN" != "1" ]; then \
+        (tar -cf - \
+            /huggingface.co/BAAI/bge-large-zh-v1.5 \
+            /huggingface.co/BAAI/bge-reranker-v2-m3 \
+            /huggingface.co/maidalun1020/bce-embedding-base_v1 \
+            /huggingface.co/maidalun1020/bce-reranker-base_v1 \
+            | tar -xf - --strip-components=2 -C /root/.ragflow) \
+    fi
 
-RUN --mount=type=cache,id=ragflow_base_apt,target=/var/cache/apt,sharing=locked \
-    apt update && apt-get --no-install-recommends install -y ca-certificates
+# https://github.com/chrismattmann/tika-python
+# This is the only way to run python-tika without internet access. Without this set, the default is to check the tika version and pull latest every time from Apache.
+RUN --mount=type=bind,from=infiniflow/ragflow_deps:latest,source=/,target=/deps \
+    cp -r /deps/nltk_data /root/ && \
+    cp /deps/tika-server-standard-3.0.0.jar /deps/tika-server-standard-3.0.0.jar.md5 /ragflow/ && \
+    cp /deps/cl100k_base.tiktoken /ragflow/9b5ad71b2ce5302211f9c61530b329a4922fc6a4
 
-# If you download Python modules too slow, you can use a pip mirror site to speed up apt and poetry
-RUN sed -i 's|http://archive.ubuntu.com|https://mirrors.tuna.tsinghua.edu.cn|g' /etc/apt/sources.list.d/ubuntu.sources
-ENV POETRY_PYPI_MIRROR_URL=https://pypi.tuna.tsinghua.edu.cn/simple/
+ENV TIKA_SERVER_JAR="file:///ragflow/tika-server-standard-3.0.0.jar"
+ENV DEBIAN_FRONTEND=noninteractive
 
-RUN --mount=type=cache,id=ragflow_base_apt,target=/var/cache/apt,sharing=locked \
-    apt update && apt install -y curl libpython3-dev nginx libglib2.0-0 libglx-mesa0 pkg-config libicu-dev libgdiplus python3-pip python3-poetry \
-    && pip3 install --user --break-system-packages poetry-plugin-pypi-mirror --index-url https://pypi.tuna.tsinghua.edu.cn/simple/ \
-    && rm -rf /var/lib/apt/lists/*
+# Setup apt
+# Python package and implicit dependencies:
+# opencv-python: libglib2.0-0 libglx-mesa0 libgl1
+# aspose-slides: pkg-config libicu-dev libgdiplus         libssl1.1_1.1.1f-1ubuntu2_amd64.deb
+# python-pptx:   default-jdk                              tika-server-standard-3.0.0.jar
+# selenium:      libatk-bridge2.0-0                       chrome-linux64-121-0-6167-85
+# Building C extensions: libpython3-dev libgtk-4-1 libnss3 xdg-utils libgbm-dev
+RUN --mount=type=cache,id=ragflow_apt,target=/var/cache/apt,sharing=locked \
+    if [ "$NEED_MIRROR" == "1" ]; then \
+        sed -i 's|http://archive.ubuntu.com|https://mirrors.tuna.tsinghua.edu.cn|g' /etc/apt/sources.list; \
+    fi; \
+    rm -f /etc/apt/apt.conf.d/docker-clean && \
+    echo 'Binary::apt::APT::Keep-Downloaded-Packages "true";' > /etc/apt/apt.conf.d/keep-cache && \
+    chmod 1777 /tmp && \
+    apt update && \
+    apt --no-install-recommends install -y ca-certificates && \
+    apt update && \
+    apt install -y libglib2.0-0 libglx-mesa0 libgl1 && \
+    apt install -y pkg-config libicu-dev libgdiplus && \
+    apt install -y default-jdk && \
+    apt install -y libatk-bridge2.0-0 && \
+    apt install -y libpython3-dev libgtk-4-1 libnss3 xdg-utils libgbm-dev && \
+    apt install -y python3-pip pipx nginx unzip curl wget git vim less
+
+RUN if [ "$NEED_MIRROR" == "1" ]; then \
+        pip3 config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple && \
+        pip3 config set global.trusted-host pypi.tuna.tsinghua.edu.cn; \
+        mkdir -p /etc/uv && \
+        echo "[[index]]" > /etc/uv/uv.toml && \
+        echo 'url = "https://pypi.tuna.tsinghua.edu.cn/simple"' >> /etc/uv/uv.toml && \
+        echo "default = true" >> /etc/uv/uv.toml; \
+    fi; \
+    pipx install uv
+
+ENV PYTHONDONTWRITEBYTECODE=1 DOTNET_SYSTEM_GLOBALIZATION_INVARIANT=1
+ENV PATH=/root/.local/bin:$PATH
+
+# nodejs 12.22 on Ubuntu 22.04 is too old
+RUN --mount=type=cache,id=ragflow_apt,target=/var/cache/apt,sharing=locked \
+    curl -fsSL https://deb.nodesource.com/setup_20.x | bash - && \
+    apt purge -y nodejs npm cargo && \
+    apt autoremove -y && \
+    apt update && \
+    apt install -y nodejs
+
+# A modern version of cargo is needed for the latest version of the Rust compiler.
+RUN apt update && apt install -y curl build-essential \
+    && if [ "$NEED_MIRROR" == "1" ]; then \
+         # Use TUNA mirrors for rustup/rust dist files
+         export RUSTUP_DIST_SERVER="https://mirrors.tuna.tsinghua.edu.cn/rustup"; \
+         export RUSTUP_UPDATE_ROOT="https://mirrors.tuna.tsinghua.edu.cn/rustup/rustup"; \
+         echo "Using TUNA mirrors for Rustup."; \
+       fi; \
+    # Force curl to use HTTP/1.1
+    curl --proto '=https' --tlsv1.2 --http1.1 -sSf https://sh.rustup.rs | bash -s -- -y --profile minimal \
+    && echo 'export PATH="/root/.cargo/bin:${PATH}"' >> /root/.bashrc
+
+ENV PATH="/root/.cargo/bin:${PATH}"
+
+RUN cargo --version && rustc --version
+
+# Add msssql ODBC driver
+# macOS ARM64 environment, install msodbcsql18.
+# general x86_64 environment, install msodbcsql17.
+RUN --mount=type=cache,id=ragflow_apt,target=/var/cache/apt,sharing=locked \
+    curl https://packages.microsoft.com/keys/microsoft.asc | apt-key add - && \
+    curl https://packages.microsoft.com/config/ubuntu/22.04/prod.list > /etc/apt/sources.list.d/mssql-release.list && \
+    apt update && \
+    arch="$(uname -m)"; \
+    if [ "$arch" = "arm64" ] || [ "$arch" = "aarch64" ]; then \
+        # ARM64 (macOS/Apple Silicon or Linux aarch64)
+        ACCEPT_EULA=Y apt install -y unixodbc-dev msodbcsql18; \
+    else \
+        # x86_64 or others
+        ACCEPT_EULA=Y apt install -y unixodbc-dev msodbcsql17; \
+    fi || \
+    { echo "Failed to install ODBC driver"; exit 1; }
+
+
+
+# Add dependencies of selenium
+RUN --mount=type=bind,from=infiniflow/ragflow_deps:latest,source=/chrome-linux64-121-0-6167-85,target=/chrome-linux64.zip \
+    unzip /chrome-linux64.zip && \
+    mv chrome-linux64 /opt/chrome && \
+    ln -s /opt/chrome/chrome /usr/local/bin/
+RUN --mount=type=bind,from=infiniflow/ragflow_deps:latest,source=/chromedriver-linux64-121-0-6167-85,target=/chromedriver-linux64.zip \
+    unzip -j /chromedriver-linux64.zip chromedriver-linux64/chromedriver && \
+    mv chromedriver /usr/local/bin/ && \
+    rm -f /usr/bin/google-chrome
 
 # https://forum.aspose.com/t/aspose-slides-for-net-no-usable-version-of-libssl-found-with-linux-server/271344/13
 # aspose-slides on linux/arm64 is unavailable
-RUN --mount=type=bind,source=libssl1.1_1.1.1f-1ubuntu2_amd64.deb,target=/root/libssl1.1_1.1.1f-1ubuntu2_amd64.deb \
-    if [ "${ARCH}" = "amd64" ]; then \
-        dpkg -i /root/libssl1.1_1.1.1f-1ubuntu2_amd64.deb; \
+RUN --mount=type=bind,from=infiniflow/ragflow_deps:latest,source=/,target=/deps \
+    if [ "$(uname -m)" = "x86_64" ]; then \
+        dpkg -i /deps/libssl1.1_1.1.1f-1ubuntu2_amd64.deb; \
+    elif [ "$(uname -m)" = "aarch64" ]; then \
+        dpkg -i /deps/libssl1.1_1.1.1f-1ubuntu2_arm64.deb; \
     fi
 
-ENV PYTHONDONTWRITEBYTECODE=1 DOTNET_SYSTEM_GLOBALIZATION_INVARIANT=1
-
-# Configure Poetry
-ENV POETRY_NO_INTERACTION=1
-ENV POETRY_VIRTUALENVS_IN_PROJECT=true
-ENV POETRY_VIRTUALENVS_CREATE=true
-ENV POETRY_REQUESTS_TIMEOUT=15
 
 # builder stage
 FROM base AS builder
@@ -43,24 +143,38 @@ USER root
 
 WORKDIR /ragflow
 
-RUN --mount=type=cache,id=ragflow_builder_apt,target=/var/cache/apt,sharing=locked \
-    apt update && apt install -y nodejs npm cargo && \
-    rm -rf /var/lib/apt/lists/*
+# install dependencies from uv.lock file
+COPY pyproject.toml uv.lock ./
+
+# https://github.com/astral-sh/uv/issues/10462
+# uv records index url into uv.lock but doesn't failover among multiple indexes
+RUN --mount=type=cache,id=ragflow_uv,target=/root/.cache/uv,sharing=locked \
+    if [ "$NEED_MIRROR" == "1" ]; then \
+        sed -i 's|pypi.org|pypi.tuna.tsinghua.edu.cn|g' uv.lock; \
+    else \
+        sed -i 's|pypi.tuna.tsinghua.edu.cn|pypi.org|g' uv.lock; \
+    fi; \
+    if [ "$LIGHTEN" == "1" ]; then \
+        uv sync --python 3.10 --frozen; \
+    else \
+        uv sync --python 3.10 --frozen --all-extras; \
+    fi
 
 COPY web web
 COPY docs docs
-RUN --mount=type=cache,id=ragflow_builder_npm,target=/root/.npm,sharing=locked \
-    cd web && npm i --force && npm run build
+RUN --mount=type=cache,id=ragflow_npm,target=/root/.npm,sharing=locked \
+    cd web && npm install && npm run build
 
-# install dependencies from poetry.lock file
-COPY pyproject.toml poetry.toml poetry.lock ./
+COPY .git /ragflow/.git
 
-RUN --mount=type=cache,id=ragflow_builder_poetry,target=/root/.cache/pypoetry,sharing=locked \
-    if [ "$LIGHTEN" -eq 0 ]; then \
-        poetry install --sync --no-root --with=full; \
+RUN version_info=$(git describe --tags --match=v* --first-parent --always); \
+    if [ "$LIGHTEN" == "1" ]; then \
+        version_info="$version_info slim"; \
     else \
-        poetry install --sync --no-root; \
-    fi
+        version_info="$version_info full"; \
+    fi; \
+    echo "RAGFlow version: $version_info"; \
+    echo $version_info > /ragflow/VERSION
 
 # production stage
 FROM base AS production
@@ -68,42 +182,6 @@ USER root
 
 WORKDIR /ragflow
 
-# Install python packages' dependencies
-# cv2 requires libGL.so.1
-RUN --mount=type=cache,id=ragflow_production_apt,target=/var/cache/apt,sharing=locked \
-    apt update && apt install -y --no-install-recommends nginx libgl1 vim less && \
-    rm -rf /var/lib/apt/lists/*
-
-COPY web web
-COPY api api
-COPY conf conf
-COPY deepdoc deepdoc
-COPY rag rag
-COPY agent agent
-COPY graphrag graphrag
-COPY pyproject.toml poetry.toml poetry.lock ./
-
-# Copy models downloaded via download_deps.py
-RUN mkdir -p /ragflow/rag/res/deepdoc /root/.ragflow
-RUN --mount=type=bind,source=huggingface.co,target=/huggingface.co \
-    tar --exclude='.*' -cf - \
-        /huggingface.co/InfiniFlow/text_concat_xgb_v1.0 \
-        /huggingface.co/InfiniFlow/deepdoc \
-        | tar -xf - --strip-components=3 -C /ragflow/rag/res/deepdoc
-RUN --mount=type=bind,source=huggingface.co,target=/huggingface.co \
-    tar -cf - \
-        /huggingface.co/BAAI/bge-large-zh-v1.5 \
-        /huggingface.co/BAAI/bge-reranker-v2-m3 \
-        /huggingface.co/maidalun1020/bce-embedding-base_v1 \
-        /huggingface.co/maidalun1020/bce-reranker-base_v1 \
-        | tar -xf - --strip-components=2 -C /root/.ragflow
-
-# Copy nltk data downloaded via download_deps.py
-COPY nltk_data /root/nltk_data
-
-# Copy compiled web pages
-COPY --from=builder /ragflow/web/dist /ragflow/web/dist
-
 # Copy Python environment and packages
 ENV VIRTUAL_ENV=/ragflow/.venv
 COPY --from=builder ${VIRTUAL_ENV} ${VIRTUAL_ENV}
@@ -111,7 +189,21 @@ ENV PATH="${VIRTUAL_ENV}/bin:${PATH}"
 
 ENV PYTHONPATH=/ragflow/
 
-COPY docker/entrypoint.sh ./entrypoint.sh
-RUN chmod +x ./entrypoint.sh
+COPY web web
+COPY api api
+COPY conf conf
+COPY deepdoc deepdoc
+COPY rag rag
+COPY agent agent
+COPY graphrag graphrag
+COPY pyproject.toml uv.lock ./
 
+COPY docker/service_conf.yaml.template ./conf/service_conf.yaml.template
+COPY docker/entrypoint.sh docker/entrypoint-parser.sh ./
+RUN chmod +x ./entrypoint*.sh
+
+# Copy compiled web pages
+COPY --from=builder /ragflow/web/dist /ragflow/web/dist
+
+COPY --from=builder /ragflow/VERSION /ragflow/VERSION
 ENTRYPOINT ["./entrypoint.sh"]

Dockerfile.deps

@@ -0,0 +1,10 @@
+# This builds an image that contains the resources needed by Dockerfile
+#
+FROM scratch
+
+# Copy resources downloaded via download_deps.py
+COPY chromedriver-linux64-121-0-6167-85 chrome-linux64-121-0-6167-85 cl100k_base.tiktoken libssl1.1_1.1.1f-1ubuntu2_amd64.deb libssl1.1_1.1.1f-1ubuntu2_arm64.deb tika-server-standard-3.0.0.jar tika-server-standard-3.0.0.jar.md5 libssl*.deb /
+
+COPY nltk_data /nltk_data
+
+COPY huggingface.co /huggingface.co

Dockerfile.scratch.oc9

@@ -53,6 +53,7 @@ RUN conda run -n py11 python -m nltk.downloader wordnet
 ENV PYTHONPATH=/ragflow/
 ENV HF_ENDPOINT=https://hf-mirror.com
 
+COPY docker/service_conf.yaml.template ./conf/service_conf.yaml.template
 ADD docker/entrypoint.sh ./entrypoint.sh
 RUN chmod +x ./entrypoint.sh
 

Dockerfile.slim

@@ -1,109 +0,0 @@
-# base stage
-FROM ubuntu:24.04 AS base
-USER root
-
-ARG ARCH=amd64
-ENV LIGHTEN=1
-
-WORKDIR /ragflow
-
-RUN rm -f /etc/apt/apt.conf.d/docker-clean \
-    && echo 'Binary::apt::APT::Keep-Downloaded-Packages "true";' > /etc/apt/apt.conf.d/keep-cache
-
-RUN --mount=type=cache,id=ragflow_base_apt,target=/var/cache/apt,sharing=locked \
-    apt update && apt-get --no-install-recommends install -y ca-certificates
-
-# If you download Python modules too slow, you can use a pip mirror site to speed up apt and poetry
-RUN sed -i 's|http://archive.ubuntu.com|https://mirrors.tuna.tsinghua.edu.cn|g' /etc/apt/sources.list.d/ubuntu.sources
-ENV POETRY_PYPI_MIRROR_URL=https://pypi.tuna.tsinghua.edu.cn/simple/
-
-RUN --mount=type=cache,id=ragflow_base_apt,target=/var/cache/apt,sharing=locked \
-    apt update && apt install -y curl libpython3-dev nginx libglib2.0-0 libglx-mesa0 pkg-config libicu-dev libgdiplus python3-pip python3-poetry \
-    && pip3 install --user --break-system-packages poetry-plugin-pypi-mirror --index-url https://pypi.tuna.tsinghua.edu.cn/simple/ \
-    && rm -rf /var/lib/apt/lists/*
-
-# https://forum.aspose.com/t/aspose-slides-for-net-no-usable-version-of-libssl-found-with-linux-server/271344/13
-# aspose-slides on linux/arm64 is unavailable
-RUN if [ "${ARCH}" = "amd64" ]; then \
-        curl -o libssl1.deb http://archive.ubuntu.com/ubuntu/pool/main/o/openssl/libssl1.1_1.1.1f-1ubuntu2_amd64.deb && dpkg -i libssl1.deb && rm -f libssl1.deb; \
-    fi
-
-ENV PYTHONDONTWRITEBYTECODE=1 DOTNET_SYSTEM_GLOBALIZATION_INVARIANT=1
-
-# Configure Poetry
-ENV POETRY_NO_INTERACTION=1
-ENV POETRY_VIRTUALENVS_IN_PROJECT=true
-ENV POETRY_VIRTUALENVS_CREATE=true
-ENV POETRY_REQUESTS_TIMEOUT=15
-
-# builder stage
-FROM base AS builder
-USER root
-
-WORKDIR /ragflow
-
-RUN --mount=type=cache,id=ragflow_builder_apt,target=/var/cache/apt,sharing=locked \
-    apt update && apt install -y nodejs npm cargo && \
-    rm -rf /var/lib/apt/lists/*
-
-COPY web web
-COPY docs docs
-RUN --mount=type=cache,id=ragflow_builder_npm,target=/root/.npm,sharing=locked \
-    cd web && npm i && npm run build
-
-# install dependencies from poetry.lock file
-COPY pyproject.toml poetry.toml poetry.lock ./
-
-RUN --mount=type=cache,id=ragflow_builder_poetry,target=/root/.cache/pypoetry,sharing=locked \
-    if [ "$LIGHTEN" -eq 0 ]; then \
-        poetry install --sync --no-root --with=full; \
-    else \
-        poetry install --sync --no-root; \
-    fi
-
-# production stage
-FROM base AS production
-USER root
-
-WORKDIR /ragflow
-
-# Install python packages' dependencies
-# cv2 requires libGL.so.1
-RUN --mount=type=cache,id=ragflow_production_apt,target=/var/cache/apt,sharing=locked \
-    apt update && apt install -y --no-install-recommends nginx libgl1 vim less && \
-    rm -rf /var/lib/apt/lists/*
-
-COPY web web
-COPY api api
-COPY conf conf
-COPY deepdoc deepdoc
-COPY rag rag
-COPY agent agent
-COPY graphrag graphrag
-COPY pyproject.toml poetry.toml poetry.lock ./
-
-# Copy models downloaded via download_deps.py
-RUN mkdir -p /ragflow/rag/res/deepdoc /root/.ragflow
-RUN --mount=type=bind,source=huggingface.co,target=/huggingface.co \
-    tar --exclude='.*' -cf - \
-        /huggingface.co/InfiniFlow/text_concat_xgb_v1.0 \
-        /huggingface.co/InfiniFlow/deepdoc \
-        | tar -xf - --strip-components=3 -C /ragflow/rag/res/deepdoc
-
-# Copy nltk data downloaded via download_deps.py
-COPY nltk_data /root/nltk_data
-
-# Copy compiled web pages
-COPY --from=builder /ragflow/web/dist /ragflow/web/dist
-
-# Copy Python environment and packages
-ENV VIRTUAL_ENV=/ragflow/.venv
-COPY --from=builder ${VIRTUAL_ENV} ${VIRTUAL_ENV}
-ENV PATH="${VIRTUAL_ENV}/bin:${PATH}"
-
-ENV PYTHONPATH=/ragflow/
-
-COPY docker/entrypoint.sh ./entrypoint.sh
-RUN chmod +x ./entrypoint.sh
-
-ENTRYPOINT ["./entrypoint.sh"]

README.md

@@ -7,8 +7,11 @@
 <p align="center">
   <a href="./README.md">English</a> |
   <a href="./README_zh.md">简体中文</a> |
+  <a href="./README_tzh.md">繁体中文</a> |
   <a href="./README_ja.md">日本語</a> |
-  <a href="./README_ko.md">한국어</a>
+  <a href="./README_ko.md">한국어</a> |
+  <a href="./README_id.md">Bahasa Indonesia</a> |
+  <a href="/README_pt_br.md">Português (Brasil)</a>
 </p>
 
 <p align="center">
@@ -19,7 +22,7 @@
         <img alt="Static Badge" src="https://img.shields.io/badge/Online-Demo-4e6b99">
     </a>
     <a href="https://hub.docker.com/r/infiniflow/ragflow" target="_blank">
-        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.13.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.13.0">
+        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.16.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.16.0">
     </a>
     <a href="https://github.com/infiniflow/ragflow/releases/latest">
         <img src="https://img.shields.io/github/v/release/infiniflow/ragflow?color=blue&label=Latest%20Release" alt="Latest Release">
@@ -31,14 +34,14 @@
 
 <h4 align="center">
   <a href="https://ragflow.io/docs/dev/">Document</a> |
-  <a href="https://github.com/infiniflow/ragflow/issues/162">Roadmap</a> |
+  <a href="https://github.com/infiniflow/ragflow/issues/4214">Roadmap</a> |
   <a href="https://twitter.com/infiniflowai">Twitter</a> |
   <a href="https://discord.gg/4XxujFgUN7">Discord</a> |
   <a href="https://demo.ragflow.io">Demo</a>
 </h4>
 
 <details open>
-<summary></b>📕 Table of Contents</b></summary>
+<summary><b>📕 Table of Contents</b></summary>
 
 - 💡 [What is RAGFlow?](#-what-is-ragflow)
 - 🎮 [Demo](#-demo)
@@ -67,23 +70,27 @@ data.
 ## 🎮 Demo
 
 Try our demo at [https://demo.ragflow.io](https://demo.ragflow.io).
+
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
 <img src="https://github.com/infiniflow/ragflow/assets/7248/2f6baa3e-1092-4f11-866d-36f6a9d075e5" width="1200"/>
-<img src="https://github.com/infiniflow/ragflow/assets/12318111/b083d173-dadc-4ea9-bdeb-180d7df514eb" width="1200"/>
+<img src="https://github.com/user-attachments/assets/504bbbf1-c9f7-4d83-8cc5-e9cb63c26db6" width="1200"/>
 </div>
 
 ## 🔥 Latest Updates
 
-- 2024-09-29 Optimizes multi-round conversations.
-- 2024-09-13 Adds search mode for knowledge base Q&A.
-- 2024-09-09 Adds a medical consultant agent template.
+- 2025-02-05 Updates the model list of 'SILICONFLOW' and adds support for Deepseek-R1/DeepSeek-V3.
+- 2025-01-26 Optimizes knowledge graph extraction and application, offering various configuration options.
+- 2024-12-18 Upgrades Document Layout Analysis model in Deepdoc.
+- 2024-12-04 Adds support for pagerank score in knowledge base.
+- 2024-11-22 Adds more variables to Agent.
+- 2024-11-01 Adds keyword extraction and related question generation to the parsed chunks to improve the accuracy of retrieval.
 - 2024-08-22 Support text to SQL statements through RAG.
-- 2024-08-02 Supports GraphRAG inspired by [graphrag](https://github.com/microsoft/graphrag) and mind map.
 
 ## 🎉 Stay Tuned
 
 ⭐️ Star our repository to stay up-to-date with exciting new features and improvements! Get instant notifications for new
 releases! 🌟
+
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
 <img src="https://github.com/user-attachments/assets/18c9707e-b8aa-4caf-a154-037089c105ba" width="1200"/>
 </div>
@@ -132,7 +139,7 @@ releases! 🌟
 - Disk >= 50 GB
 - Docker >= 24.0.0 & Docker Compose >= v2.26.1
   > If you have not installed Docker on your local machine (Windows, Mac, or Linux),
-  see [Install Docker Engine](https://docs.docker.com/engine/install/).
+  > see [Install Docker Engine](https://docs.docker.com/engine/install/).
 
 ### 🚀 Start up the server
 
@@ -152,7 +159,7 @@ releases! 🌟
    > ```
    >
    > This change will be reset after a system reboot. To ensure your change remains permanent, add or update the
-   `vm.max_map_count` value in **/etc/sysctl.conf** accordingly:
+   > `vm.max_map_count` value in **/etc/sysctl.conf** accordingly:
    >
    > ```bash
    > vm.max_map_count=262144
@@ -164,29 +171,21 @@ releases! 🌟
    $ git clone https://github.com/infiniflow/ragflow.git
    ```
 
-3. Build the pre-built Docker images and start up the server:
+3. Start up the server using the pre-built Docker images:
 
-   > The command below downloads the dev version Docker image for RAGFlow slim (`dev-slim`). Note that RAGFlow slim
-   Docker images do not include embedding models or Python libraries and hence are approximately 1GB in size.
+   > The command below downloads the `v0.16.0-slim` edition of the RAGFlow Docker image. Refer to the following table for descriptions of different RAGFlow editions. To download an RAGFlow edition different from `v0.16.0-slim`, update the `RAGFLOW_IMAGE` variable accordingly in **docker/.env** before using `docker compose` to start the server. For example: set `RAGFLOW_IMAGE=infiniflow/ragflow:v0.16.0` for the full edition `v0.16.0`.
 
    ```bash
-   $ cd ragflow/docker
-   $ docker compose -f docker-compose.yml up -d
+   $ cd ragflow
+   $ docker compose -f docker/docker-compose.yml up -d
    ```
 
-   > - To download a RAGFlow slim Docker image of a specific version, update the `RAGFlow_IMAGE` variable in *
-       *docker/.env** to your desired version. For example, `RAGFLOW_IMAGE=infiniflow/ragflow:v0.13.0-slim`. After
-       making this change, rerun the command above to initiate the download.
-   > - To download the dev version of RAGFlow Docker image *including* embedding models and Python libraries, update the
-       `RAGFlow_IMAGE` variable in **docker/.env** to `RAGFLOW_IMAGE=infiniflow/ragflow:dev`. After making this change,
-       rerun the command above to initiate the download.
-   > - To download a specific version of RAGFlow Docker image *including* embedding models and Python libraries, update
-       the `RAGFlow_IMAGE` variable in **docker/.env** to your desired version. For example,
-       `RAGFLOW_IMAGE=infiniflow/ragflow:v0.13.0`. After making this change, rerun the command above to initiate the
-       download.
-
-   > **NOTE:** A RAGFlow Docker image that includes embedding models and Python libraries is approximately 9GB in size
-   and may take significantly longer time to load.
+   | RAGFlow image tag | Image size (GB) | Has embedding models? | Stable?                  |
+   |-------------------|-----------------|-----------------------|--------------------------|
+   | v0.16.0           | &approx;9       | :heavy_check_mark:    | Stable release           |
+   | v0.16.0-slim      | &approx;2       | ❌                     | Stable release           |
+   | nightly           | &approx;9       | :heavy_check_mark:    | _Unstable_ nightly build |
+   | nightly-slim      | &approx;2       | ❌                     | _Unstable_ nightly build |
 
 4. Check the server status after having the server up and running:
 
@@ -209,13 +208,14 @@ releases! 🌟
     * Running on http://x.x.x.x:9380
     INFO:werkzeug:Press CTRL+C to quit
    ```
-   > If you skip this confirmation step and directly log in to RAGFlow, your browser may prompt a `network abnormal`
-   error because, at that moment, your RAGFlow may not be fully initialized.
+
+   > If you skip this confirmation step and directly log in to RAGFlow, your browser may prompt a `network anormal`
+   > error because, at that moment, your RAGFlow may not be fully initialized.
 
 5. In your web browser, enter the IP address of your server and log in to RAGFlow.
    > With the default settings, you only need to enter `http://IP_OF_YOUR_MACHINE` (**sans** port number) as the default
-   HTTP serving port `80` can be omitted when using the default configurations.
-6. In [service_conf.yaml](./docker/service_conf.yaml), select the desired LLM factory in `user_default_llm` and update
+   > HTTP serving port `80` can be omitted when using the default configurations.
+6. In [service_conf.yaml.template](./docker/service_conf.yaml.template), select the desired LLM factory in `user_default_llm` and update
    the `API_KEY` field with the corresponding API key.
 
    > See [llm_api_key_setup](https://ragflow.io/docs/dev/llm_api_key_setup) for more information.
@@ -228,17 +228,11 @@ When it comes to system configurations, you will need to manage the following fi
 
 - [.env](./docker/.env): Keeps the fundamental setups for the system, such as `SVR_HTTP_PORT`, `MYSQL_PASSWORD`, and
   `MINIO_PASSWORD`.
-- [service_conf.yaml](./docker/service_conf.yaml): Configures the back-end services.
-- [docker-compose.yml](./docker/docker-compose.yml): The system relies
-  on [docker-compose.yml](./docker/docker-compose.yml) to start up.
-
-You must ensure that changes to the [.env](./docker/.env) file are in line with what are in
-the [service_conf.yaml](./docker/service_conf.yaml) file.
+- [service_conf.yaml.template](./docker/service_conf.yaml.template): Configures the back-end services. The environment variables in this file will be automatically populated when the Docker container starts. Any environment variables set within the Docker container will be available for use, allowing you to customize service behavior based on the deployment environment.
+- [docker-compose.yml](./docker/docker-compose.yml): The system relies on [docker-compose.yml](./docker/docker-compose.yml) to start up.
 
 > The [./docker/README](./docker/README.md) file provides a detailed description of the environment settings and service
-> configurations, and you are REQUIRED to ensure that all environment settings listed in
-> the [./docker/README](./docker/README.md) file are aligned with the corresponding configurations in
-> the [service_conf.yaml](./docker/service_conf.yaml) file.
+> configurations which can be used as `${ENV_VARS}` in the [service_conf.yaml.template](./docker/service_conf.yaml.template) file.
 
 To update the default HTTP serving port (80), go to [docker-compose.yml](./docker/docker-compose.yml) and change `80:80`
 to `<YOUR_SERVING_PORT>:80`.
@@ -249,16 +243,35 @@ Updates to the above configurations require a reboot of all containers to take e
 > $ docker compose -f docker/docker-compose.yml up -d
 > ```
 
+### Switch doc engine from Elasticsearch to Infinity
+
+RAGFlow uses Elasticsearch by default for storing full text and vectors. To switch to [Infinity](https://github.com/infiniflow/infinity/), follow these steps:
+
+1. Stop all running containers:
+
+   ```bash
+   $ docker compose -f docker/docker-compose.yml down -v
+   ```
+
+2. Set `DOC_ENGINE` in **docker/.env** to `infinity`.
+
+3. Start the containers:
+
+   ```bash
+   $ docker compose -f docker/docker-compose.yml up -d
+   ```
+
+> [!WARNING]
+> Switching to Infinity on a Linux/arm64 machine is not yet officially supported.
+
 ## 🔧 Build a Docker image without embedding models
 
-This image is approximately 1 GB in size and relies on external LLM and embedding services.
+This image is approximately 2 GB in size and relies on external LLM and embedding services.
 
 ```bash
 git clone https://github.com/infiniflow/ragflow.git
 cd ragflow/
-pip3 install huggingface-hub nltk
-python3 download_deps.py
-docker build -f Dockerfile.slim -t infiniflow/ragflow:dev-slim .
+docker build --build-arg LIGHTEN=1 -f Dockerfile -t infiniflow/ragflow:nightly-slim .
 ```
 
 ## 🔧 Build a Docker image including embedding models
@@ -268,36 +281,36 @@ This image is approximately 9 GB in size. As it includes embedding models, it re
 ```bash
 git clone https://github.com/infiniflow/ragflow.git
 cd ragflow/
-pip3 install huggingface-hub nltk
-python3 download_deps.py
-docker build -f Dockerfile -t infiniflow/ragflow:dev .
+docker build -f Dockerfile -t infiniflow/ragflow:nightly .
 ```
 
 ## 🔨 Launch service from source for development
 
-1. Install Poetry, or skip this step if it is already installed:
+1. Install uv, or skip this step if it is already installed:
+
    ```bash
-   curl -sSL https://install.python-poetry.org | python3 -
+   pipx install uv
    ```
 
 2. Clone the source code and install Python dependencies:
+
    ```bash
    git clone https://github.com/infiniflow/ragflow.git
    cd ragflow/
-   export POETRY_VIRTUALENVS_CREATE=true POETRY_VIRTUALENVS_IN_PROJECT=true
-   ~/.local/bin/poetry install --sync --no-root # install RAGFlow dependent python modules
+   uv sync --python 3.10 --all-extras # install RAGFlow dependent python modules
    ```
 
 3. Launch the dependent services (MinIO, Elasticsearch, Redis, and MySQL) using Docker Compose:
+
    ```bash
    docker compose -f docker/docker-compose-base.yml up -d
    ```
 
-   Add the following line to `/etc/hosts` to resolve all hosts specified in **docker/service_conf.yaml** to `127.0.0.1`:
+   Add the following line to `/etc/hosts` to resolve all hosts specified in **docker/.env** to `127.0.0.1`:
+
    ```
-   127.0.0.1       es01 mysql minio redis
+   127.0.0.1       es01 infinity mysql minio redis
    ```
-   In **docker/service_conf.yaml**, update mysql port to `5455` and es port to `1200`, as specified in **docker/.env**.
 
 4. If you cannot access HuggingFace, set the `HF_ENDPOINT` environment variable to use a mirror site:
 
@@ -306,6 +319,7 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
    ```
 
 5. Launch backend service:
+
    ```bash
    source .venv/bin/activate
    export PYTHONPATH=$(pwd)
@@ -315,10 +329,10 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
 6. Install frontend dependencies:
    ```bash
    cd web
-   npm install --force
+   npm install
    ```
-7. Configure frontend to update `proxy.target` in **.umirc.ts** to `http://127.0.0.1:9380`:
-8. Launch frontend service:
+7. Launch frontend service:
+
    ```bash
    npm run dev
    ```
@@ -336,7 +350,7 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
 
 ## 📜 Roadmap
 
-See the [RAGFlow Roadmap 2024](https://github.com/infiniflow/ragflow/issues/162)
+See the [RAGFlow Roadmap 2025](https://github.com/infiniflow/ragflow/issues/4214)
 
 ## 🏄 Community
 

README_id.md

@@ -0,0 +1,333 @@
+<div align="center">
+<a href="https://demo.ragflow.io/">
+<img src="web/src/assets/logo-with-text.png" width="520" alt="Logo ragflow">
+</a>
+</div>
+
+<p align="center">
+  <a href="./README.md">English</a> |
+  <a href="./README_zh.md">简体中文</a> |
+  <a href="./README_tzh.md">繁体中文</a> |
+  <a href="./README_ja.md">日本語</a> |
+  <a href="./README_ko.md">한국어</a> |
+  <a href="./README_id.md">Bahasa Indonesia</a> |
+  <a href="/README_pt_br.md">Português (Brasil)</a>
+</p>
+
+<p align="center">
+    <a href="https://x.com/intent/follow?screen_name=infiniflowai" target="_blank">
+        <img src="https://img.shields.io/twitter/follow/infiniflow?logo=X&color=%20%23f5f5f5" alt="Ikuti di X (Twitter)">
+    </a>
+    <a href="https://demo.ragflow.io" target="_blank">
+        <img alt="Lencana Daring" src="https://img.shields.io/badge/Online-Demo-4e6b99">
+    </a>
+    <a href="https://hub.docker.com/r/infiniflow/ragflow" target="_blank">
+        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.16.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.16.0">
+    </a>
+    <a href="https://github.com/infiniflow/ragflow/releases/latest">
+        <img src="https://img.shields.io/github/v/release/infiniflow/ragflow?color=blue&label=Rilis%20Terbaru" alt="Rilis Terbaru">
+    </a>
+    <a href="https://github.com/infiniflow/ragflow/blob/main/LICENSE">
+        <img height="21" src="https://img.shields.io/badge/Lisensi-Apache--2.0-ffffff?labelColor=d4eaf7&color=2e6cc4" alt="Lisensi">
+    </a>
+</p>
+
+<h4 align="center">
+  <a href="https://ragflow.io/docs/dev/">Dokumentasi</a> |
+  <a href="https://github.com/infiniflow/ragflow/issues/4214">Peta Jalan</a> |
+  <a href="https://twitter.com/infiniflowai">Twitter</a> |
+  <a href="https://discord.gg/4XxujFgUN7">Discord</a> |
+  <a href="https://demo.ragflow.io">Demo</a>
+</h4>
+
+<details open>
+<summary></b>📕 Daftar Isi</b></summary>
+
+- 💡 [Apa Itu RAGFlow?](#-apa-itu-ragflow)
+- 🎮 [Demo](#-demo)
+- 📌 [Pembaruan Terbaru](#-pembaruan-terbaru)
+- 🌟 [Fitur Utama](#-fitur-utama)
+- 🔎 [Arsitektur Sistem](#-arsitektur-sistem)
+- 🎬 [Mulai](#-mulai)
+- 🔧 [Konfigurasi](#-konfigurasi)
+- 🔧 [Membangun Image Docker tanpa Model Embedding](#-membangun-image-docker-tanpa-model-embedding)
+- 🔧 [Membangun Image Docker dengan Model Embedding](#-membangun-image-docker-dengan-model-embedding)
+- 🔨 [Meluncurkan aplikasi dari Sumber untuk Pengembangan](#-meluncurkan-aplikasi-dari-sumber-untuk-pengembangan)
+- 📚 [Dokumentasi](#-dokumentasi)
+- 📜 [Peta Jalan](#-peta-jalan)
+- 🏄 [Komunitas](#-komunitas)
+- 🙌 [Kontribusi](#-kontribusi)
+
+</details>
+
+## 💡 Apa Itu RAGFlow?
+
+[RAGFlow](https://ragflow.io/) adalah mesin RAG (Retrieval-Augmented Generation) open-source berbasis pemahaman dokumen yang mendalam. Platform ini menyediakan alur kerja RAG yang efisien untuk bisnis dengan berbagai skala, menggabungkan LLM (Large Language Models) untuk menyediakan kemampuan tanya-jawab yang benar dan didukung oleh referensi dari data terstruktur kompleks.
+
+## 🎮 Demo
+
+Coba demo kami di [https://demo.ragflow.io](https://demo.ragflow.io).
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/7248/2f6baa3e-1092-4f11-866d-36f6a9d075e5" width="1200"/>
+<img src="https://github.com/user-attachments/assets/504bbbf1-c9f7-4d83-8cc5-e9cb63c26db6" width="1200"/>
+</div>
+
+## 🔥 Pembaruan Terbaru
+
+- 2025-02-05 Memperbarui daftar model 'SILICONFLOW' dan menambahkan dukungan untuk Deepseek-R1/DeepSeek-V3.
+- 2025-01-26 Optimalkan ekstraksi dan penerapan grafik pengetahuan dan sediakan berbagai opsi konfigurasi.
+- 2024-12-18 Meningkatkan model Analisis Tata Letak Dokumen di Deepdoc.
+- 2024-12-04 Mendukung skor pagerank ke basis pengetahuan.
+- 2024-11-22 Peningkatan definisi dan penggunaan variabel di Agen.
+- 2024-11-01 Penambahan ekstraksi kata kunci dan pembuatan pertanyaan terkait untuk meningkatkan akurasi pengambilan.
+- 2024-08-22 Dukungan untuk teks ke pernyataan SQL melalui RAG.
+
+## 🎉 Tetap Terkini
+
+⭐️ Star repositori kami untuk tetap mendapat informasi tentang fitur baru dan peningkatan menarik! 🌟
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/user-attachments/assets/18c9707e-b8aa-4caf-a154-037089c105ba" width="1200"/>
+</div>
+
+## 🌟 Fitur Utama
+
+### 🍭 **"Kualitas Masuk, Kualitas Keluar"**
+
+- Ekstraksi pengetahuan berbasis pemahaman dokumen mendalam dari data tidak terstruktur dengan format yang rumit.
+- Menemukan "jarum di tumpukan data" dengan token yang hampir tidak terbatas.
+
+### 🍱 **Pemotongan Berbasis Template**
+
+- Cerdas dan dapat dijelaskan.
+- Banyak pilihan template yang tersedia.
+
+### 🌱 **Referensi yang Didasarkan pada Data untuk Mengurangi Hallusinasi**
+
+- Visualisasi pemotongan teks memungkinkan intervensi manusia.
+- Tampilan cepat referensi kunci dan referensi yang dapat dilacak untuk mendukung jawaban yang didasarkan pada fakta.
+
+### 🍔 **Kompatibilitas dengan Sumber Data Heterogen**
+
+- Mendukung Word, slide, excel, txt, gambar, salinan hasil scan, data terstruktur, halaman web, dan banyak lagi.
+
+### 🛀 **Alur Kerja RAG yang Otomatis dan Mudah**
+
+- Orkestrasi RAG yang ramping untuk bisnis kecil dan besar.
+- LLM yang dapat dikonfigurasi serta model embedding.
+- Peringkat ulang berpasangan dengan beberapa pengambilan ulang.
+- API intuitif untuk integrasi yang mudah dengan bisnis.
+
+## 🔎 Arsitektur Sistem
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/12318111/d6ac5664-c237-4200-a7c2-a4a00691b485" width="1000"/>
+</div>
+
+## 🎬 Mulai
+
+### 📝 Prasyarat
+
+- CPU >= 4 inti
+- RAM >= 16 GB
+- Disk >= 50 GB
+- Docker >= 24.0.0 & Docker Compose >= v2.26.1
+
+### 🚀 Menjalankan Server
+
+1. Pastikan `vm.max_map_count` >= 262144:
+
+   > Untuk memeriksa nilai `vm.max_map_count`:
+   >
+   > ```bash
+   > $ sysctl vm.max_map_count
+   > ```
+   >
+   > Jika nilainya kurang dari 262144, setel ulang `vm.max_map_count` ke setidaknya 262144:
+   >
+   > ```bash
+   > # Dalam contoh ini, kita atur menjadi 262144:
+   > $ sudo sysctl -w vm.max_map_count=262144
+   > ```
+   >
+   > Perubahan ini akan hilang setelah sistem direboot. Untuk membuat perubahan ini permanen, tambahkan atau perbarui nilai
+   > `vm.max_map_count` di **/etc/sysctl.conf**:
+   >
+   > ```bash
+   > vm.max_map_count=262144
+   > ```
+
+2. Clone repositori:
+
+   ```bash
+   $ git clone https://github.com/infiniflow/ragflow.git
+   ```
+
+3. Bangun image Docker pre-built dan jalankan server:
+
+   > Perintah di bawah ini mengunduh edisi v0.16.0-slim dari gambar Docker RAGFlow. Silakan merujuk ke tabel berikut untuk deskripsi berbagai edisi RAGFlow. Untuk mengunduh edisi RAGFlow yang berbeda dari v0.16.0-slim, perbarui variabel RAGFLOW_IMAGE di docker/.env sebelum menggunakan docker compose untuk memulai server. Misalnya, atur RAGFLOW_IMAGE=infiniflow/ragflow:v0.16.0 untuk edisi lengkap v0.16.0.
+
+   ```bash
+   $ cd ragflow
+   $ docker compose -f docker/docker-compose.yml up -d
+   ```
+
+   | RAGFlow image tag | Image size (GB) | Has embedding models? | Stable?                  |
+   | ----------------- | --------------- | --------------------- | ------------------------ |
+   | v0.16.0           | &approx;9       | :heavy_check_mark:    | Stable release           |
+   | v0.16.0-slim      | &approx;2       | ❌                    | Stable release           |
+   | nightly           | &approx;9       | :heavy_check_mark:    | _Unstable_ nightly build |
+   | nightly-slim      | &approx;2       | ❌                    | _Unstable_ nightly build |
+
+4. Periksa status server setelah server aktif dan berjalan:
+
+   ```bash
+   $ docker logs -f ragflow-server
+   ```
+
+   _Output berikut menandakan bahwa sistem berhasil diluncurkan:_
+
+   ```bash
+
+         ____   ___    ______ ______ __
+        / __ \ /   |  / ____// ____// /____  _      __
+       / /_/ // /| | / / __ / /_   / // __ \| | /| / /
+      / _, _// ___ |/ /_/ // __/  / // /_/ /| |/ |/ /
+     /_/ |_|/_/  |_|\____//_/    /_/ \____/ |__/|__/
+
+    * Running on all addresses (0.0.0.0)
+    * Running on http://127.0.0.1:9380
+    * Running on http://x.x.x.x:9380
+    INFO:werkzeug:Press CTRL+C to quit
+   ```
+
+   > Jika Anda melewatkan langkah ini dan langsung login ke RAGFlow, browser Anda mungkin menampilkan error `network anormal`
+   > karena RAGFlow mungkin belum sepenuhnya siap.
+
+5. Buka browser web Anda, masukkan alamat IP server Anda, dan login ke RAGFlow.
+   > Dengan pengaturan default, Anda hanya perlu memasukkan `http://IP_DEVICE_ANDA` (**tanpa** nomor port) karena
+   > port HTTP default `80` bisa dihilangkan saat menggunakan konfigurasi default.
+6. Dalam [service_conf.yaml.template](./docker/service_conf.yaml.template), pilih LLM factory yang diinginkan di `user_default_llm` dan perbarui
+   bidang `API_KEY` dengan kunci API yang sesuai.
+
+   > Lihat [llm_api_key_setup](https://ragflow.io/docs/dev/llm_api_key_setup) untuk informasi lebih lanjut.
+
+   _Sistem telah siap digunakan!_
+
+## 🔧 Konfigurasi
+
+Untuk konfigurasi sistem, Anda perlu mengelola file-file berikut:
+
+- [.env](./docker/.env): Menyimpan pengaturan dasar sistem, seperti `SVR_HTTP_PORT`, `MYSQL_PASSWORD`, dan
+  `MINIO_PASSWORD`.
+- [service_conf.yaml.template](./docker/service_conf.yaml.template): Mengonfigurasi aplikasi backend.
+- [docker-compose.yml](./docker/docker-compose.yml): Sistem ini bergantung pada [docker-compose.yml](./docker/docker-compose.yml) untuk memulai.
+
+Untuk memperbarui port HTTP default (80), buka [docker-compose.yml](./docker/docker-compose.yml) dan ubah `80:80`
+menjadi `<YOUR_SERVING_PORT>:80`.
+
+Pembaruan konfigurasi ini memerlukan reboot semua kontainer agar efektif:
+
+> ```bash
+> $ docker compose -f docker/docker-compose.yml up -d
+> ```
+
+## 🔧 Membangun Docker Image tanpa Model Embedding
+
+Image ini berukuran sekitar 2 GB dan bergantung pada aplikasi LLM eksternal dan embedding.
+
+```bash
+git clone https://github.com/infiniflow/ragflow.git
+cd ragflow/
+docker build --build-arg LIGHTEN=1 -f Dockerfile -t infiniflow/ragflow:nightly-slim .
+```
+
+## 🔧 Membangun Docker Image Termasuk Model Embedding
+
+Image ini berukuran sekitar 9 GB. Karena sudah termasuk model embedding, ia hanya bergantung pada aplikasi LLM eksternal.
+
+```bash
+git clone https://github.com/infiniflow/ragflow.git
+cd ragflow/
+docker build -f Dockerfile -t infiniflow/ragflow:nightly .
+```
+
+## 🔨 Menjalankan Aplikasi dari untuk Pengembangan
+
+1. Instal uv, atau lewati langkah ini jika sudah terinstal:
+
+   ```bash
+   pipx install uv
+   ```
+
+2. Clone kode sumber dan instal dependensi Python:
+
+   ```bash
+   git clone https://github.com/infiniflow/ragflow.git
+   cd ragflow/
+   uv sync --python 3.10 --all-extras # install RAGFlow dependent python modules
+   ```
+
+3. Jalankan aplikasi yang diperlukan (MinIO, Elasticsearch, Redis, dan MySQL) menggunakan Docker Compose:
+
+   ```bash
+   docker compose -f docker/docker-compose-base.yml up -d
+   ```
+
+   Tambahkan baris berikut ke `/etc/hosts` untuk memetakan semua host yang ditentukan di **conf/service_conf.yaml** ke `127.0.0.1`:
+
+   ```
+   127.0.0.1       es01 infinity mysql minio redis
+   ```
+
+4. Jika Anda tidak dapat mengakses HuggingFace, atur variabel lingkungan `HF_ENDPOINT` untuk menggunakan situs mirror:
+
+   ```bash
+   export HF_ENDPOINT=https://hf-mirror.com
+   ```
+
+5. Jalankan aplikasi backend:
+
+   ```bash
+   source .venv/bin/activate
+   export PYTHONPATH=$(pwd)
+   bash docker/launch_backend_service.sh
+   ```
+
+6. Instal dependensi frontend:
+   ```bash
+   cd web
+   npm install
+   ```
+7. Jalankan aplikasi frontend:
+
+   ```bash
+   npm run dev
+   ```
+
+   _Output berikut menandakan bahwa sistem berhasil diluncurkan:_
+
+   ![](https://github.com/user-attachments/assets/0daf462c-a24d-4496-a66f-92533534e187)
+
+## 📚 Dokumentasi
+
+- [Quickstart](https://ragflow.io/docs/dev/)
+- [Panduan Pengguna](https://ragflow.io/docs/dev/category/guides)
+- [Referensi](https://ragflow.io/docs/dev/category/references)
+- [FAQ](https://ragflow.io/docs/dev/faq)
+
+## 📜 Roadmap
+
+Lihat [Roadmap RAGFlow 2025](https://github.com/infiniflow/ragflow/issues/4214)
+
+## 🏄 Komunitas
+
+- [Discord](https://discord.gg/4XxujFgUN7)
+- [Twitter](https://twitter.com/infiniflowai)
+- [GitHub Discussions](https://github.com/orgs/infiniflow/discussions)
+
+## 🙌 Kontribusi
+
+RAGFlow berkembang melalui kolaborasi open-source. Dalam semangat ini, kami menerima kontribusi dari komunitas.
+Jika Anda ingin berpartisipasi, tinjau terlebih dahulu [Panduan Kontribusi](./CONTRIBUTING.md).

README_ja.md

@@ -7,8 +7,11 @@
 <p align="center">
   <a href="./README.md">English</a> |
   <a href="./README_zh.md">简体中文</a> |
+  <a href="./README_tzh.md">繁体中文</a> |
   <a href="./README_ja.md">日本語</a> |
-  <a href="./README_ko.md">한국어</a>
+  <a href="./README_ko.md">한국어</a> |
+  <a href="./README_id.md">Bahasa Indonesia</a> |
+  <a href="/README_pt_br.md">Português (Brasil)</a>
 </p>
 
 <p align="center">
@@ -19,7 +22,7 @@
         <img alt="Static Badge" src="https://img.shields.io/badge/Online-Demo-4e6b99">
     </a>
     <a href="https://hub.docker.com/r/infiniflow/ragflow" target="_blank">
-        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.13.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.13.0">
+        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.16.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.16.0">
     </a>
     <a href="https://github.com/infiniflow/ragflow/releases/latest">
         <img src="https://img.shields.io/github/v/release/infiniflow/ragflow?color=blue&label=Latest%20Release" alt="Latest Release">
@@ -29,10 +32,9 @@
     </a>
 </p>
 
-
 <h4 align="center">
   <a href="https://ragflow.io/docs/dev/">Document</a> |
-  <a href="https://github.com/infiniflow/ragflow/issues/162">Roadmap</a> |
+  <a href="https://github.com/infiniflow/ragflow/issues/4214">Roadmap</a> |
   <a href="https://twitter.com/infiniflowai">Twitter</a> |
   <a href="https://discord.gg/4XxujFgUN7">Discord</a> |
   <a href="https://demo.ragflow.io">Demo</a>
@@ -45,22 +47,26 @@
 ## 🎮 Demo
 
 デモをお試しください：[https://demo.ragflow.io](https://demo.ragflow.io)。
+
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
 <img src="https://github.com/infiniflow/ragflow/assets/7248/2f6baa3e-1092-4f11-866d-36f6a9d075e5" width="1200"/>
-<img src="https://github.com/infiniflow/ragflow/assets/12318111/b083d173-dadc-4ea9-bdeb-180d7df514eb" width="1200"/>
+<img src="https://github.com/user-attachments/assets/504bbbf1-c9f7-4d83-8cc5-e9cb63c26db6" width="1200"/>
 </div>
 
-
 ## 🔥 最新情報
 
-- 2024-09-29 マルチラウンドダイアログを最適化。
-- 2024-09-13 ナレッジベース Q&A の検索モードを追加しました。
-- 2024-09-09 エージェントに医療相談テンプレートを追加しました。
+- 2025-02-05 シリコン フローの St およびモデル リストを更新し、Deep Seek-R1/Deep Seek-V3 のサポートを追加しました。
+- 2025-01-26 ナレッジ グラフの抽出と適用を最適化し、さまざまな構成オプションを提供します。
+- 2024-12-18 Deepdoc のドキュメント レイアウト分析モデルをアップグレードします。
+- 2024-12-04 ナレッジ ベースへのページランク スコアをサポートしました。
+- 2024-11-22 エージェントでの変数の定義と使用法を改善しました。
+- 2024-11-01 再現の精度を向上させるために、解析されたチャンクにキーワード抽出と関連質問の生成を追加しました。
 - 2024-08-22 RAG を介して SQL ステートメントへのテキストをサポートします。
-- 2024-08-02 [graphrag](https://github.com/microsoft/graphrag) からインスピレーションを得た GraphRAG とマインド マップをサポートします。
 
 ## 🎉 続きを楽しみに
+
 ⭐️ リポジトリをスター登録して、エキサイティングな新機能やアップデートを最新の状態に保ちましょう！すべての新しいリリースに関する即時通知を受け取れます！ 🌟
+
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
 <img src="https://github.com/user-attachments/assets/18c9707e-b8aa-4caf-a154-037089c105ba" width="1200"/>
 </div>
@@ -140,18 +146,19 @@
 
 3. ビルド済みの Docker イメージをビルドし、サーバーを起動する:
 
-   > 以下のコマンドは、RAGFlow slim（`dev-slim`）の開発版Dockerイメージをダウンロードします。RAGFlow slimのDockerイメージには、埋め込みモデルやPythonライブラリが含まれていないため、サイズは約1GBです。
+   > 以下のコマンドは、RAGFlow Docker イメージの v0.16.0-slim エディションをダウンロードします。異なる RAGFlow エディションの説明については、以下の表を参照してください。v0.16.0-slim とは異なるエディションをダウンロードするには、docker/.env ファイルの RAGFLOW_IMAGE 変数を適宜更新し、docker compose を使用してサーバーを起動してください。例えば、完全版 v0.16.0 をダウンロードするには、RAGFLOW_IMAGE=infiniflow/ragflow:v0.16.0 と設定します。
 
    ```bash
-   $ cd ragflow/docker
-   $ docker compose -f docker-compose.yml up -d
+   $ cd ragflow
+   $ docker compose -f docker/docker-compose.yml up -d
    ```
 
-   > - 特定のバージョンのRAGFlow slim Dockerイメージをダウンロードするには、**docker/.env**内の`RAGFlow_IMAGE`変数を希望のバージョンに更新します。例えば、`RAGFLOW_IMAGE=infiniflow/ragflow:v0.13.0`とします。この変更を行った後、上記のコマンドを再実行してダウンロードを開始してください。
-   > - RAGFlowの埋め込みモデルとPythonライブラリを含む開発版Dockerイメージをダウンロードするには、**docker/.env**内の`RAGFlow_IMAGE`変数を`RAGFLOW_IMAGE=infiniflow/ragflow:dev`に更新します。この変更を行った後、上記のコマンドを再実行してダウンロードを開始してください。
-   > - 特定のバージョンのRAGFlow Dockerイメージ（埋め込みモデルとPythonライブラリを含む）をダウンロードするには、**docker/.env**内の`RAGFlow_IMAGE`変数を希望のバージョンに更新します。例えば、`RAGFLOW_IMAGE=infiniflow/ragflow:v0.13.0`とします。この変更を行った後、上記のコマンドを再実行してダウンロードを開始してください。  
-   
-   > **NOTE:** 埋め込みモデルとPythonライブラリを含むRAGFlow Dockerイメージのサイズは約9GBであり、読み込みにかなりの時間がかかる場合があります。
+   | RAGFlow image tag | Image size (GB) | Has embedding models? | Stable?                  |
+   | ----------------- | --------------- | --------------------- | ------------------------ |
+   | v0.16.0           | &approx;9       | :heavy_check_mark:    | Stable release           |
+   | v0.16.0-slim      | &approx;2       | ❌                    | Stable release           |
+   | nightly           | &approx;9       | :heavy_check_mark:    | _Unstable_ nightly build |
+   | nightly-slim      | &approx;2       | ❌                    | _Unstable_ nightly build |
 
 4. サーバーを立ち上げた後、サーバーの状態を確認する:
 
@@ -173,11 +180,12 @@
     * Running on http://x.x.x.x:9380
     INFO:werkzeug:Press CTRL+C to quit
    ```
+
    > もし確認ステップをスキップして直接 RAGFlow にログインした場合、その時点で RAGFlow が完全に初期化されていない可能性があるため、ブラウザーがネットワーク異常エラーを表示するかもしれません。
 
 5. ウェブブラウザで、プロンプトに従ってサーバーの IP アドレスを入力し、RAGFlow にログインします。
    > デフォルトの設定を使用する場合、デフォルトの HTTP サービングポート `80` は省略できるので、与えられたシナリオでは、`http://IP_OF_YOUR_MACHINE`（ポート番号は省略）だけを入力すればよい。
-6. [service_conf.yaml](./docker/service_conf.yaml) で、`user_default_llm` で希望の LLM ファクトリを選択し、`API_KEY` フィールドを対応する API キーで更新する。
+6. [service_conf.yaml.template](./docker/service_conf.yaml.template) で、`user_default_llm` で希望の LLM ファクトリを選択し、`API_KEY` フィールドを対応する API キーで更新する。
 
    > 詳しくは [llm_api_key_setup](https://ragflow.io/docs/dev/llm_api_key_setup) を参照してください。
 
@@ -188,12 +196,12 @@
 システムコンフィグに関しては、以下のファイルを管理する必要がある:
 
 - [.env](./docker/.env): `SVR_HTTP_PORT`、`MYSQL_PASSWORD`、`MINIO_PASSWORD` などのシステムの基本設定を保持する。
-- [service_conf.yaml](./docker/service_conf.yaml): バックエンドのサービスを設定します。
+- [service_conf.yaml.template](./docker/service_conf.yaml.template): バックエンドのサービスを設定します。
 - [docker-compose.yml](./docker/docker-compose.yml): システムの起動は [docker-compose.yml](./docker/docker-compose.yml) に依存している。
 
-[.env](./docker/.env) ファイルの変更が [service_conf.yaml](./docker/service_conf.yaml) ファイルの内容と一致していることを確認する必要があります。
+[.env](./docker/.env) ファイルの変更が [service_conf.yaml.template](./docker/service_conf.yaml.template) ファイルの内容と一致していることを確認する必要があります。
 
-> [./docker/README](./docker/README.md) ファイルは環境設定とサービスコンフィグの詳細な説明を提供し、[./docker/README](./docker/README.md) ファイルに記載されている全ての環境設定が [service_conf.yaml](./docker/service_conf.yaml) ファイルの対応するコンフィグと一致していることを確認することが義務付けられています。
+> [./docker/README](./docker/README.md) ファイル ./docker/README には、service_conf.yaml.template ファイルで ${ENV_VARS} として使用できる環境設定とサービス構成の詳細な説明が含まれています。
 
 デフォルトの HTTP サービングポート(80)を更新するには、[docker-compose.yml](./docker/docker-compose.yml) にアクセスして、`80:80` を `<YOUR_SERVING_PORT>:80` に変更します。
 
@@ -203,6 +211,23 @@
 > $ docker compose -f docker/docker-compose.yml up -d
 > ```
 
+### Elasticsearch から Infinity にドキュメントエンジンを切り替えます
+
+RAGFlow はデフォルトで Elasticsearch を使用して全文とベクトルを保存します。［Infinity］に切り替え（https://github.com/infiniflow/infinity/)、次の手順に従います。
+
+1. 実行中のすべてのコンテナを停止するには：
+   ```bash
+   $ docker compose -f docker/docker-compose.yml down -v
+   ```
+2. **docker/.env** の「DOC \_ ENGINE」を「infinity」に設定します。
+
+3. 起動コンテナ：
+   ```bash
+   $ docker compose -f docker/docker-compose.yml up -d
+   ```
+   > [!WARNING]  
+   > Linux/arm64 マシンでの Infinity への切り替えは正式にサポートされていません。
+
 ## 🔧 ソースコードで Docker イメージを作成（埋め込みモデルなし）
 
 この Docker イメージのサイズは約 1GB で、外部の大モデルと埋め込みサービスに依存しています。
@@ -210,9 +235,7 @@
 ```bash
 git clone https://github.com/infiniflow/ragflow.git
 cd ragflow/
-pip3 install huggingface-hub nltk
-python3 download_deps.py
-docker build -f Dockerfile.slim -t infiniflow/ragflow:dev-slim .
+docker build --build-arg LIGHTEN=1 -f Dockerfile -t infiniflow/ragflow:nightly-slim .
 ```
 
 ## 🔧 ソースコードをコンパイルした Docker イメージ（埋め込みモデルを含む）
@@ -222,36 +245,36 @@ docker build -f Dockerfile.slim -t infiniflow/ragflow:dev-slim .
 ```bash
 git clone https://github.com/infiniflow/ragflow.git
 cd ragflow/
-pip3 install huggingface-hub nltk
-python3 download_deps.py
-docker build -f Dockerfile -t infiniflow/ragflow:dev .
+docker build -f Dockerfile -t infiniflow/ragflow:nightly .
 ```
 
 ## 🔨 ソースコードからサービスを起動する方法
 
-1. Poetry をインストールする。すでにインストールされている場合は、このステップをスキップしてください:
+1. uv をインストールする。すでにインストールされている場合は、このステップをスキップしてください:
+
    ```bash
-   curl -sSL https://install.python-poetry.org | python3 -
+   pipx install uv
    ```
 
 2. ソースコードをクローンし、Python の依存関係をインストールする:
+
    ```bash
    git clone https://github.com/infiniflow/ragflow.git
    cd ragflow/
-   export POETRY_VIRTUALENVS_CREATE=true POETRY_VIRTUALENVS_IN_PROJECT=true
-   ~/.local/bin/poetry install --sync --no-root # install RAGFlow dependent python modules
+   uv sync --python 3.10 --all-extras # install RAGFlow dependent python modules
    ```
 
 3. Docker Compose を使用して依存サービス（MinIO、Elasticsearch、Redis、MySQL）を起動する:
+
    ```bash
    docker compose -f docker/docker-compose-base.yml up -d
    ```
 
-   `/etc/hosts` に以下の行を追加して、**docker/service_conf.yaml** に指定されたすべてのホストを `127.0.0.1` に解決します:  
+   `/etc/hosts` に以下の行を追加して、**conf/service_conf.yaml** に指定されたすべてのホストを `127.0.0.1` に解決します:
+
    ```
-   127.0.0.1       es01 mysql minio redis
+   127.0.0.1       es01 infinity mysql minio redis
    ```
-   **docker/service_conf.yaml** で mysql のポートを `5455` に、es のポートを `1200` に更新します（**docker/.env** に指定された通り）.
 
 4. HuggingFace にアクセスできない場合は、`HF_ENDPOINT` 環境変数を設定してミラーサイトを使用してください:
 
@@ -260,6 +283,7 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
    ```
 
 5. バックエンドサービスを起動する:
+
    ```bash
    source .venv/bin/activate
    export PYTHONPATH=$(pwd)
@@ -269,10 +293,10 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
 6. フロントエンドの依存関係をインストールする:
    ```bash
    cd web
-   npm install --force
+   npm install
    ```
-7. フロントエンドを設定し、**.umirc.ts** の `proxy.target` を `http://127.0.0.1:9380` に更新します:
-8. フロントエンドサービスを起動する:  
+7. フロントエンドサービスを起動する:
+
    ```bash
    npm run dev
    ```
@@ -290,7 +314,7 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
 
 ## 📜 ロードマップ
 
-[RAGFlow ロードマップ 2024](https://github.com/infiniflow/ragflow/issues/162) を参照
+[RAGFlow ロードマップ 2025](https://github.com/infiniflow/ragflow/issues/4214) を参照
 
 ## 🏄 コミュニティ
 

README_ko.md

@@ -7,8 +7,11 @@
 <p align="center">
   <a href="./README.md">English</a> |
   <a href="./README_zh.md">简体中文</a> |
+  <a href="./README_tzh.md">繁体中文</a> |
   <a href="./README_ja.md">日本語</a> |
   <a href="./README_ko.md">한국어</a> |
+  <a href="./README_id.md">Bahasa Indonesia</a> |
+  <a href="/README_pt_br.md">Português (Brasil)</a>
 </p>
 
 <p align="center">
@@ -19,7 +22,7 @@
         <img alt="Static Badge" src="https://img.shields.io/badge/Online-Demo-4e6b99">
     </a>
     <a href="https://hub.docker.com/r/infiniflow/ragflow" target="_blank">
-        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.13.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.13.0">
+        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.16.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.16.0">
     </a>
     <a href="https://github.com/infiniflow/ragflow/releases/latest">
         <img src="https://img.shields.io/github/v/release/infiniflow/ragflow?color=blue&label=Latest%20Release" alt="Latest Release">
@@ -29,76 +32,74 @@
     </a>
 </p>
 
-
 <h4 align="center">
   <a href="https://ragflow.io/docs/dev/">Document</a> |
-  <a href="https://github.com/infiniflow/ragflow/issues/162">Roadmap</a> |
+  <a href="https://github.com/infiniflow/ragflow/issues/4214">Roadmap</a> |
   <a href="https://twitter.com/infiniflowai">Twitter</a> |
   <a href="https://discord.gg/4XxujFgUN7">Discord</a> |
   <a href="https://demo.ragflow.io">Demo</a>
 </h4>
 
-
 ## 💡 RAGFlow란?
 
 [RAGFlow](https://ragflow.io/)는 심층 문서 이해에 기반한 오픈소스 RAG (Retrieval-Augmented Generation) 엔진입니다. 이 엔진은 대규모 언어 모델(LLM)과 결합하여 정확한 질문 응답 기능을 제공하며, 다양한 복잡한 형식의 데이터에서 신뢰할 수 있는 출처를 바탕으로 한 인용을 통해 이를 뒷받침합니다. RAGFlow는 규모에 상관없이 모든 기업에 최적화된 RAG 워크플로우를 제공합니다.
 
-
-
 ## 🎮 데모
+
 데모를 [https://demo.ragflow.io](https://demo.ragflow.io)에서 실행해 보세요.
+
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
 <img src="https://github.com/infiniflow/ragflow/assets/7248/2f6baa3e-1092-4f11-866d-36f6a9d075e5" width="1200"/>
-<img src="https://github.com/infiniflow/ragflow/assets/12318111/b083d173-dadc-4ea9-bdeb-180d7df514eb" width="1200"/>
+<img src="https://github.com/user-attachments/assets/504bbbf1-c9f7-4d83-8cc5-e9cb63c26db6" width="1200"/>
 </div>
 
-
 ## 🔥 업데이트
 
-- 2024-09-29 다단계 대화를 최적화합니다.
-  
-- 2024-09-13 지식베이스 Q&A 검색 모드를 추가합니다.
-  
-- 2024-09-09 Agent에 의료상담 템플릿을 추가하였습니다.
+- 2025-02-05 'SILICONFLOW' 모델 목록을 업데이트하고 Deepseek-R1/DeepSeek-V3에 대한 지원을 추가합니다.
+- 2025-01-26 지식 그래프 추출 및 적용을 최적화하고 다양한 구성 옵션을 제공합니다.
+- 2024-12-18 Deepdoc의 문서 레이아웃 분석 모델 업그레이드.
+- 2024-12-04 지식베이스에 대한 페이지랭크 점수를 지원합니다.
 
+- 2024-11-22 에이전트의 변수 정의 및 사용을 개선했습니다.
+- 2024-11-01 파싱된 청크에 키워드 추출 및 관련 질문 생성을 추가하여 재현율을 향상시킵니다.
 - 2024-08-22 RAG를 통해 SQL 문에 텍스트를 지원합니다.
 
-- 2024-08-02: [graphrag](https://github.com/microsoft/graphrag)와 마인드맵에서 영감을 받은 GraphRAG를 지원합니다.
-
-
 ## 🎉 계속 지켜봐 주세요
+
 ⭐️우리의 저장소를 즐겨찾기에 등록하여 흥미로운 새로운 기능과 업데이트를 최신 상태로 유지하세요! 모든 새로운 릴리스에 대한 즉시 알림을 받으세요! 🌟
+
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
 <img src="https://github.com/user-attachments/assets/18c9707e-b8aa-4caf-a154-037089c105ba" width="1200"/>
 </div>
 
-
 ## 🌟 주요 기능
 
 ### 🍭 **"Quality in, quality out"**
+
 - [심층 문서 이해](./deepdoc/README.md)를 기반으로 복잡한 형식의 비정형 데이터에서 지식을 추출합니다.
 - 문자 그대로 무한한 토큰에서 "데이터 속의 바늘"을 찾아냅니다.
 
 ### 🍱 **템플릿 기반의 chunking**
+
 - 똑똑하고 설명 가능한 방식.
 - 다양한 템플릿 옵션을 제공합니다.
 
-
 ### 🌱 **할루시네이션을 줄인 신뢰할 수 있는 인용**
+
 - 텍스트 청킹을 시각화하여 사용자가 개입할 수 있도록 합니다.
 - 중요한 참고 자료와 추적 가능한 인용을 빠르게 확인하여 신뢰할 수 있는 답변을 지원합니다.
 
-
 ### 🍔 **다른 종류의 데이터 소스와의 호환성**
+
 - 워드, 슬라이드, 엑셀, 텍스트 파일, 이미지, 스캔본, 구조화된 데이터, 웹 페이지 등을 지원합니다.
 
 ### 🛀 **자동화되고 손쉬운 RAG 워크플로우**
+
 - 개인 및 대규모 비즈니스에 맞춘 효율적인 RAG 오케스트레이션.
 - 구성 가능한 LLM 및 임베딩 모델.
 - 다중 검색과 결합된 re-ranking.
 - 비즈니스와 원활하게 통합할 수 있는 직관적인 API.
 
-
 ## 🔎 시스템 아키텍처
 
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
@@ -106,17 +107,19 @@
 </div>
 
 ## 🎬 시작하기
+
 ### 📝 사전 준비 사항
+
 - CPU >= 4 cores
 - RAM >= 16 GB
 - Disk >= 50 GB
 - Docker >= 24.0.0 & Docker Compose >= v2.26.1
-  > 로컬 머신(Windows, Mac, Linux)에 Docker가 설치되지 않은 경우, [Docker 엔진 설치]((https://docs.docker.com/engine/install/))를 참조하세요.
-
+  > 로컬 머신(Windows, Mac, Linux)에 Docker가 설치되지 않은 경우, [Docker 엔진 설치](<(https://docs.docker.com/engine/install/)>)를 참조하세요.
 
 ### 🚀 서버 시작하기
 
 1. `vm.max_map_count`가 262144 이상인지 확인하세요:
+
    > `vm.max_map_count`의 값을 아래 명령어를 통해 확인하세요:
    >
    > ```bash
@@ -144,19 +147,19 @@
 
 3. 미리 빌드된 Docker 이미지를 생성하고 서버를 시작하세요:
 
-   > 아래의 명령은 RAGFlow slim(dev-slim)의 개발 버전 Docker 이미지를 다운로드합니다. RAGFlow slim Docker 이미지에는 임베딩 모델이나 Python 라이브러리가 포함되어 있지 않으므로 크기는 약 1GB입니다.
+   > 아래 명령어는 RAGFlow Docker 이미지의 v0.16.0-slim 버전을 다운로드합니다. 다양한 RAGFlow 버전에 대한 설명은 다음 표를 참조하십시오. v0.16.0-slim과 다른 RAGFlow 버전을 다운로드하려면, docker/.env 파일에서 RAGFLOW_IMAGE 변수를 적절히 업데이트한 후 docker compose를 사용하여 서버를 시작하십시오. 예를 들어, 전체 버전인 v0.16.0을 다운로드하려면 RAGFLOW_IMAGE=infiniflow/ragflow:v0.16.0로 설정합니다.
 
    ```bash
-   $ cd ragflow/docker
-   $ docker compose -f docker-compose.yml up -d
+   $ cd ragflow
+   $ docker compose -f docker/docker-compose.yml up -d
    ```
 
-   > - 특정 버전의 RAGFlow slim Docker 이미지를 다운로드하려면, **docker/.env**에서 `RAGFlow_IMAGE` 변수를 원하는 버전으로 업데이트하세요. 예를 들어, `RAGFLOW_IMAGE=infiniflow/ragflow:v0.13.0-slim`으로 설정합니다. 이 변경을 완료한 후, 위의 명령을 다시 실행하여 다운로드를 시작하세요.
-   > - RAGFlow의 임베딩 모델과 Python 라이브러리를 포함한 개발 버전 Docker 이미지를 다운로드하려면, **docker/.env**에서 `RAGFlow_IMAGE` 변수를 `RAGFLOW_IMAGE=infiniflow/ragflow:dev`로 업데이트하세요. 이 변경을 완료한 후, 위의 명령을 다시 실행하여 다운로드를 시작하세요.
-   > - 특정 버전의 RAGFlow Docker 이미지를 임베딩 모델과 Python 라이브러리를 포함하여 다운로드하려면, **docker/.env**에서 `RAGFlow_IMAGE` 변수를 원하는 버전으로 업데이트하세요. 예를 들어, `RAGFLOW_IMAGE=infiniflow/ragflow:v0.13.0` 로 설정합니다. 이 변경을 완료한 후, 위의 명령을 다시 실행하여 다운로드를 시작하세요.  
-   
-   > **NOTE:** 임베딩 모델과 Python 라이브러리를 포함한 RAGFlow Docker 이미지의 크기는 약 9GB이며, 로드하는 데 상당히 오랜 시간이 걸릴 수 있습니다.
-
+   | RAGFlow image tag | Image size (GB) | Has embedding models? | Stable?                  |
+   | ----------------- | --------------- | --------------------- | ------------------------ |
+   | v0.16.0           | &approx;9       | :heavy_check_mark:    | Stable release           |
+   | v0.16.0-slim      | &approx;2       | ❌                    | Stable release           |
+   | nightly           | &approx;9       | :heavy_check_mark:    | _Unstable_ nightly build |
+   | nightly-slim      | &approx;2       | ❌                    | _Unstable_ nightly build |
 
 4. 서버가 시작된 후 서버 상태를 확인하세요:
 
@@ -178,11 +181,13 @@
     * Running on http://x.x.x.x:9380
     INFO:werkzeug:Press CTRL+C to quit
    ```
-   > 만약 확인 단계를 건너뛰고 바로 RAGFlow에 로그인하면, RAGFlow가 완전히 초기화되지 않았기 때문에 브라우저에서 `network abnormal` 오류가 발생할 수 있습니다.
+
+   > 만약 확인 단계를 건너뛰고 바로 RAGFlow에 로그인하면, RAGFlow가 완전히 초기화되지 않았기 때문에 브라우저에서 `network anormal` 오류가 발생할 수 있습니다.
 
 5. 웹 브라우저에 서버의 IP 주소를 입력하고 RAGFlow에 로그인하세요.
    > 기본 설정을 사용할 경우, `http://IP_OF_YOUR_MACHINE`만 입력하면 됩니다 (포트 번호는 제외). 기본 HTTP 서비스 포트 `80`은 기본 구성으로 사용할 때 생략할 수 있습니다.
-6. [service_conf.yaml](./docker/service_conf.yaml) 파일에서 원하는 LLM 팩토리를 `user_default_llm`에 선택하고, `API_KEY` 필드를 해당 API 키로 업데이트하세요.
+6. [service_conf.yaml.template](./docker/service_conf.yaml.template) 파일에서 원하는 LLM 팩토리를 `user_default_llm`에 선택하고, `API_KEY` 필드를 해당 API 키로 업데이트하세요.
+
    > 자세한 내용은 [llm_api_key_setup](https://ragflow.io/docs/dev/llm_api_key_setup)를 참조하세요.
 
    _이제 쇼가 시작됩니다!_
@@ -192,12 +197,12 @@
 시스템 설정과 관련하여 다음 파일들을 관리해야 합니다:
 
 - [.env](./docker/.env): `SVR_HTTP_PORT`, `MYSQL_PASSWORD`, `MINIO_PASSWORD`와 같은 시스템의 기본 설정을 포함합니다.
-- [service_conf.yaml](./docker/service_conf.yaml): 백엔드 서비스를 구성합니다.
+- [service_conf.yaml.template](./docker/service_conf.yaml.template): 백엔드 서비스를 구성합니다.
 - [docker-compose.yml](./docker/docker-compose.yml): 시스템은 [docker-compose.yml](./docker/docker-compose.yml)을 사용하여 시작됩니다.
 
-[.env](./docker/.env) 파일의 변경 사항이 [service_conf.yaml](./docker/service_conf.yaml) 파일의 내용과 일치하도록 해야 합니다.
+[.env](./docker/.env) 파일의 변경 사항이 [service_conf.yaml.template](./docker/service_conf.yaml.template) 파일의 내용과 일치하도록 해야 합니다.
 
-> [./docker/README](./docker/README.md) 파일에는 환경 설정과 서비스 구성에 대한 자세한 설명이 있으며, [./docker/README](./docker/README.md) 파일에 나열된 모든 환경 설정이 [service_conf.yaml](./docker/service_conf.yaml) 파일의 해당 구성과 일치하도록 해야 합니다.
+> [./docker/README](./docker/README.md) 파일 ./docker/README은 service_conf.yaml.template 파일에서 ${ENV_VARS}로 사용할 수 있는 환경 설정과 서비스 구성에 대한 자세한 설명을 제공합니다.
 
 기본 HTTP 서비스 포트(80)를 업데이트하려면 [docker-compose.yml](./docker/docker-compose.yml) 파일에서 `80:80`을 `<YOUR_SERVING_PORT>:80`으로 변경하세요.
 
@@ -207,6 +212,22 @@
 > $ docker compose -f docker/docker-compose.yml up -d
 > ```
 
+### Elasticsearch 에서 Infinity 로 문서 엔진 전환
+
+RAGFlow 는 기본적으로 Elasticsearch 를 사용하여 전체 텍스트 및 벡터를 저장합니다. [Infinity]로 전환(https://github.com/infiniflow/infinity/), 다음 절차를 따르십시오.
+
+1. 실행 중인 모든 컨테이너를 중지합니다.
+   ```bash
+   $docker compose-f docker/docker-compose.yml down -v
+   ```
+2. **docker/.env**의 "DOC_ENGINE" 을 "infinity" 로 설정합니다.
+3. 컨테이너 부팅:
+   ```bash
+   $docker compose-f docker/docker-compose.yml up -d
+   ```
+   > [!WARNING]
+   > Linux/arm64 시스템에서 Infinity로 전환하는 것은 공식적으로 지원되지 않습니다.
+
 ## 🔧 소스 코드로 Docker 이미지를 컴파일합니다(임베딩 모델 포함하지 않음)
 
 이 Docker 이미지의 크기는 약 1GB이며, 외부 대형 모델과 임베딩 서비스에 의존합니다.
@@ -214,9 +235,7 @@
 ```bash
 git clone https://github.com/infiniflow/ragflow.git
 cd ragflow/
-pip3 install huggingface-hub nltk
-python3 download_deps.py
-docker build -f Dockerfile.slim -t infiniflow/ragflow:dev-slim .
+docker build --build-arg LIGHTEN=1 -f Dockerfile -t infiniflow/ragflow:nightly-slim .
 ```
 
 ## 🔧 소스 코드로 Docker 이미지를 컴파일합니다(임베딩 모델 포함)
@@ -226,36 +245,36 @@ docker build -f Dockerfile.slim -t infiniflow/ragflow:dev-slim .
 ```bash
 git clone https://github.com/infiniflow/ragflow.git
 cd ragflow/
-pip3 install huggingface-hub nltk
-python3 download_deps.py
-docker build -f Dockerfile -t infiniflow/ragflow:dev .
+docker build -f Dockerfile -t infiniflow/ragflow:nightly .
 ```
 
 ## 🔨 소스 코드로 서비스를 시작합니다.
 
-1. Poetry를 설치하거나 이미 설치된 경우 이 단계를 건너뜁니다:
+1. uv를 설치하거나 이미 설치된 경우 이 단계를 건너뜁니다:
+
    ```bash
-   curl -sSL https://install.python-poetry.org | python3 -
+   pipx install uv
    ```
 
 2. 소스 코드를 클론하고 Python 의존성을 설치합니다:
+
    ```bash
    git clone https://github.com/infiniflow/ragflow.git
    cd ragflow/
-   export POETRY_VIRTUALENVS_CREATE=true POETRY_VIRTUALENVS_IN_PROJECT=true
-   ~/.local/bin/poetry install --sync --no-root # install RAGFlow dependent python modules
+   uv sync --python 3.10 --all-extras # install RAGFlow dependent python modules
    ```
 
 3. Docker Compose를 사용하여 의존 서비스(MinIO, Elasticsearch, Redis 및 MySQL)를 시작합니다:
+
    ```bash
    docker compose -f docker/docker-compose-base.yml up -d
    ```
 
-   `/etc/hosts` 에 다음 줄을 추가하여 **docker/service_conf.yaml** 에 지정된 모든 호스트를 `127.0.0.1` 로 해결합니다:  
+   `/etc/hosts` 에 다음 줄을 추가하여 **conf/service_conf.yaml** 에 지정된 모든 호스트를 `127.0.0.1` 로 해결합니다:
+
    ```
-   127.0.0.1       es01 mysql minio redis
+   127.0.0.1       es01 infinity mysql minio redis
    ```
-   **docker/service_conf.yaml** 에서 mysql 포트를 `5455` 로, es 포트를 `1200` 으로 업데이트합니다( **docker/.env** 에 지정된 대로).
 
 4. HuggingFace에 접근할 수 없는 경우, `HF_ENDPOINT` 환경 변수를 설정하여 미러 사이트를 사용하세요:
 
@@ -264,6 +283,7 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
    ```
 
 5. 백엔드 서비스를 시작합니다:
+
    ```bash
    source .venv/bin/activate
    export PYTHONPATH=$(pwd)
@@ -273,10 +293,10 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
 6. 프론트엔드 의존성을 설치합니다:
    ```bash
    cd web
-   npm install --force
+   npm install
    ```
-7. **.umirc.ts** 에서 `proxy.target` 을 `http://127.0.0.1:9380` 으로 업데이트합니다:
-8. 프론트엔드 서비스를 시작합니다:  
+7. 프론트엔드 서비스를 시작합니다:
+
    ```bash
    npm run dev
    ```
@@ -294,7 +314,7 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
 
 ## 📜 로드맵
 
-[RAGFlow 로드맵 2024](https://github.com/infiniflow/ragflow/issues/162)을 확인하세요.
+[RAGFlow 로드맵 2025](https://github.com/infiniflow/ragflow/issues/4214)을 확인하세요.
 
 ## 🏄 커뮤니티
 

README_pt_br.md

@@ -0,0 +1,354 @@
+<div align="center">
+<a href="https://demo.ragflow.io/">
+<img src="web/src/assets/logo-with-text.png" width="520" alt="ragflow logo">
+</a>
+</div>
+
+<p align="center">
+  <a href="./README.md">English</a> |
+  <a href="./README_zh.md">简体中文</a> |
+  <a href="./README_tzh.md">繁体中文</a> |
+  <a href="./README_ja.md">日本語</a> |
+  <a href="./README_ko.md">한국어</a> |
+  <a href="./README_id.md">Bahasa Indonesia</a> |
+  <a href="/README_pt_br.md">Português (Brasil)</a>
+</p>
+
+<p align="center">
+    <a href="https://x.com/intent/follow?screen_name=infiniflowai" target="_blank">
+        <img src="https://img.shields.io/twitter/follow/infiniflow?logo=X&color=%20%23f5f5f5" alt="seguir no X(Twitter)">
+    </a>
+    <a href="https://demo.ragflow.io" target="_blank">
+        <img alt="Badge Estático" src="https://img.shields.io/badge/Online-Demo-4e6b99">
+    </a>
+    <a href="https://hub.docker.com/r/infiniflow/ragflow" target="_blank">
+        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.16.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.16.0">
+    </a>
+    <a href="https://github.com/infiniflow/ragflow/releases/latest">
+        <img src="https://img.shields.io/github/v/release/infiniflow/ragflow?color=blue&label=Última%20Relese" alt="Última Versão">
+    </a>
+    <a href="https://github.com/infiniflow/ragflow/blob/main/LICENSE">
+        <img height="21" src="https://img.shields.io/badge/License-Apache--2.0-ffffff?labelColor=d4eaf7&color=2e6cc4" alt="licença">
+    </a>
+</p>
+
+<h4 align="center">
+  <a href="https://ragflow.io/docs/dev/">Documentação</a> |
+  <a href="https://github.com/infiniflow/ragflow/issues/4214">Roadmap</a> |
+  <a href="https://twitter.com/infiniflowai">Twitter</a> |
+  <a href="https://discord.gg/4XxujFgUN7">Discord</a> |
+  <a href="https://demo.ragflow.io">Demo</a>
+</h4>
+
+<details open>
+<summary></b>📕 Índice</b></summary>
+
+- 💡 [O que é o RAGFlow?](#-o-que-é-o-ragflow)
+- 🎮 [Demo](#-demo)
+- 📌 [Últimas Atualizações](#-últimas-atualizações)
+- 🌟 [Principais Funcionalidades](#-principais-funcionalidades)
+- 🔎 [Arquitetura do Sistema](#-arquitetura-do-sistema)
+- 🎬 [Primeiros Passos](#-primeiros-passos)
+- 🔧 [Configurações](#-configurações)
+- 🔧 [Construir uma imagem docker sem incorporar modelos](#-construir-uma-imagem-docker-sem-incorporar-modelos)
+- 🔧 [Construir uma imagem docker incluindo modelos](#-construir-uma-imagem-docker-incluindo-modelos)
+- 🔨 [Lançar serviço a partir do código-fonte para desenvolvimento](#-lançar-serviço-a-partir-do-código-fonte-para-desenvolvimento)
+- 📚 [Documentação](#-documentação)
+- 📜 [Roadmap](#-roadmap)
+- 🏄 [Comunidade](#-comunidade)
+- 🙌 [Contribuindo](#-contribuindo)
+
+</details>
+
+## 💡 O que é o RAGFlow?
+
+[RAGFlow](https://ragflow.io/) é um mecanismo RAG (Geração Aumentada por Recuperação) de código aberto baseado em entendimento profundo de documentos. Ele oferece um fluxo de trabalho RAG simplificado para empresas de qualquer porte, combinando LLMs (Modelos de Linguagem de Grande Escala) para fornecer capacidades de perguntas e respostas verídicas, respaldadas por citações bem fundamentadas de diversos dados complexos formatados.
+
+## 🎮 Demo
+
+Experimente nossa demo em [https://demo.ragflow.io](https://demo.ragflow.io).
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/7248/2f6baa3e-1092-4f11-866d-36f6a9d075e5" width="1200"/>
+<img src="https://github.com/user-attachments/assets/504bbbf1-c9f7-4d83-8cc5-e9cb63c26db6" width="1200"/>
+</div>
+
+## 🔥 Últimas Atualizações
+
+- 05-02-2025 Atualiza a lista de modelos de 'SILICONFLOW' e adiciona suporte para Deepseek-R1/DeepSeek-V3.
+- 26-01-2025 Otimize a extração e aplicação de gráficos de conhecimento e forneça uma variedade de opções de configuração.
+- 18-12-2024 Atualiza o modelo de Análise de Layout de Documentos no Deepdoc.
+- 04-12-2024 Adiciona suporte para pontuação de pagerank na base de conhecimento.
+- 22-11-2024 Adiciona mais variáveis para o Agente.
+- 01-11-2024 Adiciona extração de palavras-chave e geração de perguntas relacionadas aos blocos analisados para melhorar a precisão da recuperação.
+- 22-08-2024 Suporta conversão de texto para comandos SQL via RAG.
+
+## 🎉 Fique Ligado
+
+⭐️ Dê uma estrela no nosso repositório para se manter atualizado com novas funcionalidades e melhorias empolgantes! Receba notificações instantâneas sobre novos lançamentos! 🌟
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/user-attachments/assets/18c9707e-b8aa-4caf-a154-037089c105ba" width="1200"/>
+</div>
+
+## 🌟 Principais Funcionalidades
+
+### 🍭 **"Qualidade entra, qualidade sai"**
+
+- Extração de conhecimento baseada em [entendimento profundo de documentos](./deepdoc/README.md) a partir de dados não estruturados com formatos complicados.
+- Encontra a "agulha no palheiro de dados" de literalmente tokens ilimitados.
+
+### 🍱 **Fragmentação baseada em templates**
+
+- Inteligente e explicável.
+- Muitas opções de templates para escolher.
+
+### 🌱 **Citações fundamentadas com menos alucinações**
+
+- Visualização da fragmentação de texto para permitir intervenção humana.
+- Visualização rápida das referências chave e citações rastreáveis para apoiar respostas fundamentadas.
+
+### 🍔 **Compatibilidade com fontes de dados heterogêneas**
+
+- Suporta Word, apresentações, excel, txt, imagens, cópias digitalizadas, dados estruturados, páginas da web e mais.
+
+### 🛀 **Fluxo de trabalho RAG automatizado e sem esforço**
+
+- Orquestração RAG simplificada voltada tanto para negócios pessoais quanto grandes empresas.
+- Modelos LLM e de incorporação configuráveis.
+- Múltiplas recuperações emparelhadas com reclassificação fundida.
+- APIs intuitivas para integração sem problemas com os negócios.
+
+## 🔎 Arquitetura do Sistema
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/12318111/d6ac5664-c237-4200-a7c2-a4a00691b485" width="1000"/>
+</div>
+
+## 🎬 Primeiros Passos
+
+### 📝 Pré-requisitos
+
+- CPU >= 4 núcleos
+- RAM >= 16 GB
+- Disco >= 50 GB
+- Docker >= 24.0.0 & Docker Compose >= v2.26.1
+  > Se você não instalou o Docker na sua máquina local (Windows, Mac ou Linux), veja [Instalar Docker Engine](https://docs.docker.com/engine/install/).
+
+### 🚀 Iniciar o servidor
+
+1.  Certifique-se de que `vm.max_map_count` >= 262144:
+
+    > Para verificar o valor de `vm.max_map_count`:
+    >
+    > ```bash
+    > $ sysctl vm.max_map_count
+    > ```
+    >
+    > Se necessário, redefina `vm.max_map_count` para um valor de pelo menos 262144:
+    >
+    > ```bash
+    > # Neste caso, defina para 262144:
+    > $ sudo sysctl -w vm.max_map_count=262144
+    > ```
+    >
+    > Essa mudança será resetada após a reinicialização do sistema. Para garantir que a alteração permaneça permanente, adicione ou atualize o valor de `vm.max_map_count` em **/etc/sysctl.conf**:
+    >
+    > ```bash
+    > vm.max_map_count=262144
+    > ```
+
+2.  Clone o repositório:
+
+    ```bash
+    $ git clone https://github.com/infiniflow/ragflow.git
+    ```
+
+3.  Inicie o servidor usando as imagens Docker pré-compiladas:
+
+    > O comando abaixo baixa a edição `v0.16.0-slim` da imagem Docker do RAGFlow. Consulte a tabela a seguir para descrições de diferentes edições do RAGFlow. Para baixar uma edição do RAGFlow diferente da `v0.16.0-slim`, atualize a variável `RAGFLOW_IMAGE` conforme necessário no **docker/.env** antes de usar `docker compose` para iniciar o servidor. Por exemplo: defina `RAGFLOW_IMAGE=infiniflow/ragflow:v0.16.0` para a edição completa `v0.16.0`.
+
+    ```bash
+    $ cd ragflow
+    $ docker compose -f docker/docker-compose.yml up -d
+    ```
+
+    | Tag da imagem RAGFlow | Tamanho da imagem (GB) | Possui modelos de incorporação? | Estável?                 |
+    | --------------------- | ---------------------- | ------------------------------- | ------------------------ |
+    | v0.16.0               | ~9                     | :heavy_check_mark:              | Lançamento estável       |
+    | v0.16.0-slim          | ~2                     | ❌                              | Lançamento estável       |
+    | nightly               | ~9                     | :heavy_check_mark:              | _Instável_ build noturno |
+    | nightly-slim          | ~2                     | ❌                              | _Instável_ build noturno |
+
+4.  Verifique o status do servidor após tê-lo iniciado:
+
+    ```bash
+    $ docker logs -f ragflow-server
+    ```
+
+    _O seguinte resultado confirma o lançamento bem-sucedido do sistema:_
+
+    ```bash
+         ____   ___    ______ ______ __
+        / __ \ /   |  / ____// ____// /____  _      __
+       / /_/ // /| | / / __ / /_   / // __ \| | /| / /
+      / _, _// ___ |/ /_/ // __/  / // /_/ /| |/ |/ /
+     /_/ |_|/_/  |_|\____//_/    /_/ \____/ |__/|__/
+
+     * Rodando em todos os endereços (0.0.0.0)
+     * Rodando em http://127.0.0.1:9380
+     * Rodando em http://x.x.x.x:9380
+     INFO:werkzeug:Pressione CTRL+C para sair
+    ```
+
+    > Se você pular essa etapa de confirmação e acessar diretamente o RAGFlow, seu navegador pode exibir um erro `network anormal`, pois, nesse momento, seu RAGFlow pode não estar totalmente inicializado.
+
+5.  No seu navegador, insira o endereço IP do seu servidor e faça login no RAGFlow.
+
+    > Com as configurações padrão, você só precisa digitar `http://IP_DO_SEU_MÁQUINA` (**sem** o número da porta), pois a porta HTTP padrão `80` pode ser omitida ao usar as configurações padrão.
+
+6.  Em [service_conf.yaml.template](./docker/service_conf.yaml.template), selecione a fábrica LLM desejada em `user_default_llm` e atualize o campo `API_KEY` com a chave de API correspondente.
+
+    > Consulte [llm_api_key_setup](https://ragflow.io/docs/dev/llm_api_key_setup) para mais informações.
+
+_O show está no ar!_
+
+## 🔧 Configurações
+
+Quando se trata de configurações do sistema, você precisará gerenciar os seguintes arquivos:
+
+- [.env](./docker/.env): Contém as configurações fundamentais para o sistema, como `SVR_HTTP_PORT`, `MYSQL_PASSWORD` e `MINIO_PASSWORD`.
+- [service_conf.yaml.template](./docker/service_conf.yaml.template): Configura os serviços de back-end. As variáveis de ambiente neste arquivo serão automaticamente preenchidas quando o contêiner Docker for iniciado. Quaisquer variáveis de ambiente definidas dentro do contêiner Docker estarão disponíveis para uso, permitindo personalizar o comportamento do serviço com base no ambiente de implantação.
+- [docker-compose.yml](./docker/docker-compose.yml): O sistema depende do [docker-compose.yml](./docker/docker-compose.yml) para iniciar.
+
+> O arquivo [./docker/README](./docker/README.md) fornece uma descrição detalhada das configurações do ambiente e dos serviços, que podem ser usadas como `${ENV_VARS}` no arquivo [service_conf.yaml.template](./docker/service_conf.yaml.template).
+
+Para atualizar a porta HTTP de serviço padrão (80), vá até [docker-compose.yml](./docker/docker-compose.yml) e altere `80:80` para `<SUA_PORTA_DE_SERVIÇO>:80`.
+
+Atualizações nas configurações acima exigem um reinício de todos os contêineres para que tenham efeito:
+
+> ```bash
+> $ docker compose -f docker/docker-compose.yml up -d
+> ```
+
+### Mudar o mecanismo de documentos de Elasticsearch para Infinity
+
+O RAGFlow usa o Elasticsearch por padrão para armazenar texto completo e vetores. Para mudar para o [Infinity](https://github.com/infiniflow/infinity/), siga estas etapas:
+
+1. Pare todos os contêineres em execução:
+
+   ```bash
+   $ docker compose -f docker/docker-compose.yml down -v
+   ```
+
+2. Defina `DOC_ENGINE` no **docker/.env** para `infinity`.
+
+3. Inicie os contêineres:
+
+   ```bash
+   $ docker compose -f docker/docker-compose.yml up -d
+   ```
+
+> [!ATENÇÃO]
+> A mudança para o Infinity em uma máquina Linux/arm64 ainda não é oficialmente suportada.
+
+## 🔧 Criar uma imagem Docker sem modelos de incorporação
+
+Esta imagem tem cerca de 2 GB de tamanho e depende de serviços externos de LLM e incorporação.
+
+```bash
+git clone https://github.com/infiniflow/ragflow.git
+cd ragflow/
+docker build --build-arg LIGHTEN=1 -f Dockerfile -t infiniflow/ragflow:nightly-slim .
+```
+
+## 🔧 Criar uma imagem Docker incluindo modelos de incorporação
+
+Esta imagem tem cerca de 9 GB de tamanho. Como inclui modelos de incorporação, depende apenas de serviços externos de LLM.
+
+```bash
+git clone https://github.com/infiniflow/ragflow.git
+cd ragflow/
+docker build -f Dockerfile -t infiniflow/ragflow:nightly .
+```
+
+## 🔨 Lançar o serviço a partir do código-fonte para desenvolvimento
+
+1. Instale o `uv`, ou pule esta etapa se ele já estiver instalado:
+
+   ```bash
+   pipx install uv
+   ```
+
+2. Clone o código-fonte e instale as dependências Python:
+
+   ```bash
+   git clone https://github.com/infiniflow/ragflow.git
+   cd ragflow/
+   uv sync --python 3.10 --all-extras # instala os módulos Python dependentes do RAGFlow
+   ```
+
+3. Inicie os serviços dependentes (MinIO, Elasticsearch, Redis e MySQL) usando Docker Compose:
+
+   ```bash
+   docker compose -f docker/docker-compose-base.yml up -d
+   ```
+
+   Adicione a seguinte linha ao arquivo `/etc/hosts` para resolver todos os hosts especificados em **docker/.env** para `127.0.0.1`:
+
+   ```
+   127.0.0.1       es01 infinity mysql minio redis
+   ```
+
+4. Se não conseguir acessar o HuggingFace, defina a variável de ambiente `HF_ENDPOINT` para usar um site espelho:
+
+   ```bash
+   export HF_ENDPOINT=https://hf-mirror.com
+   ```
+
+5. Lance o serviço de back-end:
+
+   ```bash
+   source .venv/bin/activate
+   export PYTHONPATH=$(pwd)
+   bash docker/launch_backend_service.sh
+   ```
+
+6. Instale as dependências do front-end:
+
+   ```bash
+   cd web
+   npm install
+   ```
+
+7. Lance o serviço de front-end:
+
+   ```bash
+   npm run dev
+   ```
+
+   _O seguinte resultado confirma o lançamento bem-sucedido do sistema:_
+
+   ![](https://github.com/user-attachments/assets/0daf462c-a24d-4496-a66f-92533534e187)
+
+## 📚 Documentação
+
+- [Início rápido](https://ragflow.io/docs/dev/)
+- [Guia do usuário](https://ragflow.io/docs/dev/category/guides)
+- [Referências](https://ragflow.io/docs/dev/category/references)
+- [FAQ](https://ragflow.io/docs/dev/faq)
+
+## 📜 Roadmap
+
+Veja o [RAGFlow Roadmap 2025](https://github.com/infiniflow/ragflow/issues/4214)
+
+## 🏄 Comunidade
+
+- [Discord](https://discord.gg/4XxujFgUN7)
+- [Twitter](https://twitter.com/infiniflowai)
+- [GitHub Discussions](https://github.com/orgs/infiniflow/discussions)
+
+## 🙌 Contribuindo
+
+O RAGFlow prospera por meio da colaboração de código aberto. Com esse espírito, abraçamos contribuições diversas da comunidade.
+Se você deseja fazer parte, primeiro revise nossas [Diretrizes de Contribuição](./CONTRIBUTING.md).

README_tzh.md

@@ -0,0 +1,353 @@
+<div align="center">
+<a href="https://demo.ragflow.io/">
+<img src="web/src/assets/logo-with-text.png" width="350" alt="ragflow logo">
+</a>
+</div>
+
+<p align="center">
+  <a href="./README.md">English</a> |
+  <a href="./README_zh.md">简体中文</a> |
+  <a href="./README_ja.md">日本語</a> |
+  <a href="./README_ko.md">한국어</a> |
+  <a href="./README_id.md">Bahasa Indonesia</a> |
+  <a href="/README_pt_br.md">Português (Brasil)</a>
+</p>
+
+<p align="center">
+    <a href="https://x.com/intent/follow?screen_name=infiniflowai" target="_blank">
+        <img src="https://img.shields.io/twitter/follow/infiniflow?logo=X&color=%20%23f5f5f5" alt="follow on X(Twitter)">
+    </a>
+    <a href="https://demo.ragflow.io" target="_blank">
+        <img alt="Static Badge" src="https://img.shields.io/badge/Online-Demo-4e6b99">
+    </a>
+    <a href="https://hub.docker.com/r/infiniflow/ragflow" target="_blank">
+        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.16.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.16.0">
+    </a>
+    <a href="https://github.com/infiniflow/ragflow/releases/latest">
+        <img src="https://img.shields.io/github/v/release/infiniflow/ragflow?color=blue&label=Latest%20Release" alt="Latest Release">
+    </a>
+    <a href="https://github.com/infiniflow/ragflow/blob/main/LICENSE">
+        <img height="21" src="https://img.shields.io/badge/License-Apache--2.0-ffffff?labelColor=d4eaf7&color=2e6cc4" alt="license">
+    </a>
+</p>
+
+<h4 align="center">
+  <a href="https://ragflow.io/docs/dev/">Document</a> |
+  <a href="https://github.com/infiniflow/ragflow/issues/4214">Roadmap</a> |
+  <a href="https://twitter.com/infiniflowai">Twitter</a> |
+  <a href="https://discord.gg/4XxujFgUN7">Discord</a> |
+  <a href="https://demo.ragflow.io">Demo</a>
+</h4>
+
+## 💡 RAGFlow 是什麼？
+
+[RAGFlow](https://ragflow.io/) 是一款基於深度文件理解所建構的開源 RAG（Retrieval-Augmented Generation）引擎。 RAGFlow 可以為各種規模的企業及個人提供一套精簡的 RAG 工作流程，結合大語言模型（LLM）針對用戶各類不同的複雜格式數據提供可靠的問答以及有理有據的引用。
+
+## 🎮 Demo 試用
+
+請登入網址 [https://demo.ragflow.io](https://demo.ragflow.io) 試用 demo。
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/7248/2f6baa3e-1092-4f11-866d-36f6a9d075e5" width="1200"/>
+<img src="https://github.com/user-attachments/assets/504bbbf1-c9f7-4d83-8cc5-e9cb63c26db6" width="1200"/>
+</div>
+
+## 🔥 近期更新
+
+- 2025-02-05 更新「SILICONFLOW」的型號清單並新增 Deepseek-R1/DeepSeek-V3 的支援。
+- 2025-01-26 最佳化知識圖譜的擷取與應用，提供了多種配置選擇。
+- 2024-12-18 升級了 Deepdoc 的文檔佈局分析模型。
+- 2024-12-04 支援知識庫的 Pagerank 分數。
+- 2024-11-22 完善了 Agent 中的變數定義和使用。
+- 2024-11-01 對解析後的 chunk 加入關鍵字抽取和相關問題產生以提高回想的準確度。
+- 2024-08-22 支援用 RAG 技術實現從自然語言到 SQL 語句的轉換。
+
+## 🎉 關注項目
+
+⭐️ 點擊右上角的 Star 追蹤 RAGFlow，可以取得最新發布的即時通知 !🌟
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/user-attachments/assets/18c9707e-b8aa-4caf-a154-037089c105ba" width="1200"/>
+</div>
+
+## 🌟 主要功能
+
+### 🍭 **"Quality in, quality out"**
+
+- 基於[深度文件理解](./deepdoc/README.md)，能夠從各類複雜格式的非結構化資料中提取真知灼見。
+- 真正在無限上下文（token）的場景下快速完成大海撈針測試。
+
+### 🍱 **基於模板的文字切片**
+
+- 不只是智能，更重要的是可控可解釋。
+- 多種文字範本可供選擇
+
+### 🌱 **有理有據、最大程度降低幻覺（hallucination）**
+
+- 文字切片過程視覺化，支援手動調整。
+- 有理有據：答案提供關鍵引用的快照並支持追根溯源。
+
+### 🍔 **相容各類異質資料來源**
+
+- 支援豐富的文件類型，包括 Word 文件、PPT、excel 表格、txt 檔案、圖片、PDF、影印件、影印件、結構化資料、網頁等。
+
+### 🛀 **全程無憂、自動化的 RAG 工作流程**
+
+- 全面優化的 RAG 工作流程可以支援從個人應用乃至超大型企業的各類生態系統。
+- 大語言模型 LLM 以及向量模型皆支援配置。
+- 基於多路召回、融合重排序。
+- 提供易用的 API，可輕鬆整合到各類企業系統。
+
+## 🔎 系統架構
+
+<div align="center" style="margin-top:20px;margin-bottom:20px;">
+<img src="https://github.com/infiniflow/ragflow/assets/12318111/d6ac5664-c237-4200-a7c2-a4a00691b485" width="1000"/>
+</div>
+
+## 🎬 快速開始
+
+### 📝 前提條件
+
+- CPU >= 4 核
+- RAM >= 16 GB
+- Disk >= 50 GB
+- Docker >= 24.0.0 & Docker Compose >= v2.26.1
+  > 如果你並沒有在本機安裝 Docker（Windows、Mac，或 Linux）, 可以參考文件 [Install Docker Engine](https://docs.docker.com/engine/install/) 自行安裝。
+
+### 🚀 啟動伺服器
+
+1. 確保 `vm.max_map_count` 不小於 262144：
+
+   > 如需確認 `vm.max_map_count` 的大小：
+   >
+   > ```bash
+   > $ sysctl vm.max_map_count
+   > ```
+   >
+   > 如果 `vm.max_map_count` 的值小於 262144，可以進行重設：
+   >
+   > ```bash
+   > # 這裡我們設為 262144:
+   > $ sudo sysctl -w vm.max_map_count=262144
+   > ```
+   >
+   > 你的改動會在下次系統重新啟動時被重置。如果希望做永久改動，還需要在 **/etc/sysctl.conf** 檔案裡把 `vm.max_map_count` 的值再相應更新一遍：
+   >
+   > ```bash
+   > vm.max_map_count=262144
+   > ```
+
+2. 克隆倉庫：
+
+   ```bash
+   $ git clone https://github.com/infiniflow/ragflow.git
+   ```
+
+3. 進入 **docker** 資料夾，利用事先編譯好的 Docker 映像啟動伺服器：
+
+   > 執行以下指令會自動下載 RAGFlow slim Docker 映像 `v0.16.0-slim`。請參考下表查看不同 Docker 發行版的說明。如需下載不同於 `v0.16.0-slim` 的 Docker 映像，請在執行 `docker compose` 啟動服務之前先更新 **docker/.env** 檔案內的 `RAGFLOW_IMAGE` 變數。例如，你可以透過設定 `RAGFLOW_IMAGE=infiniflow/ragflow:v0.16.0` 來下載 RAGFlow 鏡像的 `v0.16.0` 完整發行版。
+
+   ```bash
+   $ cd ragflow
+   $ docker compose -f docker/docker-compose.yml up -d
+   ```
+
+   | RAGFlow image tag | Image size (GB) | Has embedding models? | Stable?                  |
+   | ----------------- | --------------- | --------------------- | ------------------------ |
+   | v0.16.0           | &approx;9       | :heavy_check_mark:    | Stable release           |
+   | v0.16.0-slim      | &approx;2       | ❌                    | Stable release           |
+   | nightly           | &approx;9       | :heavy_check_mark:    | _Unstable_ nightly build |
+   | nightly-slim      | &approx;2       | ❌                    | _Unstable_ nightly build |
+
+   > [!TIP]
+   > 如果你遇到 Docker 映像檔拉不下來的問題，可以在 **docker/.env** 檔案內根據變數 `RAGFLOW_IMAGE` 的註解提示選擇華為雲或阿里雲的對應映像。
+   >
+   > - 華為雲鏡像名：`swr.cn-north-4.myhuaweicloud.com/infiniflow/ragflow`
+   > - 阿里雲鏡像名：`registry.cn-hangzhou.aliyuncs.com/infiniflow/ragflow`
+
+4. 伺服器啟動成功後再次確認伺服器狀態：
+
+   ```bash
+   $ docker logs -f ragflow-server
+   ```
+
+   _出現以下介面提示說明伺服器啟動成功：_
+
+   ```bash
+        ____   ___    ______ ______ __
+       / __ \ /   |  / ____// ____// /____  _      __
+      / /_/ // /| | / / __ / /_   / // __ \| | /| / /
+     / _, _// ___ |/ /_/ // __/  / // /_/ /| |/ |/ /
+    /_/ |_|/_/  |_|\____//_/    /_/ \____/ |__/|__/
+
+    * Running on all addresses (0.0.0.0)
+    * Running on http://127.0.0.1:9380
+    * Running on http://x.x.x.x:9380
+    INFO:werkzeug:Press CTRL+C to quit
+   ```
+
+   > 如果您跳過這一步驟系統確認步驟就登入 RAGFlow，你的瀏覽器有可能會提示 `network anormal` 或 `網路異常`，因為 RAGFlow 可能並未完全啟動成功。
+
+5. 在你的瀏覽器中輸入你的伺服器對應的 IP 位址並登入 RAGFlow。
+   > 上面這個範例中，您只需輸入 http://IP_OF_YOUR_MACHINE 即可：未改動過設定則無需輸入連接埠（預設的 HTTP 服務連接埠 80）。
+6. 在 [service_conf.yaml.template](./docker/service_conf.yaml.template) 檔案的 `user_default_llm` 欄位設定 LLM factory，並在 `API_KEY` 欄填入和你選擇的大模型相對應的 API key。
+
+   > 詳見 [llm_api_key_setup](https://ragflow.io/docs/dev/llm_api_key_setup)。
+
+   _好戲開始，接著奏樂接著舞！ _
+
+## 🔧 系統配置
+
+系統配置涉及以下三份文件：
+
+- [.env](./docker/.env)：存放一些基本的系統環境變量，例如 `SVR_HTTP_PORT`、`MYSQL_PASSWORD`、`MINIO_PASSWORD` 等。
+- [service_conf.yaml.template](./docker/service_conf.yaml.template)：設定各類別後台服務。
+- [docker-compose.yml](./docker/docker-compose.yml): 系統依賴該檔案完成啟動。
+
+請務必確保 [.env](./docker/.env) 檔案中的變數設定與 [service_conf.yaml.template](./docker/service_conf.yaml.template) 檔案中的設定保持一致！
+
+如果無法存取映像網站 hub.docker.com 或模型網站 huggingface.co，請依照 [.env](./docker/.env) 註解修改 `RAGFLOW_IMAGE` 和 `HF_ENDPOINT`。
+
+> [./docker/README](./docker/README.md) 解釋了 [service_conf.yaml.template](./docker/service_conf.yaml.template) 用到的環境變數設定和服務配置。
+
+如需更新預設的 HTTP 服務連接埠(80), 可以在[docker-compose.yml](./docker/docker-compose.yml) 檔案中將配置`80:80` 改為`<YOUR_SERVING_PORT>:80` 。
+
+> 所有系統配置都需要透過系統重新啟動生效：
+>
+> ```bash
+> $ docker compose -f docker/docker-compose.yml up -d
+> ```
+
+###把文檔引擎從 Elasticsearch 切換成為 Infinity
+
+RAGFlow 預設使用 Elasticsearch 儲存文字和向量資料. 如果要切換為 [Infinity](https://github.com/infiniflow/infinity/), 可以按照下面步驟進行:
+
+1. 停止所有容器運作:
+
+   ```bash
+   $ docker compose -f docker/docker-compose.yml down -v
+   ```
+
+2. 設定 **docker/.env** 目錄中的 `DOC_ENGINE` 為 `infinity`.
+
+3. 啟動容器:
+
+   ```bash
+   $ docker compose -f docker/docker-compose.yml up -d
+   ```
+
+> [!WARNING]
+> Infinity 目前官方並未正式支援在 Linux/arm64 架構下的機器上運行.
+
+## 🔧 原始碼編譯 Docker 映像（不含 embedding 模型）
+
+本 Docker 映像大小約 2 GB 左右並且依賴外部的大模型和 embedding 服務。
+
+```bash
+git clone https://github.com/infiniflow/ragflow.git
+cd ragflow/
+docker build --build-arg LIGHTEN=1 --build-arg NEED_MIRROR=1 -f Dockerfile -t infiniflow/ragflow:nightly-slim .
+```
+
+## 🔧 原始碼編譯 Docker 映像（包含 embedding 模型）
+
+本 Docker 大小約 9 GB 左右。由於已包含 embedding 模型，所以只需依賴外部的大模型服務即可。
+
+```bash
+git clone https://github.com/infiniflow/ragflow.git
+cd ragflow/
+docker build --build-arg NEED_MIRROR=1 -f Dockerfile -t infiniflow/ragflow:nightly .
+```
+
+## 🔨 以原始碼啟動服務
+
+1. 安裝 uv。如已安裝，可跳過此步驟：
+
+   ```bash
+   pipx install uv
+   export UV_INDEX=https://pypi.tuna.tsinghua.edu.cn/simple
+   ```
+
+2. 下載原始碼並安裝 Python 依賴：
+
+   ```bash
+   git clone https://github.com/infiniflow/ragflow.git
+   cd ragflow/
+   uv sync --python 3.10 --all-extras # install RAGFlow dependent python modules
+   ```
+
+3. 透過 Docker Compose 啟動依賴的服務（MinIO, Elasticsearch, Redis, and MySQL）：
+
+   ```bash
+   docker compose -f docker/docker-compose-base.yml up -d
+   ```
+
+   在 `/etc/hosts` 中加入以下程式碼，將 **conf/service_conf.yaml** 檔案中的所有 host 位址都解析為 `127.0.0.1`：
+
+   ```
+   127.0.0.1 es01 infinity mysql minio redis
+   ```
+
+4. 如果無法存取 HuggingFace，可以把環境變數 `HF_ENDPOINT` 設為對應的鏡像網站：
+
+   ```bash
+   export HF_ENDPOINT=https://hf-mirror.com
+   ```
+
+5.啟動後端服務：
+『`bash
+source .venv/bin/activate
+export PYTHONPATH=$(pwd)
+bash docker/launch_backend_service.sh
+
+```
+
+6. 安裝前端依賴：
+『`bash
+cd web
+npm install
+```
+
+7. 啟動前端服務：
+   『`bash
+   npm run dev
+
+   ```
+
+   以下界面說明系統已成功啟動：_
+
+   ![](https://github.com/user-attachments/assets/0daf462c-a24d-4496-a66f-92533534e187)
+   ```
+
+## 📚 技術文檔
+
+- [Quickstart](https://ragflow.io/docs/dev/)
+- [User guide](https://ragflow.io/docs/dev/category/guides)
+- [References](https://ragflow.io/docs/dev/category/references)
+- [FAQ](https://ragflow.io/docs/dev/faq)
+
+## 📜 路線圖
+
+詳見 [RAGFlow Roadmap 2025](https://github.com/infiniflow/ragflow/issues/4214) 。
+
+## 🏄 開源社群
+
+- [Discord](https://discord.gg/4XxujFgUN7)
+- [Twitter](https://twitter.com/infiniflowai)
+- [GitHub Discussions](https://github.com/orgs/infiniflow/discussions)
+
+## 🙌 貢獻指南
+
+RAGFlow 只有透過開源協作才能蓬勃發展。秉持這項精神,我們歡迎來自社區的各種貢獻。如果您有意參與其中,請查閱我們的 [貢獻者指南](./CONTRIBUTING.md) 。
+
+## 🤝 商務合作
+
+- [預約諮詢](https://aao615odquw.feishu.cn/share/base/form/shrcnjw7QleretCLqh1nuPo1xxh)
+
+## 👥 加入社區
+
+掃二維碼加入 RAGFlow 小助手，進 RAGFlow 交流群。
+
+<p align="center">
+  <img src="https://github.com/infiniflow/ragflow/assets/7248/bccf284f-46f2-4445-9809-8f1030fb7585" width=50% height=50%>
+</p>

README_zh.md

@@ -7,8 +7,11 @@
 <p align="center">
   <a href="./README.md">English</a> |
   <a href="./README_zh.md">简体中文</a> |
+  <a href="./README_tzh.md">繁体中文</a> |
   <a href="./README_ja.md">日本語</a> |
-  <a href="./README_ko.md">한국어</a>
+  <a href="./README_ko.md">한국어</a> |
+  <a href="./README_id.md">Bahasa Indonesia</a> |
+  <a href="/README_pt_br.md">Português (Brasil)</a>
 </p>
 
 <p align="center">
@@ -19,7 +22,7 @@
         <img alt="Static Badge" src="https://img.shields.io/badge/Online-Demo-4e6b99">
     </a>
     <a href="https://hub.docker.com/r/infiniflow/ragflow" target="_blank">
-        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.13.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.13.0">
+        <img src="https://img.shields.io/badge/docker_pull-ragflow:v0.16.0-brightgreen" alt="docker pull infiniflow/ragflow:v0.16.0">
     </a>
     <a href="https://github.com/infiniflow/ragflow/releases/latest">
         <img src="https://img.shields.io/github/v/release/infiniflow/ragflow?color=blue&label=Latest%20Release" alt="Latest Release">
@@ -29,10 +32,9 @@
     </a>
 </p>
 
-
 <h4 align="center">
   <a href="https://ragflow.io/docs/dev/">Document</a> |
-  <a href="https://github.com/infiniflow/ragflow/issues/162">Roadmap</a> |
+  <a href="https://github.com/infiniflow/ragflow/issues/4214">Roadmap</a> |
   <a href="https://twitter.com/infiniflowai">Twitter</a> |
   <a href="https://discord.gg/4XxujFgUN7">Discord</a> |
   <a href="https://demo.ragflow.io">Demo</a>
@@ -45,27 +47,30 @@
 ## 🎮 Demo 试用
 
 请登录网址 [https://demo.ragflow.io](https://demo.ragflow.io) 试用 demo。
+
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
 <img src="https://github.com/infiniflow/ragflow/assets/7248/2f6baa3e-1092-4f11-866d-36f6a9d075e5" width="1200"/>
-<img src="https://github.com/infiniflow/ragflow/assets/12318111/b083d173-dadc-4ea9-bdeb-180d7df514eb" width="1200"/>
+<img src="https://github.com/user-attachments/assets/504bbbf1-c9f7-4d83-8cc5-e9cb63c26db6" width="1200"/>
 </div>
 
-
 ## 🔥 近期更新
 
-- 2024-09-29 优化多轮对话.
-- 2024-09-13 增加知识库问答搜索模式。
-- 2024-09-09 在 Agent 中加入医疗问诊模板。
+- 2025-02-05 更新硅基流动的模型列表，增加了对 Deepseek-R1/DeepSeek-V3 的支持。
+- 2025-01-26 优化知识图谱的提取和应用，提供了多种配置选择。
+- 2024-12-18 升级了 Deepdoc 的文档布局分析模型。
+- 2024-12-04 支持知识库的 Pagerank 分数。
+- 2024-11-22 完善了 Agent 中的变量定义和使用。
+- 2024-11-01 对解析后的 chunk 加入关键词抽取和相关问题生成以提高召回的准确度。
 - 2024-08-22 支持用 RAG 技术实现从自然语言到 SQL 语句的转换。
-- 2024-08-02 支持 GraphRAG 启发于 [graphrag](https://github.com/microsoft/graphrag) 和思维导图。
 
 ## 🎉 关注项目
+
 ⭐️ 点击右上角的 Star 关注 RAGFlow，可以获取最新发布的实时通知 !🌟
+
 <div align="center" style="margin-top:20px;margin-bottom:20px;">
 <img src="https://github.com/user-attachments/assets/18c9707e-b8aa-4caf-a154-037089c105ba" width="1200"/>
 </div>
 
-
 ## 🌟 主要功能
 
 ### 🍭 **"Quality in, quality out"**
@@ -141,17 +146,25 @@
 
 3. 进入 **docker** 文件夹，利用提前编译好的 Docker 镜像启动服务器：
 
-   > 运行以下命令会自动下载 dev 版的 RAGFlow slim Docker 镜像（`dev-slim`），该镜像并不包含 embedding 模型以及一些 Python 库，因此镜像大小约 1GB。
+   > 运行以下命令会自动下载 RAGFlow slim Docker 镜像 `v0.16.0-slim`。请参考下表查看不同 Docker 发行版的描述。如需下载不同于 `v0.16.0-slim` 的 Docker 镜像，请在运行 `docker compose` 启动服务之前先更新 **docker/.env** 文件内的 `RAGFLOW_IMAGE` 变量。比如，你可以通过设置 `RAGFLOW_IMAGE=infiniflow/ragflow:v0.16.0` 来下载 RAGFlow 镜像的 `v0.16.0` 完整发行版。
 
    ```bash
-   $ cd ragflow/docker
-   $ docker compose -f docker-compose.yml up -d
+   $ cd ragflow
+   $ docker compose -f docker/docker-compose.yml up -d
    ```
 
-   > - 如果你想下载并运行特定版本的 RAGFlow slim Docker 镜像，请在 **docker/.env** 文件中找到 `RAGFLOW_IMAGE` 变量，将其改为对应版本。例如 `RAGFLOW_IMAGE=infiniflow/ragflow:v0.13.0-slim`，然后再运行上述命令。
-   > - 如果您想安装内置 embedding 模型和 Python 库的 dev 版本的 Docker 镜像，需要将 **docker/.env** 文件中的 `RAGFLOW_IMAGE` 变量修改为： `RAGFLOW_IMAGE=infiniflow/ragflow:dev`。
-   > - 如果您想安装内置 embedding 模型和 Python 库的指定版本的 RAGFlow Docker 镜像，需要将 **docker/.env** 文件中的 `RAGFLOW_IMAGE` 变量修改为： `RAGFLOW_IMAGE=infiniflow/ragflow:v0.13.0`。修改后，再运行上面的命令。
-   > **注意：** 安装内置 embedding 模型和 Python 库的指定版本的 RAGFlow Docker 镜像大小约 9 GB，可能需要更长时间下载，请耐心等待。
+   | RAGFlow image tag | Image size (GB) | Has embedding models? | Stable?                  |
+   | ----------------- | --------------- | --------------------- | ------------------------ |
+   | v0.16.0           | &approx;9       | :heavy_check_mark:    | Stable release           |
+   | v0.16.0-slim      | &approx;2       | ❌                    | Stable release           |
+   | nightly           | &approx;9       | :heavy_check_mark:    | _Unstable_ nightly build |
+   | nightly-slim      | &approx;2       | ❌                    | _Unstable_ nightly build |
+
+   > [!TIP]
+   > 如果你遇到 Docker 镜像拉不下来的问题，可以在 **docker/.env** 文件内根据变量 `RAGFLOW_IMAGE` 的注释提示选择华为云或者阿里云的相应镜像。
+   >
+   > - 华为云镜像名：`swr.cn-north-4.myhuaweicloud.com/infiniflow/ragflow`
+   > - 阿里云镜像名：`registry.cn-hangzhou.aliyuncs.com/infiniflow/ragflow`
 
 4. 服务器启动成功后再次确认服务器状态：
 
@@ -173,11 +186,12 @@
     * Running on http://x.x.x.x:9380
     INFO:werkzeug:Press CTRL+C to quit
    ```
-   > 如果您跳过这一步系统确认步骤就登录 RAGFlow，你的浏览器有可能会提示 `network abnormal` 或 `网络异常`，因为 RAGFlow 可能并未完全启动成功。
+
+   > 如果您跳过这一步系统确认步骤就登录 RAGFlow，你的浏览器有可能会提示 `network anormal` 或 `网络异常`，因为 RAGFlow 可能并未完全启动成功。
 
 5. 在你的浏览器中输入你的服务器对应的 IP 地址并登录 RAGFlow。
    > 上面这个例子中，您只需输入 http://IP_OF_YOUR_MACHINE 即可：未改动过配置则无需输入端口（默认的 HTTP 服务端口 80）。
-6. 在 [service_conf.yaml](./docker/service_conf.yaml) 文件的 `user_default_llm` 栏配置 LLM factory，并在 `API_KEY` 栏填写和你选择的大模型相对应的 API key。
+6. 在 [service_conf.yaml.template](./docker/service_conf.yaml.template) 文件的 `user_default_llm` 栏配置 LLM factory，并在 `API_KEY` 栏填写和你选择的大模型相对应的 API key。
 
    > 详见 [llm_api_key_setup](https://ragflow.io/docs/dev/llm_api_key_setup)。
 
@@ -188,33 +202,52 @@
 系统配置涉及以下三份文件：
 
 - [.env](./docker/.env)：存放一些基本的系统环境变量，比如 `SVR_HTTP_PORT`、`MYSQL_PASSWORD`、`MINIO_PASSWORD` 等。
-- [service_conf.yaml](./docker/service_conf.yaml)：配置各类后台服务。
+- [service_conf.yaml.template](./docker/service_conf.yaml.template)：配置各类后台服务。
 - [docker-compose.yml](./docker/docker-compose.yml): 系统依赖该文件完成启动。
 
-请务必确保 [.env](./docker/.env) 文件中的变量设置与 [service_conf.yaml](./docker/service_conf.yaml) 文件中的配置保持一致！
+请务必确保 [.env](./docker/.env) 文件中的变量设置与 [service_conf.yaml.template](./docker/service_conf.yaml.template) 文件中的配置保持一致！
 
 如果不能访问镜像站点 hub.docker.com 或者模型站点 huggingface.co，请按照 [.env](./docker/.env) 注释修改 `RAGFLOW_IMAGE` 和 `HF_ENDPOINT`。
 
-> [./docker/README](./docker/README.md) 文件提供了环境变量设置和服务配置的详细信息。请**一定要**确保 [./docker/README](./docker/README.md) 文件当中列出来的环境变量的值与 [service_conf.yaml](./docker/service_conf.yaml) 文件当中的系统配置保持一致。
+> [./docker/README](./docker/README.md) 解释了 [service_conf.yaml.template](./docker/service_conf.yaml.template) 用到的环境变量设置和服务配置。
 
 如需更新默认的 HTTP 服务端口(80), 可以在 [docker-compose.yml](./docker/docker-compose.yml) 文件中将配置 `80:80` 改为 `<YOUR_SERVING_PORT>:80`。
 
 > 所有系统配置都需要通过系统重启生效：
 >
 > ```bash
-> $ docker compose -f docker-compose.yml up -d
+> $ docker compose -f docker/docker-compose.yml up -d
 > ```
 
+### 把文档引擎从 Elasticsearch 切换成为 Infinity
+
+RAGFlow 默认使用 Elasticsearch 存储文本和向量数据. 如果要切换为 [Infinity](https://github.com/infiniflow/infinity/), 可以按照下面步骤进行:
+
+1. 停止所有容器运行:
+
+   ```bash
+   $ docker compose -f docker/docker-compose.yml down -v
+   ```
+
+2. 设置 **docker/.env** 目录中的 `DOC_ENGINE` 为 `infinity`.
+
+3. 启动容器:
+
+   ```bash
+   $ docker compose -f docker/docker-compose.yml up -d
+   ```
+
+> [!WARNING]
+> Infinity 目前官方并未正式支持在 Linux/arm64 架构下的机器上运行.
+
 ## 🔧 源码编译 Docker 镜像（不含 embedding 模型）
 
-本 Docker 镜像大小约 1 GB 左右并且依赖外部的大模型和 embedding 服务。
+本 Docker 镜像大小约 2 GB 左右并且依赖外部的大模型和 embedding 服务。
 
 ```bash
 git clone https://github.com/infiniflow/ragflow.git
 cd ragflow/
-pip3 install huggingface-hub nltk
-python3 download_deps.py
-docker build -f Dockerfile.slim -t infiniflow/ragflow:dev-slim .
+docker build --build-arg LIGHTEN=1 --build-arg NEED_MIRROR=1 -f Dockerfile -t infiniflow/ragflow:nightly-slim .
 ```
 
 ## 🔧 源码编译 Docker 镜像（包含 embedding 模型）
@@ -224,36 +257,37 @@ docker build -f Dockerfile.slim -t infiniflow/ragflow:dev-slim .
 ```bash
 git clone https://github.com/infiniflow/ragflow.git
 cd ragflow/
-pip3 install huggingface-hub nltk
-python3 download_deps.py
-docker build -f Dockerfile -t infiniflow/ragflow:dev .
+docker build --build-arg NEED_MIRROR=1 -f Dockerfile -t infiniflow/ragflow:nightly .
 ```
 
 ## 🔨 以源代码启动服务
 
-1. 安装 Poetry。如已经安装，可跳过本步骤：  
+1. 安装 uv。如已经安装，可跳过本步骤：
+
    ```bash
-   curl -sSL https://install.python-poetry.org | python3 -
+   pipx install uv
+   export UV_INDEX=https://pypi.tuna.tsinghua.edu.cn/simple
    ```
 
 2. 下载源代码并安装 Python 依赖：
+
    ```bash
    git clone https://github.com/infiniflow/ragflow.git
    cd ragflow/
-   export POETRY_VIRTUALENVS_CREATE=true POETRY_VIRTUALENVS_IN_PROJECT=true
-   ~/.local/bin/poetry install --sync --no-root # install RAGFlow dependent python modules
+   uv sync --python 3.10 --all-extras # install RAGFlow dependent python modules
    ```
 
 3. 通过 Docker Compose 启动依赖的服务（MinIO, Elasticsearch, Redis, and MySQL）：
+
    ```bash
    docker compose -f docker/docker-compose-base.yml up -d
    ```
 
-   在 `/etc/hosts` 中添加以下代码，将 **docker/service_conf.yaml** 文件中的所有 host 地址都解析为 `127.0.0.1`：  
+   在 `/etc/hosts` 中添加以下代码，将 **conf/service_conf.yaml** 文件中的所有 host 地址都解析为 `127.0.0.1`：
+
    ```
-   127.0.0.1       es01 mysql minio redis
+   127.0.0.1       es01 infinity mysql minio redis
    ```
-   在文件 **docker/service_conf.yaml** 中，对照 **docker/.env** 的配置将 mysql 端口更新为 `5455`，es 端口更新为 `1200`。
 
 4. 如果无法访问 HuggingFace，可以把环境变量 `HF_ENDPOINT` 设成相应的镜像站点：
 
@@ -262,6 +296,7 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
    ```
 
 5. 启动后端服务：
+
    ```bash
    source .venv/bin/activate
    export PYTHONPATH=$(pwd)
@@ -271,10 +306,10 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
 6. 安装前端依赖：
    ```bash
    cd web
-   npm install --force
+   npm install
    ```
-7. 配置前端，将 **.umirc.ts** 的 `proxy.target` 更新为 `http://127.0.0.1:9380`：  
-8. 启动前端服务：  
+7. 启动前端服务：
+
    ```bash
    npm run dev
    ```
@@ -292,7 +327,7 @@ docker build -f Dockerfile -t infiniflow/ragflow:dev .
 
 ## 📜 路线图
 
-详见 [RAGFlow Roadmap 2024](https://github.com/infiniflow/ragflow/issues/162) 。
+详见 [RAGFlow Roadmap 2025](https://github.com/infiniflow/ragflow/issues/4214) 。
 
 ## 🏄 开源社区
 
@@ -315,4 +350,3 @@ RAGFlow 只有通过开源协作才能蓬勃发展。秉持这一精神,我们
 <p align="center">
   <img src="https://github.com/infiniflow/ragflow/assets/7248/bccf284f-46f2-4445-9809-8f1030fb7585" width=50% height=50%>
 </p>
-

agent/README.md

@@ -10,7 +10,7 @@ It is used to compose a complex work flow or agent.
 And this graph is beyond the DAG that we can use circles to describe our agent or work flow.
 Under this folder, we propose a test tool ./test/client.py which can test the DSLs such as json files in folder ./test/dsl_examples.
 Please use this client at the same folder you start RAGFlow. If it's run by Docker, please go into the container before running the client.
-Otherwise, correct configurations in conf/service_conf.yaml is essential.
+Otherwise, correct configurations in service_conf.yaml is essential.
 
 ```bash
 PYTHONPATH=path/to/ragflow python graph/test/client.py -h

agent/README_zh.md

@@ -11,7 +11,7 @@
 在这个文件夹下，我们提出了一个测试工具 ./test/client.py，
 它可以测试像文件夹./test/dsl_examples下一样的DSL文件。
 请在启动 RAGFlow 的同一文件夹中使用此客户端。如果它是通过 Docker 运行的，请在运行客户端之前进入容器。
-否则，正确配置 conf/service_conf.yaml 文件是必不可少的。
+否则，正确配置 service_conf.yaml 文件是必不可少的。
 
 ```bash
 PYTHONPATH=path/to/ragflow python graph/test/client.py -h

agent/__init__.py

@@ -0,0 +1,18 @@
+#
+#  Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+
+from beartype.claw import beartype_this_package
+beartype_this_package()

agent/canvas.py

@@ -13,9 +13,8 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-import importlib
+import logging
 import json
-import traceback
 from abc import ABC
 from copy import deepcopy
 from functools import partial
@@ -24,7 +23,6 @@ import pandas as pd
 
 from agent.component import component_class
 from agent.component.base import ComponentBase
-from agent.settings import flow_logger, DEBUG
 
 
 class Canvas(ABC):
@@ -88,7 +86,8 @@ class Canvas(ABC):
                         }
                     },
                     "downstream": [],
-                    "upstream": []
+                    "upstream": [],
+                    "parent_id": ""
                 }
             },
             "history": [],
@@ -139,7 +138,8 @@ class Canvas(ABC):
             "components": {}
         }
         for k in self.dsl.keys():
-            if k in ["components"]:continue
+            if k in ["components"]:
+                continue
             dsl[k] = deepcopy(self.dsl[k])
 
         for k, cpn in self.components.items():
@@ -162,8 +162,13 @@ class Canvas(ABC):
             self.components[k]["obj"].reset()
         self._embed_id = ""
 
+    def get_compnent_name(self, cid):
+        for n in self.dsl["graph"]["nodes"]:
+            if cid == n["id"]:
+                return n["data"]["name"]
+        return ""
+
     def run(self, **kwargs):
-        ans = ""
         if self.answer:
             cpn_id = self.answer[0]
             self.answer.pop(0)
@@ -173,71 +178,107 @@ class Canvas(ABC):
                 ans = ComponentBase.be_output(str(e))
             self.path[-1].append(cpn_id)
             if kwargs.get("stream"):
-                assert isinstance(ans, partial)
-                return ans
-            self.history.append(("assistant", ans.to_dict("records")))
-            return ans
+                for an in ans():
+                    yield an
+            else:
+                yield ans
+            return
 
         if not self.path:
             self.components["begin"]["obj"].run(self.history, **kwargs)
             self.path.append(["begin"])
 
         self.path.append([])
+
         ran = -1
+        waiting = []
+        without_dependent_checking = []
 
         def prepare2run(cpns):
             nonlocal ran, ans
             for c in cpns:
-                if self.path[-1] and c == self.path[-1][-1]: continue
+                if self.path[-1] and c == self.path[-1][-1]:
+                    continue
                 cpn = self.components[c]["obj"]
                 if cpn.component_name == "Answer":
                     self.answer.append(c)
                 else:
-                    if DEBUG: print("RUN: ", c)
-                    if cpn.component_name == "Generate":
+                    logging.debug(f"Canvas.prepare2run: {c}")
+                    if c not in without_dependent_checking:
                         cpids = cpn.get_dependent_components()
-                        if any([c not in self.path[-1] for c in cpids]):
+                        if any([cc not in self.path[-1] for cc in cpids]):
+                            if c not in waiting:
+                                waiting.append(c)
                             continue
+                    yield "*'{}'* is running...🕞".format(self.get_compnent_name(c))
+
+                    if cpn.component_name.lower() == "iteration":
+                        st_cpn = cpn.get_start()
+                        assert st_cpn, "Start component not found for Iteration."
+                        if not st_cpn["obj"].end():
+                            cpn = st_cpn["obj"]
+                            c = cpn._id
+
+                    try:
                         ans = cpn.run(self.history, **kwargs)
+                    except Exception as e:
+                        logging.exception(f"Canvas.run got exception: {e}")
                         self.path[-1].append(c)
                         ran += 1
+                        raise e
+                    self.path[-1].append(c)
+
+            ran += 1
+
+        downstream = self.components[self.path[-2][-1]]["downstream"]
+        if not downstream and self.components[self.path[-2][-1]].get("parent_id"):
+            cid = self.path[-2][-1]
+            pid = self.components[cid]["parent_id"]
+            o, _ = self.components[cid]["obj"].output(allow_partial=False)
+            oo, _ = self.components[pid]["obj"].output(allow_partial=False)
+            self.components[pid]["obj"].set(pd.concat([oo, o], ignore_index=True))
+            downstream = [pid]
+
+        for m in prepare2run(downstream):
+            yield {"content": m, "running_status": True}
 
-        prepare2run(self.components[self.path[-2][-1]]["downstream"])
         while 0 <= ran < len(self.path[-1]):
-            if DEBUG: print(ran, self.path)
+            logging.debug(f"Canvas.run: {ran} {self.path}")
             cpn_id = self.path[-1][ran]
             cpn = self.get_component(cpn_id)
-            if not cpn["downstream"]: break
+            if not any([cpn["downstream"], cpn.get("parent_id"), waiting]):
+                break
 
             loop = self._find_loop()
-            if loop: raise OverflowError(f"Too much loops: {loop}")
+            if loop:
+                raise OverflowError(f"Too much loops: {loop}")
 
             if cpn["obj"].component_name.lower() in ["switch", "categorize", "relevant"]:
                 switch_out = cpn["obj"].output()[1].iloc[0, 0]
                 assert switch_out in self.components, \
                     "{}'s output: {} not valid.".format(cpn_id, switch_out)
-                try:
-                    prepare2run([switch_out])
-                except Exception as e:
-                    for p in [c for p in self.path for c in p][::-1]:
-                        if p.lower().find("answer") >= 0:
-                            self.get_component(p)["obj"].set_exception(e)
-                            prepare2run([p])
-                            break
-                    traceback.print_exc()
-                    break
+                for m in prepare2run([switch_out]):
+                    yield {"content": m, "running_status": True}
                 continue
 
-            try:
-                prepare2run(cpn["downstream"])
-            except Exception as e:
-                for p in [c for p in self.path for c in p][::-1]:
-                    if p.lower().find("answer") >= 0:
-                        self.get_component(p)["obj"].set_exception(e)
-                        prepare2run([p])
-                        break
-                traceback.print_exc()
-                break
+            downstream = cpn["downstream"]
+            if not downstream and cpn.get("parent_id"):
+                pid = cpn["parent_id"]
+                _, o = cpn["obj"].output(allow_partial=False)
+                _, oo = self.components[pid]["obj"].output(allow_partial=False)
+                self.components[pid]["obj"].set_output(pd.concat([oo.dropna(axis=1), o.dropna(axis=1)], ignore_index=True))
+                downstream = [pid]
+
+            for m in prepare2run(downstream):
+                yield {"content": m, "running_status": True}
+
+            if ran >= len(self.path[-1]) and waiting:
+                without_dependent_checking = waiting
+                waiting = []
+                for m in prepare2run(without_dependent_checking):
+                    yield {"content": m, "running_status": True}
+                without_dependent_checking = []
+                ran -= 1
 
         if self.answer:
             cpn_id = self.answer[0]
@@ -246,11 +287,13 @@ class Canvas(ABC):
             self.path[-1].append(cpn_id)
             if kwargs.get("stream"):
                 assert isinstance(ans, partial)
-                return ans
+                for an in ans():
+                    yield an
+            else:
+                yield ans
 
-            self.history.append(("assistant", ans.to_dict("records")))
-
-        return ans
+        else:
+            raise Exception("The dialog flow has no way to interact with you. Please add an 'Interact' component to the end of the flow.")
 
     def get_component(self, cpn_id):
         return self.components[cpn_id]
@@ -261,8 +304,10 @@ class Canvas(ABC):
     def get_history(self, window_size):
         convs = []
         for role, obj in self.history[window_size * -1:]:
-            convs.append({"role": role, "content": (obj if role == "user" else
-                    '\n'.join([str(s) for s in pd.DataFrame(obj)['content']]))})
+            if isinstance(obj, list) and obj and all([isinstance(o, dict) for o in obj]):
+                convs.append({"role": role, "content": '\n'.join([str(s.get("content", "")) for s in obj])})
+            else:
+                convs.append({"role": role, "content": str(obj)})
         return convs
 
     def add_user_input(self, question):
@@ -276,19 +321,22 @@ class Canvas(ABC):
 
     def _find_loop(self, max_loops=6):
         path = self.path[-1][::-1]
-        if len(path) < 2: return False
+        if len(path) < 2:
+            return False
 
         for i in range(len(path)):
-            if path[i].lower().find("answer") >= 0:
+            if path[i].lower().find("answer") == 0 or path[i].lower().find("iterationitem") == 0:
                 path = path[:i]
                 break
 
-        if len(path) < 2: return False
+        if len(path) < 2:
+            return False
 
-        for l in range(2, len(path) // 2):
-            pat = ",".join(path[0:l])
+        for loc in range(2, len(path) // 2):
+            pat = ",".join(path[0:loc])
             path_str = ",".join(path)
-            if len(pat) >= len(path_str): return False
+            if len(pat) >= len(path_str):
+                return False
             loop = max_loops
             while path_str.find(pat) == 0 and loop >= 0:
                 loop -= 1
@@ -296,10 +344,23 @@ class Canvas(ABC):
                     return False
                 path_str = path_str[len(pat)+1:]
             if loop < 0:
-                pat = " => ".join([p.split(":")[0] for p in path[0:l]])
+                pat = " => ".join([p.split(":")[0] for p in path[0:loc]])
                 return pat + " => " + pat
 
         return False
 
     def get_prologue(self):
         return self.components["begin"]["obj"]._param.prologue
+
+    def set_global_param(self, **kwargs):
+        for k, v in kwargs.items():
+            for q in self.components["begin"]["obj"]._param.query:
+                if k != q["key"]:
+                    continue
+                q["value"] = v
+
+    def get_preset_param(self):
+        return self.components["begin"]["obj"]._param.query
+
+    def get_component_input_elements(self, cpnnm):
+        return self.components[cpnnm]["obj"].get_input_elements()

agent/component/__init__.py

@@ -1,3 +1,19 @@
+#
+#  Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+
 import importlib
 from .begin import Begin, BeginParam
 from .generate import Generate, GenerateParam
@@ -30,9 +46,88 @@ from .tushare import TuShare, TuShareParam
 from .akshare import AkShare, AkShareParam
 from .crawler import Crawler, CrawlerParam
 from .invoke import Invoke, InvokeParam
+from .template import Template, TemplateParam
+from .email import Email, EmailParam
+from .iteration import Iteration, IterationParam
+from .iterationitem import IterationItem, IterationItemParam
 
 
 def component_class(class_name):
     m = importlib.import_module("agent.component")
     c = getattr(m, class_name)
     return c
+
+
+__all__ = [
+    "Begin",
+    "BeginParam",
+    "Generate",
+    "GenerateParam",
+    "Retrieval",
+    "RetrievalParam",
+    "Answer",
+    "AnswerParam",
+    "Categorize",
+    "CategorizeParam",
+    "Switch",
+    "SwitchParam",
+    "Relevant",
+    "RelevantParam",
+    "Message",
+    "MessageParam",
+    "RewriteQuestion",
+    "RewriteQuestionParam",
+    "KeywordExtract",
+    "KeywordExtractParam",
+    "Concentrator",
+    "ConcentratorParam",
+    "Baidu",
+    "BaiduParam",
+    "DuckDuckGo",
+    "DuckDuckGoParam",
+    "Wikipedia",
+    "WikipediaParam",
+    "PubMed",
+    "PubMedParam",
+    "ArXiv",
+    "ArXivParam",
+    "Google",
+    "GoogleParam",
+    "Bing",
+    "BingParam",
+    "GoogleScholar",
+    "GoogleScholarParam",
+    "DeepL",
+    "DeepLParam",
+    "GitHub",
+    "GitHubParam",
+    "BaiduFanyi",
+    "BaiduFanyiParam",
+    "QWeather",
+    "QWeatherParam",
+    "ExeSQL",
+    "ExeSQLParam",
+    "YahooFinance",
+    "YahooFinanceParam",
+    "WenCai",
+    "WenCaiParam",
+    "Jin10",
+    "Jin10Param",
+    "TuShare",
+    "TuShareParam",
+    "AkShare",
+    "AkShareParam",
+    "Crawler",
+    "CrawlerParam",
+    "Invoke",
+    "InvokeParam",
+    "Iteration",
+    "IterationParam",
+    "IterationItem",
+    "IterationItemParam",
+    "Template",
+    "TemplateParam",
+    "Email",
+    "EmailParam",
+    "component_class"
+]

agent/component/akshare.py

@@ -15,7 +15,6 @@
 #
 from abc import ABC
 import pandas as pd
-import akshare as ak
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -36,6 +35,7 @@ class AkShare(ComponentBase, ABC):
     component_name = "AkShare"
 
     def _run(self, history, **kwargs):
+        import akshare as ak
         ans = self.get_input()
         ans = ",".join(ans["content"]) if "content" in ans else ""
         if not ans:

agent/component/answer.py

@@ -16,6 +16,7 @@
 import random
 from abc import ABC
 from functools import partial
+from typing import Tuple, Union
 
 import pandas as pd
 
@@ -76,4 +77,13 @@ class Answer(ComponentBase, ABC):
     def set_exception(self, e):
         self.exception = e
 
+    def output(self, allow_partial=True) -> Tuple[str, Union[pd.DataFrame, partial]]:
+        if allow_partial:
+            return super.output()
+
+        for r, c in self._canvas.history[::-1]:
+            if r == "user":
+                return self._param.output_var_name, pd.DataFrame([{"content": c}])
+
+        self._param.output_var_name, pd.DataFrame([])
 

agent/component/arxiv.py

@@ -13,13 +13,12 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 import arxiv
 import pandas as pd
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 
-
 class ArXivParam(ComponentParamBase):
     """
     Define the ArXiv component parameters.
@@ -65,5 +64,5 @@ class ArXiv(ComponentBase, ABC):
             return ArXiv.be_output("")
 
         df = pd.DataFrame(arxiv_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"df: {str(df)}")
         return df

agent/component/baidu.py

@@ -13,13 +13,11 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-import random
+import logging
 from abc import ABC
-from functools import partial
 import pandas as pd
 import requests
 import re
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -46,7 +44,7 @@ class Baidu(ComponentBase, ABC):
             return Baidu.be_output("")
 
         try:
-            url = 'https://www.baidu.com/s?wd=' + ans + '&rn=' + str(self._param.top_n)
+            url = 'http://www.baidu.com/s?wd=' + ans + '&rn=' + str(self._param.top_n)
             headers = {
                 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.104 Safari/537.36'}
             response = requests.get(url=url, headers=headers)
@@ -64,6 +62,6 @@ class Baidu(ComponentBase, ABC):
             return Baidu.be_output("")
 
         df = pd.DataFrame(baidu_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"df: {str(df)}")
         return df
 

agent/component/baidufanyi.py

@@ -39,9 +39,6 @@ class BaiduFanyiParam(ComponentParamBase):
         self.check_empty(self.appid, "BaiduFanyi APPID")
         self.check_empty(self.secret_key, "BaiduFanyi Secret Key")
         self.check_valid_value(self.trans_type, "Translate type", ['translate', 'fieldtranslate'])
-        self.check_valid_value(self.trans_type, "Translate domain",
-                               ['it', 'finance', 'machinery', 'senimed', 'novel', 'academic', 'aerospace', 'wiki',
-                                'news', 'law', 'contract'])
         self.check_valid_value(self.source_lang, "Source language",
                                ['auto', 'zh', 'en', 'yue', 'wyw', 'jp', 'kor', 'fra', 'spa', 'th', 'ara', 'ru', 'pt',
                                 'de', 'it', 'el', 'nl', 'pl', 'bul', 'est', 'dan', 'fin', 'cs', 'rom', 'slo', 'swe',
@@ -96,3 +93,4 @@ class BaiduFanyi(ComponentBase, ABC):
 
         except Exception as e:
             BaiduFanyi.be_output("**Error**:" + str(e))
+    

agent/component/base.py

@@ -17,14 +17,13 @@ from abc import ABC
 import builtins
 import json
 import os
-from copy import deepcopy
+import logging
 from functools import partial
-from typing import List, Dict, Tuple, Union
+from typing import Tuple, Union
 
 import pandas as pd
 
 from agent import settings
-from agent.settings import flow_logger, DEBUG
 
 _FEEDED_DEPRECATED_PARAMS = "_feeded_deprecated_params"
 _DEPRECATED_PARAMS = "_deprecated_params"
@@ -36,6 +35,9 @@ class ComponentParamBase(ABC):
     def __init__(self):
         self.output_var_name = "output"
         self.message_history_window_size = 22
+        self.query = []
+        self.inputs = []
+        self.debug_inputs = []
 
     def set_name(self, name: str):
         self._name = name
@@ -81,7 +83,6 @@ class ComponentParamBase(ABC):
         return {name: True for name in self.get_feeded_deprecated_params()}
 
     def __str__(self):
-
         return json.dumps(self.as_dict(), ensure_ascii=False)
 
     def as_dict(self):
@@ -359,13 +360,13 @@ class ComponentParamBase(ABC):
 
     def _warn_deprecated_param(self, param_name, descr):
         if self._deprecated_params_set.get(param_name):
-            flow_logger.warning(
+            logging.warning(
                 f"{descr} {param_name} is deprecated and ignored in this version."
             )
 
     def _warn_to_deprecate_param(self, param_name, descr, new_param):
         if self._deprecated_params_set.get(param_name):
-            flow_logger.warning(
+            logging.warning(
                 f"{descr} {param_name} will be deprecated in future release; "
                 f"please use {new_param} instead."
             )
@@ -385,9 +386,13 @@ class ComponentBase(ABC):
         """
         return """{{
             "component_name": "{}",
-            "params": {}
+            "params": {},
+            "output": {},
+            "inputs": {}
         }}""".format(self.component_name,
-                     self._param
+                     self._param,
+                     json.dumps(json.loads(str(self._param)).get("output", {}), ensure_ascii=False),
+                     json.dumps(json.loads(str(self._param)).get("inputs", []), ensure_ascii=False)
         )
 
     def __init__(self, canvas, id, param: ComponentParamBase):
@@ -396,9 +401,17 @@ class ComponentBase(ABC):
         self._param = param
         self._param.check()
 
+    def get_dependent_components(self):
+        cpnts = set([para["component_id"].split("@")[0] for para in self._param.query \
+                     if para.get("component_id") \
+                     and para["component_id"].lower().find("answer") < 0 \
+                     and para["component_id"].lower().find("begin") < 0])
+        return list(cpnts)
+
     def run(self, history, **kwargs):
-        flow_logger.info("{}, history: {}, kwargs: {}".format(self, json.dumps(history, ensure_ascii=False),
+        logging.debug("{}, history: {}, kwargs: {}".format(self, json.dumps(history, ensure_ascii=False),
                                                               json.dumps(kwargs, ensure_ascii=False)))
+        self._param.debug_inputs = []
         try:
             res = self._run(history, **kwargs)
             self.set_output(res)
@@ -413,9 +426,14 @@ class ComponentBase(ABC):
 
     def output(self, allow_partial=True) -> Tuple[str, Union[pd.DataFrame, partial]]:
         o = getattr(self._param, self._param.output_var_name)
-        if not isinstance(o, partial) and not isinstance(o, pd.DataFrame):
-            if not isinstance(o, list): o = [o]
-            o = pd.DataFrame(o)
+        if not isinstance(o, partial):
+            if not isinstance(o, pd.DataFrame):
+                if isinstance(o, list):
+                    return self._param.output_var_name, pd.DataFrame(o)
+                if o is None:
+                    return self._param.output_var_name, pd.DataFrame()
+                return self._param.output_var_name, pd.DataFrame([{"content": str(o)}])
+            return self._param.output_var_name, o
 
         if allow_partial or not isinstance(o, partial):
             if not isinstance(o, partial) and not isinstance(o, pd.DataFrame):
@@ -426,53 +444,121 @@ class ComponentBase(ABC):
         for oo in o():
             if not isinstance(oo, pd.DataFrame):
                 outs = pd.DataFrame(oo if isinstance(oo, list) else [oo])
-            else: outs = oo
+            else:
+                outs = oo
         return self._param.output_var_name, outs
 
     def reset(self):
         setattr(self._param, self._param.output_var_name, None)
+        self._param.inputs = []
 
-    def set_output(self, v: pd.DataFrame):
+    def set_output(self, v):
         setattr(self._param, self._param.output_var_name, v)
 
     def get_input(self):
-        upstream_outs = []
+        if self._param.debug_inputs:
+            return pd.DataFrame([{"content": v["value"]} for v in self._param.debug_inputs if v.get("value")])
+
         reversed_cpnts = []
         if len(self._canvas.path) > 1:
             reversed_cpnts.extend(self._canvas.path[-2])
         reversed_cpnts.extend(self._canvas.path[-1])
 
-        if DEBUG: print(self.component_name, reversed_cpnts[::-1])
+        if self._param.query:
+            self._param.inputs = []
+            outs = []
+            for q in self._param.query:
+                if q.get("component_id"):
+                    if q["component_id"].split("@")[0].lower().find("begin") >= 0:
+                        cpn_id, key = q["component_id"].split("@")
+                        for p in self._canvas.get_component(cpn_id)["obj"]._param.query:
+                            if p["key"] == key:
+                                outs.append(pd.DataFrame([{"content": p.get("value", "")}]))
+                                self._param.inputs.append({"component_id": q["component_id"],
+                                                           "content": p.get("value", "")})
+                                break
+                        else:
+                            assert False, f"Can't find parameter '{key}' for {cpn_id}"
+                        continue
+
+                    if q["component_id"].lower().find("answer") == 0:
+                        txt = []
+                        for r, c in self._canvas.history[::-1][:self._param.message_history_window_size][::-1]:
+                            txt.append(f"{r.upper()}: {c}")
+                        txt = "\n".join(txt)
+                        self._param.inputs.append({"content": txt, "component_id": q["component_id"]})
+                        outs.append(pd.DataFrame([{"content": txt}]))
+                        continue
+
+                    outs.append(self._canvas.get_component(q["component_id"])["obj"].output(allow_partial=False)[1])
+                    self._param.inputs.append({"component_id": q["component_id"],
+                                               "content": "\n".join(
+                                                   [str(d["content"]) for d in outs[-1].to_dict('records')])})
+                elif q.get("value"):
+                    self._param.inputs.append({"component_id": None, "content": q["value"]})
+                    outs.append(pd.DataFrame([{"content": q["value"]}]))
+            if outs:
+                df = pd.concat(outs, ignore_index=True)
+                if "content" in df:
+                    df = df.drop_duplicates(subset=['content']).reset_index(drop=True)
+                return df
+
+        upstream_outs = []
+
         for u in reversed_cpnts[::-1]:
-            if self.get_component_name(u) in ["switch", "concentrator"]: continue
+            if self.get_component_name(u) in ["switch", "concentrator"]:
+                continue
             if self.component_name.lower() == "generate" and self.get_component_name(u) == "retrieval":
                 o = self._canvas.get_component(u)["obj"].output(allow_partial=False)[1]
                 if o is not None:
+                    o["component_id"] = u
                     upstream_outs.append(o)
                     continue
-            if u not in self._canvas.get_component(self._id)["upstream"]: continue
+            #if self.component_name.lower()!="answer" and u not in self._canvas.get_component(self._id)["upstream"]: continue
             if self.component_name.lower().find("switch") < 0 \
                     and self.get_component_name(u) in ["relevant", "categorize"]:
                 continue
             if u.lower().find("answer") >= 0:
                 for r, c in self._canvas.history[::-1]:
                     if r == "user":
-                        upstream_outs.append(pd.DataFrame([{"content": c}]))
+                        upstream_outs.append(pd.DataFrame([{"content": c, "component_id": u}]))
                         break
                 break
             if self.component_name.lower().find("answer") >= 0 and self.get_component_name(u) in ["relevant"]:
                 continue
             o = self._canvas.get_component(u)["obj"].output(allow_partial=False)[1]
             if o is not None:
+                o["component_id"] = u
                 upstream_outs.append(o)
             break
 
-        if upstream_outs:
+        assert upstream_outs, "Can't inference the where the component input is. Please identify whose output is this component's input."
+
         df = pd.concat(upstream_outs, ignore_index=True)
         if "content" in df:
             df = df.drop_duplicates(subset=['content']).reset_index(drop=True)
+
+        self._param.inputs = []
+        for _, r in df.iterrows():
+            self._param.inputs.append({"component_id": r["component_id"], "content": r["content"]})
+
         return df
-        return pd.DataFrame(self._canvas.get_history(3)[-1:])
+
+    def get_input_elements(self):
+        assert self._param.query, "Please identify input parameters firstly."
+        eles = []
+        for q in self._param.query:
+            if q.get("component_id"):
+                cpn_id = q["component_id"]
+                if cpn_id.split("@")[0].lower().find("begin") >= 0:
+                    cpn_id, key = cpn_id.split("@")
+                    eles.extend(self._canvas.get_component(cpn_id)["obj"]._param.query)
+                    continue
+
+                eles.append({"name": self._canvas.get_compnent_name(cpn_id), "key": cpn_id})
+            else:
+                eles.append({"key": q["value"], "name": q["value"], "value": q["value"]})
+        return eles
 
     def get_stream_input(self):
         reversed_cpnts = []
@@ -481,7 +567,8 @@ class ComponentBase(ABC):
         reversed_cpnts.extend(self._canvas.path[-1])
 
         for u in reversed_cpnts[::-1]:
-            if self.get_component_name(u) in ["switch", "answer"]: continue
+            if self.get_component_name(u) in ["switch", "answer"]:
+                continue
             return self._canvas.get_component(u)["obj"].output()[1]
 
     @staticmethod
@@ -490,3 +577,10 @@ class ComponentBase(ABC):
 
     def get_component_name(self, cpn_id):
         return self._canvas.get_component(cpn_id)["obj"].component_name.lower()
+
+    def debug(self, **kwargs):
+        return self._run([], **kwargs)
+
+    def get_parent(self):
+        pid = self._canvas.get_component(self._id)["parent_id"]
+        return self._canvas.get_component(pid)["obj"]

agent/component/begin.py

@@ -26,6 +26,7 @@ class BeginParam(ComponentParamBase):
     def __init__(self):
         super().__init__()
         self.prologue = "Hi! I'm your smart assistant. What can I do for you?"
+        self.query = []
 
     def check(self):
         return True
@@ -42,7 +43,7 @@ class Begin(ComponentBase):
     def stream_output(self):
         res = {"content": self._param.prologue}
         yield res
-        self.set_output(res)
+        self.set_output(self.be_output(res))
 
 
 

agent/component/bing.py

@@ -13,13 +13,12 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 import requests
 import pandas as pd
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 
-
 class BingParam(ComponentParamBase):
     """
     Define the Bing component parameters.
@@ -81,5 +80,5 @@ class Bing(ComponentBase, ABC):
             return Bing.be_output("")
 
         df = pd.DataFrame(bing_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"df: {str(df)}")
         return df

agent/component/categorize.py

@@ -13,11 +13,11 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 from api.db import LLMType
 from api.db.services.llm_service import LLMBundle
 from agent.component import GenerateParam, Generate
-from agent.settings import DEBUG
 
 
 class CategorizeParam(GenerateParam):
@@ -34,15 +34,18 @@ class CategorizeParam(GenerateParam):
         super().check()
         self.check_empty(self.category_description, "[Categorize] Category examples")
         for k, v in self.category_description.items():
-            if not k: raise ValueError(f"[Categorize] Category name can not be empty!")
-            if not v.get("to"): raise ValueError(f"[Categorize] 'To' of category {k} can not be empty!")
+            if not k:
+                raise ValueError("[Categorize] Category name can not be empty!")
+            if not v.get("to"):
+                raise ValueError(f"[Categorize] 'To' of category {k} can not be empty!")
 
-    def get_prompt(self):
+    def get_prompt(self, chat_hist):
         cate_lines = []
         for c, desc in self.category_description.items():
-            for l in desc.get("examples", "").split("\n"):
-                if not l: continue
-                cate_lines.append("Question: {}\tCategory: {}".format(l, c))
+            for line in desc.get("examples", "").split("\n"):
+                if not line:
+                    continue
+                cate_lines.append("USER: {}\nCategory: {}".format(line, c))
         descriptions = []
         for c, desc in self.category_description.items():
             if desc.get("description"):
@@ -59,11 +62,15 @@ class CategorizeParam(GenerateParam):
         {}
         You could learn from the above examples.
         Just mention the category names, no need for any additional words.
+        
+        ---- Real Data ----
+        {}
         """.format(
             len(self.category_description.keys()),
             "/".join(list(self.category_description.keys())),
             "\n".join(descriptions),
-            "- ".join(cate_lines)
+            "- ".join(cate_lines),
+            chat_hist
         )
         return self.prompt
 
@@ -73,15 +80,18 @@ class Categorize(Generate, ABC):
 
     def _run(self, history, **kwargs):
         input = self.get_input()
-        input = "Question: " + (list(input["content"])[-1] if "content" in input else "") + "\tCategory: "
         chat_mdl = LLMBundle(self._canvas.get_tenant_id(), LLMType.CHAT, self._param.llm_id)
-        ans = chat_mdl.chat(self._param.get_prompt(), [{"role": "user", "content": input}],
+        ans = chat_mdl.chat(self._param.get_prompt(input), [{"role": "user", "content": "\nCategory: "}],
                             self._param.gen_conf())
-        if DEBUG: print(ans, ":::::::::::::::::::::::::::::::::", input)
+        logging.debug(f"input: {input}, answer: {str(ans)}")
         for c in self._param.category_description.keys():
             if ans.lower().find(c.lower()) >= 0:
                 return Categorize.be_output(self._param.category_description[c]["to"])
 
         return Categorize.be_output(list(self._param.category_description.items())[-1][1]["to"])
 
+    def debug(self, **kwargs):
+        df = self._run([], **kwargs)
+        cpn_id = df.iloc[0, 0]
+        return Categorize.be_output(self._canvas.get_compnent_name(cpn_id))
 

agent/component/cite.py

@@ -1,75 +0,0 @@
-#
-#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
-#
-#  Licensed under the Apache License, Version 2.0 (the "License");
-#  you may not use this file except in compliance with the License.
-#  You may obtain a copy of the License at
-#
-#      http://www.apache.org/licenses/LICENSE-2.0
-#
-#  Unless required by applicable law or agreed to in writing, software
-#  distributed under the License is distributed on an "AS IS" BASIS,
-#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#  See the License for the specific language governing permissions and
-#  limitations under the License.
-#
-from abc import ABC
-
-import pandas as pd
-
-from api.db import LLMType
-from api.db.services.knowledgebase_service import KnowledgebaseService
-from api.db.services.llm_service import LLMBundle
-from api.settings import retrievaler
-from agent.component.base import ComponentBase, ComponentParamBase
-
-
-class CiteParam(ComponentParamBase):
-
-    """
-    Define the Retrieval component parameters.
-    """
-    def __init__(self):
-        super().__init__()
-        self.cite_sources = []
-
-    def check(self):
-        self.check_empty(self.cite_source, "Please specify where you want to cite from.")
-
-
-class Cite(ComponentBase, ABC):
-    component_name = "Cite"
-
-    def _run(self, history, **kwargs):
-        input = "\n- ".join(self.get_input()["content"])
-        sources = [self._canvas.get_component(cpn_id).output()[1] for cpn_id in self._param.cite_source]
-        query = []
-        for role, cnt in history[::-1][:self._param.message_history_window_size]:
-            if role != "user":continue
-            query.append(cnt)
-        query = "\n".join(query)
-
-        kbs = KnowledgebaseService.get_by_ids(self._param.kb_ids)
-        if not kbs:
-            raise ValueError("Can't find knowledgebases by {}".format(self._param.kb_ids))
-        embd_nms = list(set([kb.embd_id for kb in kbs]))
-        assert len(embd_nms) == 1, "Knowledge bases use different embedding models."
-
-        embd_mdl = LLMBundle(kbs[0].tenant_id, LLMType.EMBEDDING, embd_nms[0])
-
-        rerank_mdl = None
-        if self._param.rerank_id:
-            rerank_mdl = LLMBundle(kbs[0].tenant_id, LLMType.RERANK, self._param.rerank_id)
-
-        kbinfos = retrievaler.retrieval(query, embd_mdl, kbs[0].tenant_id, self._param.kb_ids,
-                                        1, self._param.top_n,
-                                        self._param.similarity_threshold, 1 - self._param.keywords_similarity_weight,
-                                        aggs=False, rerank_mdl=rerank_mdl)
-
-        if not kbinfos["chunks"]: return pd.DataFrame()
-        df = pd.DataFrame(kbinfos["chunks"])
-        df["content"] = df["content_with_weight"]
-        del df["content_with_weight"]
-        return df
-
-

agent/component/crawler.py

@@ -17,6 +17,7 @@ from abc import ABC
 import asyncio
 from crawl4ai import AsyncWebCrawler
 from agent.component.base import ComponentBase, ComponentParamBase
+from api.utils.web_utils import is_valid_url
 
 
 class CrawlerParam(ComponentParamBase):
@@ -39,8 +40,8 @@ class Crawler(ComponentBase, ABC):
     def _run(self, history, **kwargs):
         ans = self.get_input()
         ans = " - ".join(ans["content"]) if "content" in ans else ""
-        if not ans:
-            return Crawler.be_output("")
+        if not is_valid_url(ans):
+            return Crawler.be_output("URL not valid")
         try:
             result = asyncio.run(self.get_web(ans))
 
@@ -64,7 +65,3 @@ class Crawler(ComponentBase, ABC):
             elif self._param.extract_type == 'content':
                 result.extracted_content
             return result.markdown
-            
-
-
-    

agent/component/deepl.py

@@ -14,7 +14,6 @@
 #  limitations under the License.
 #
 from abc import ABC
-import re
 from agent.component.base import ComponentBase, ComponentParamBase
 import deepl
 

agent/component/duckduckgo.py

@@ -13,10 +13,10 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 from duckduckgo_search import DDGS
 import pandas as pd
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -62,5 +62,5 @@ class DuckDuckGo(ComponentBase, ABC):
             return DuckDuckGo.be_output("")
 
         df = pd.DataFrame(duck_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug("df: {df}")
         return df

agent/component/email.py

@@ -0,0 +1,138 @@
+#
+#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+
+from abc import ABC
+import json
+import smtplib
+import logging
+from email.mime.text import MIMEText
+from email.mime.multipart import MIMEMultipart
+from email.header import Header
+from email.utils import formataddr
+from agent.component.base import ComponentBase, ComponentParamBase
+
+class EmailParam(ComponentParamBase):
+    """
+    Define the Email component parameters.
+    """
+    def __init__(self):
+        super().__init__()
+        # Fixed configuration parameters
+        self.smtp_server = ""  # SMTP server address
+        self.smtp_port = 465  # SMTP port
+        self.email = ""  # Sender email
+        self.password = ""  # Email authorization code
+        self.sender_name = ""  # Sender name
+
+    def check(self):
+        # Check required parameters
+        self.check_empty(self.smtp_server, "SMTP Server")
+        self.check_empty(self.email, "Email")
+        self.check_empty(self.password, "Password")
+        self.check_empty(self.sender_name, "Sender Name")
+
+class Email(ComponentBase, ABC):
+    component_name = "Email"
+    
+    def _run(self, history, **kwargs):
+        # Get upstream component output and parse JSON
+        ans = self.get_input()
+        content = "".join(ans["content"]) if "content" in ans else ""
+        if not content:
+            return Email.be_output("No content to send")
+
+        success = False
+        try:
+            # Parse JSON string passed from upstream
+            email_data = json.loads(content)
+            
+            # Validate required fields
+            if "to_email" not in email_data:
+                return Email.be_output("Missing required field: to_email")
+
+            # Create email object
+            msg = MIMEMultipart('alternative')
+            
+            # Properly handle sender name encoding
+            msg['From'] = formataddr((str(Header(self._param.sender_name,'utf-8')), self._param.email))
+            msg['To'] = email_data["to_email"]
+            if "cc_email" in email_data and email_data["cc_email"]:
+                msg['Cc'] = email_data["cc_email"]
+            msg['Subject'] = Header(email_data.get("subject", "No Subject"), 'utf-8').encode()
+
+            # Use content from email_data or default content
+            email_content = email_data.get("content", "No content provided")
+            # msg.attach(MIMEText(email_content, 'plain', 'utf-8'))
+            msg.attach(MIMEText(email_content, 'html', 'utf-8'))
+
+            # Connect to SMTP server and send
+            logging.info(f"Connecting to SMTP server {self._param.smtp_server}:{self._param.smtp_port}")
+            
+            context = smtplib.ssl.create_default_context()
+            with smtplib.SMTP_SSL(self._param.smtp_server, self._param.smtp_port, context=context) as server:
+                # Login
+                logging.info(f"Attempting to login with email: {self._param.email}")
+                server.login(self._param.email, self._param.password)
+                
+                # Get all recipient list
+                recipients = [email_data["to_email"]]
+                if "cc_email" in email_data and email_data["cc_email"]:
+                    recipients.extend(email_data["cc_email"].split(','))
+                
+                # Send email
+                logging.info(f"Sending email to recipients: {recipients}")
+                try:
+                    server.send_message(msg, self._param.email, recipients)
+                    success = True
+                except Exception as e:
+                    logging.error(f"Error during send_message: {str(e)}")
+                    # Try alternative method
+                    server.sendmail(self._param.email, recipients, msg.as_string())
+                    success = True
+                
+                try:
+                    server.quit()
+                except Exception as e:
+                    # Ignore errors when closing connection
+                    logging.warning(f"Non-fatal error during connection close: {str(e)}")
+
+            if success:
+                return Email.be_output("Email sent successfully")
+
+        except json.JSONDecodeError:
+            error_msg = "Invalid JSON format in input"
+            logging.error(error_msg)
+            return Email.be_output(error_msg)
+            
+        except smtplib.SMTPAuthenticationError:
+            error_msg = "SMTP Authentication failed. Please check your email and authorization code."
+            logging.error(error_msg)
+            return Email.be_output(f"Failed to send email: {error_msg}")
+            
+        except smtplib.SMTPConnectError:
+            error_msg = f"Failed to connect to SMTP server {self._param.smtp_server}:{self._param.smtp_port}"
+            logging.error(error_msg)
+            return Email.be_output(f"Failed to send email: {error_msg}")
+            
+        except smtplib.SMTPException as e:
+            error_msg = f"SMTP error occurred: {str(e)}"
+            logging.error(error_msg)
+            return Email.be_output(f"Failed to send email: {error_msg}")
+            
+        except Exception as e:
+            error_msg = f"Unexpected error: {str(e)}"
+            logging.error(error_msg)
+            return Email.be_output(f"Failed to send email: {error_msg}") 

agent/component/exesql.py

@@ -15,13 +15,17 @@
 #
 from abc import ABC
 import re
+from copy import deepcopy
+
 import pandas as pd
 import pymysql
 import psycopg2
-from agent.component.base import ComponentBase, ComponentParamBase
+from agent.component import GenerateParam, Generate
+import pyodbc
+import logging
 
 
-class ExeSQLParam(ComponentParamBase):
+class ExeSQLParam(GenerateParam):
     """
     Define the ExeSQL component parameters.
     """
@@ -38,7 +42,8 @@ class ExeSQLParam(ComponentParamBase):
         self.top_n = 30
 
     def check(self):
-        self.check_valid_value(self.db_type, "Choose DB type", ['mysql', 'postgresql', 'mariadb'])
+        super().check()
+        self.check_valid_value(self.db_type, "Choose DB type", ['mysql', 'postgresql', 'mariadb', 'mssql'])
         self.check_empty(self.database, "Database name")
         self.check_empty(self.username, "database username")
         self.check_empty(self.host, "IP Address")
@@ -46,58 +51,105 @@ class ExeSQLParam(ComponentParamBase):
         self.check_empty(self.password, "Database password")
         self.check_positive_integer(self.top_n, "Number of records")
         if self.database == "rag_flow":
-            if self.host == "ragflow-mysql": raise ValueError("The host is not accessible.")
-            if self.password == "infini_rag_flow": raise ValueError("The host is not accessible.")
+            if self.host == "ragflow-mysql":
+                raise ValueError("The host is not accessible.")
+            if self.password == "infini_rag_flow":
+                raise ValueError("The host is not accessible.")
 
 
-class ExeSQL(ComponentBase, ABC):
+class ExeSQL(Generate, ABC):
     component_name = "ExeSQL"
 
-    def _run(self, history, **kwargs):
-        if not hasattr(self, "_loop"):
-            setattr(self, "_loop", 0)
-        if self._loop >= self._param.loop:
-            self._loop = 0
-            raise Exception("Maximum loop time exceeds. Can't query the correct data via SQL statement.")
-        self._loop += 1
-
-        ans = self.get_input()
-        ans = "".join(ans["content"]) if "content" in ans else ""
-        ans = re.sub(r'^.*?SELECT ', 'SELECT ', repr(ans), flags=re.IGNORECASE)
+    def _refactor(self,ans):
+        match = re.search(r"```sql\s*(.*?)\s*```", ans, re.DOTALL)
+        if match:
+            ans = match.group(1)  # Query content
+            return ans
+        else:
+            print("no markdown")
+        ans = re.sub(r'^.*?SELECT ', 'SELECT ', (ans), flags=re.IGNORECASE)
         ans = re.sub(r';.*?SELECT ', '; SELECT ', ans, flags=re.IGNORECASE)
         ans = re.sub(r';[^;]*$', r';', ans)
         if not ans:
             raise Exception("SQL statement not found!")
+        return ans
 
+    def _run(self, history, **kwargs):
+        ans = self.get_input()
+        ans = "".join([str(a) for a in ans["content"]]) if "content" in ans else ""
+        ans = self._refactor(ans)
+        logging.info("db_type: ",self._param.db_type)
         if self._param.db_type in ["mysql", "mariadb"]:
             db = pymysql.connect(db=self._param.database, user=self._param.username, host=self._param.host,
                                  port=self._param.port, password=self._param.password)
         elif self._param.db_type == 'postgresql':
             db = psycopg2.connect(dbname=self._param.database, user=self._param.username, host=self._param.host,
                                   port=self._param.port, password=self._param.password)
-
+        elif self._param.db_type == 'mssql':
+            conn_str = (
+                r'DRIVER={ODBC Driver 17 for SQL Server};'
+                r'SERVER=' + self._param.host + ',' + str(self._param.port) + ';'
+                r'DATABASE=' + self._param.database + ';'
+                r'UID=' + self._param.username + ';'
+                r'PWD=' + self._param.password
+            )
+            db = pyodbc.connect(conn_str)
         try:
             cursor = db.cursor()
         except Exception as e:
             raise Exception("Database Connection Failed! \n" + str(e))
+        if not hasattr(self, "_loop"):
+            setattr(self, "_loop", 0)
+            self._loop += 1
+        input_list=re.split(r';', ans.replace(r"\n", " "))
         sql_res = []
-        for single_sql in re.split(r';', ans.replace(r"\n", " ")):
+        for i in range(len(input_list)):
+            single_sql=input_list[i]
+            while self._loop <= self._param.loop:
+                self._loop+=1
                 if not single_sql:
-                continue
+                    break
                 try:
+                    logging.info("single_sql: ", single_sql)
                     cursor.execute(single_sql)
                     if cursor.rowcount == 0:
-                    sql_res.append({"content": "\nTotal: 0\n No record in the database!"})
-                    continue
-                single_res = pd.DataFrame([i for i in cursor.fetchmany(size=self._param.top_n)])
+                        sql_res.append({"content": "No record in the database!"})
+                        break
+                    if self._param.db_type == 'mssql':
+                        single_res  = pd.DataFrame.from_records(cursor.fetchmany(self._param.top_n),columns = [desc[0] for desc in cursor.description])
+                    else:
+                        single_res = pd.DataFrame([i for i in cursor.fetchmany(self._param.top_n)])
                         single_res.columns = [i[0] for i in cursor.description]
-                sql_res.append({"content": "\nTotal: " + str(cursor.rowcount) + "\n" + single_res.to_markdown()})
+                    sql_res.append({"content":  single_res.to_markdown()})
+                    break
                 except Exception as e:
-                sql_res.append({"content": "**Error**:" + str(e) + "\nError SQL Statement:" + single_sql})
-                pass
+                    single_sql = self._regenerate_sql(single_sql, str(e), **kwargs)
+                    single_sql = self._refactor(single_sql)
+                    if self._loop > self._param.loop:
+                        sql_res.append({"content": "Can't query the correct data via SQL statement."})
+                        # raise Exception("Maximum loop time exceeds. Can't query the correct data via SQL statement.")
         db.close()
-
         if not sql_res:
             return ExeSQL.be_output("")
-
         return pd.DataFrame(sql_res)
+
+    def _regenerate_sql(self, failed_sql, error_message,**kwargs):
+        prompt = f'''
+        ## You are the Repair SQL Statement Helper, please modify the original SQL statement based on the SQL query error report.
+        ## The original SQL statement is as follows:{failed_sql}.
+        ## The contents of the SQL query error report is as follows:{error_message}.
+        ## Answer only the modified SQL statement. Please do not give any explanation, just answer the code.
+'''
+        self._param.prompt=prompt
+        kwargs_ = deepcopy(kwargs)
+        kwargs_["stream"] = False
+        response = Generate._run(self, [], **kwargs_)
+        try:
+            regenerated_sql = response.loc[0,"content"]
+            return regenerated_sql
+        except Exception as e:
+            logging.error(f"Failed to regenerate SQL: {e}")
+            return None
+
+    def debug(self, **kwargs):
+        return self._run([], **kwargs)

agent/component/generate.py

@@ -17,9 +17,10 @@ import re
 from functools import partial
 import pandas as pd
 from api.db import LLMType
+from api.db.services.conversation_service import structure_answer
 from api.db.services.dialog_service import message_fit_in
 from api.db.services.llm_service import LLMBundle
-from api.settings import retrievaler
+from api import settings
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -51,11 +52,16 @@ class GenerateParam(ComponentParamBase):
 
     def gen_conf(self):
         conf = {}
-        if self.max_tokens > 0: conf["max_tokens"] = self.max_tokens
-        if self.temperature > 0: conf["temperature"] = self.temperature
-        if self.top_p > 0: conf["top_p"] = self.top_p
-        if self.presence_penalty > 0: conf["presence_penalty"] = self.presence_penalty
-        if self.frequency_penalty > 0: conf["frequency_penalty"] = self.frequency_penalty
+        if self.max_tokens > 0:
+            conf["max_tokens"] = self.max_tokens
+        if self.temperature > 0:
+            conf["temperature"] = self.temperature
+        if self.top_p > 0:
+            conf["top_p"] = self.top_p
+        if self.presence_penalty > 0:
+            conf["presence_penalty"] = self.presence_penalty
+        if self.frequency_penalty > 0:
+            conf["frequency_penalty"] = self.frequency_penalty
         return conf
 
 
@@ -63,14 +69,18 @@ class Generate(ComponentBase):
     component_name = "Generate"
 
     def get_dependent_components(self):
-        cpnts = [para["component_id"] for para in self._param.parameters]
-        return cpnts
+        cpnts = set([para["component_id"].split("@")[0] for para in self._param.parameters \
+                     if para.get("component_id") \
+                     and para["component_id"].lower().find("answer") < 0 \
+                     and para["component_id"].lower().find("begin") < 0])
+        return list(cpnts)
 
     def set_cite(self, retrieval_res, answer):
         retrieval_res = retrieval_res.dropna(subset=["vector", "content_ltks"]).reset_index(drop=True)
         if "empty_response" in retrieval_res.columns:
             retrieval_res["empty_response"].fillna("", inplace=True)
-        answer, idx = retrievaler.insert_citations(answer, [ck["content_ltks"] for _, ck in retrieval_res.iterrows()],
+        answer, idx = settings.retrievaler.insert_citations(answer,
+                                                            [ck["content_ltks"] for _, ck in retrieval_res.iterrows()],
                                                             [ck["vector"] for _, ck in retrieval_res.iterrows()],
                                                             LLMBundle(self._canvas.get_tenant_id(), LLMType.EMBEDDING,
                                                                       self._canvas.get_embedding_model()), tkweight=0.7,
@@ -79,7 +89,8 @@ class Generate(ComponentBase):
         recall_docs = []
         for i in idx:
             did = retrieval_res.loc[int(i), "doc_id"]
-            if did in doc_ids: continue
+            if did in doc_ids:
+                continue
             doc_ids.add(did)
             recall_docs.append({"doc_id": did, "doc_name": retrieval_res.loc[int(i), "docnm_kwd"]})
 
@@ -92,31 +103,71 @@ class Generate(ComponentBase):
         }
 
         if answer.lower().find("invalid key") >= 0 or answer.lower().find("invalid api") >= 0:
-            answer += " Please set LLM API-Key in 'User Setting -> Model Providers -> API-Key'"
+            answer += " Please set LLM API-Key in 'User Setting -> Model providers -> API-Key'"
         res = {"content": answer, "reference": reference}
+        res = structure_answer(None, res, "", "")
 
         return res
 
+    def get_input_elements(self):
+        if self._param.parameters:
+            return [{"key": "user", "name": "Input your question here:"}, *self._param.parameters]
+
+        return [{"key": "user", "name": "Input your question here:"}]
+
     def _run(self, history, **kwargs):
         chat_mdl = LLMBundle(self._canvas.get_tenant_id(), LLMType.CHAT, self._param.llm_id)
         prompt = self._param.prompt
 
-        retrieval_res = self.get_input()
-        input = ("  - "+"\n  - ".join([c for c in retrieval_res["content"] if isinstance(c, str)])) if "content" in retrieval_res else ""
+        retrieval_res = []
+        self._param.inputs = []
         for para in self._param.parameters:
-            cpn = self._canvas.get_component(para["component_id"])["obj"]
+            if not para.get("component_id"):
+                continue
+            component_id = para["component_id"].split("@")[0]
+            if para["component_id"].lower().find("@") >= 0:
+                cpn_id, key = para["component_id"].split("@")
+                for p in self._canvas.get_component(cpn_id)["obj"]._param.query:
+                    if p["key"] == key:
+                        kwargs[para["key"]] = p.get("value", "")
+                        self._param.inputs.append(
+                            {"component_id": para["component_id"], "content": kwargs[para["key"]]})
+                        break
+                else:
+                    assert False, f"Can't find parameter '{key}' for {cpn_id}"
+                continue
+
+            cpn = self._canvas.get_component(component_id)["obj"]
             if cpn.component_name.lower() == "answer":
-                kwargs[para["key"]] = self._canvas.get_history(1)[0]["content"]
+                hist = self._canvas.get_history(1)
+                if hist:
+                    hist = hist[0]["content"]
+                else:
+                    hist = ""
+                kwargs[para["key"]] = hist
                 continue
             _, out = cpn.output(allow_partial=False)
             if "content" not in out.columns:
-                kwargs[para["key"]] = "Nothing"
+                kwargs[para["key"]] = ""
             else:
+                if cpn.component_name.lower() == "retrieval":
+                    retrieval_res.append(out)
                 kwargs[para["key"]] = "  - "+"\n - ".join([o if isinstance(o, str) else str(o) for o in out["content"]])
+            self._param.inputs.append({"component_id": para["component_id"], "content": kwargs[para["key"]]})
+
+        if retrieval_res:
+            retrieval_res = pd.concat(retrieval_res, ignore_index=True)
+        else:
+            retrieval_res = pd.DataFrame([])
 
-        kwargs["input"] = input
         for n, v in kwargs.items():
-            prompt = re.sub(r"\{%s\}" % re.escape(n), re.escape(str(v)), prompt)
+            prompt = re.sub(r"\{%s\}" % re.escape(n), str(v).replace("\\", " "), prompt)
+
+        if not self._param.inputs and prompt.find("{input}") >= 0:
+            retrieval_res = self.get_input()
+            input = ("  - " + "\n  - ".join(
+                [c for c in retrieval_res["content"] if isinstance(c, str)])) if "content" in retrieval_res else ""
+            prompt = re.sub(r"\{input\}", re.escape(input), prompt)
 
         downstreams = self._canvas.get_component(self._id)["downstream"]
         if kwargs.get("stream") and len(downstreams) == 1 and self._canvas.get_component(downstreams[0])[
@@ -129,7 +180,11 @@ class Generate(ComponentBase):
             return pd.DataFrame([res])
 
         msg = self._canvas.get_history(self._param.message_history_window_size)
+        if len(msg) < 1:
+            msg.append({"role": "user", "content": ""})
         _, msg = message_fit_in([{"role": "system", "content": prompt}, *msg], int(chat_mdl.max_length * 0.97))
+        if len(msg) < 2:
+            msg.append({"role": "user", "content": ""})
         ans = chat_mdl.chat(msg[0]["content"], msg[1:], self._param.gen_conf())
 
         if self._param.cite and "content_ltks" in retrieval_res.columns and "vector" in retrieval_res.columns:
@@ -148,7 +203,11 @@ class Generate(ComponentBase):
             return
 
         msg = self._canvas.get_history(self._param.message_history_window_size)
+        if len(msg) < 1:
+            msg.append({"role": "user", "content": ""})
         _, msg = message_fit_in([{"role": "system", "content": prompt}, *msg], int(chat_mdl.max_length * 0.97))
+        if len(msg) < 2:
+            msg.append({"role": "user", "content": ""})
         answer = ""
         for ans in chat_mdl.chat_streamly(msg[0]["content"], msg[1:], self._param.gen_conf()):
             res = {"content": ans, "reference": []}
@@ -159,4 +218,17 @@ class Generate(ComponentBase):
             res = self.set_cite(retrieval_res, answer)
             yield res
 
-        self.set_output(res)
+        self.set_output(Generate.be_output(res))
+
+    def debug(self, **kwargs):
+        chat_mdl = LLMBundle(self._canvas.get_tenant_id(), LLMType.CHAT, self._param.llm_id)
+        prompt = self._param.prompt
+
+        for para in self._param.debug_inputs:
+            kwargs[para["key"]] = para.get("value", "")
+
+        for n, v in kwargs.items():
+            prompt = re.sub(r"\{%s\}" % re.escape(n), str(v).replace("\\", " "), prompt)
+
+        ans = chat_mdl.chat(prompt, [{"role": "user", "content": kwargs.get("user", "")}], self._param.gen_conf())
+        return pd.DataFrame([ans])

agent/component/github.py

@@ -13,10 +13,10 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 import pandas as pd
 import requests
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -57,5 +57,5 @@ class GitHub(ComponentBase, ABC):
             return GitHub.be_output("")
 
         df = pd.DataFrame(github_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"df: {df}")
         return df

agent/component/google.py

@@ -13,10 +13,10 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 from serpapi import GoogleSearch
 import pandas as pd
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -85,12 +85,12 @@ class Google(ComponentBase, ABC):
                  "hl": self._param.language, "num": self._param.top_n})
             google_res = [{"content": '<a href="' + i["link"] + '">' + i["title"] + '</a>    ' + i["snippet"]} for i in
                           client.get_dict()["organic_results"]]
-        except Exception as e:
+        except Exception:
             return Google.be_output("**ERROR**: Existing Unavailable Parameters!")
 
         if not google_res:
             return Google.be_output("")
 
         df = pd.DataFrame(google_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"df: {df}")
         return df

agent/component/googlescholar.py

@@ -13,9 +13,9 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 import pandas as pd
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 from scholarly import scholarly
 
@@ -58,13 +58,13 @@ class GoogleScholar(ComponentBase, ABC):
                     'pub_url'] + '"></a> ' + "\n author: " + ",".join(pub['bib']['author']) + '\n Abstract: ' + pub[
                                                    'bib'].get('abstract', 'no abstract')})
 
-            except StopIteration or Exception as e:
-                print("**ERROR** " + str(e))
+            except StopIteration or Exception:
+                logging.exception("GoogleScholar")
                 break
 
         if not scholar_res:
             return GoogleScholar.be_output("")
 
         df = pd.DataFrame(scholar_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"df: {df}")
         return df

agent/component/invoke.py

@@ -50,11 +50,24 @@ class Invoke(ComponentBase, ABC):
         args = {}
         for para in self._param.variables:
             if para.get("component_id"):
+                if '@' in para["component_id"]:
+                    component = para["component_id"].split('@')[0]
+                    field = para["component_id"].split('@')[1]
+                    cpn = self._canvas.get_component(component)["obj"]
+                    for param in cpn._param.query:
+                        if param["key"] == field:
+                            if "value" in param:
+                                args[para["key"]] = param["value"]
+                else:
                     cpn = self._canvas.get_component(para["component_id"])["obj"]
+                    if cpn.component_name.lower() == "answer":
+                        args[para["key"]] = self._canvas.get_history(1)[0]["content"]
+                        continue
                     _, out = cpn.output(allow_partial=False)
+                    if not out.empty:
                         args[para["key"]] = "\n".join(out["content"])
             else:
-                args[para["key"]] = "\n".join(para["value"])
+                args[para["key"]] = para["value"]
 
         url = self._param.url.strip()
         if url.find("http") != 0:

agent/component/iteration.py

@@ -0,0 +1,45 @@
+#
+#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+from abc import ABC
+from agent.component.base import ComponentBase, ComponentParamBase
+
+
+class IterationParam(ComponentParamBase):
+    """
+    Define the Iteration component parameters.
+    """
+
+    def __init__(self):
+        super().__init__()
+        self.delimiter = ","
+
+    def check(self):
+        self.check_empty(self.delimiter, "Delimiter")
+
+
+class Iteration(ComponentBase, ABC):
+    component_name = "Iteration"
+
+    def get_start(self):
+        for cid in self._canvas.components.keys():
+            if self._canvas.get_component(cid)["obj"].component_name.lower() != "iterationitem":
+                continue
+            if self._canvas.get_component(cid)["parent_id"] == self._id:
+                return self._canvas.get_component(cid)
+
+    def _run(self, history, **kwargs):
+        return self.output(allow_partial=False)[1]
+

agent/component/iterationitem.py

@@ -0,0 +1,49 @@
+#
+#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+from abc import ABC
+import pandas as pd
+from agent.component.base import ComponentBase, ComponentParamBase
+
+
+class IterationItemParam(ComponentParamBase):
+    """
+    Define the IterationItem component parameters.
+    """
+    def check(self):
+        return True
+
+
+class IterationItem(ComponentBase, ABC):
+    component_name = "IterationItem"
+
+    def __init__(self, canvas, id, param: ComponentParamBase):
+        super().__init__(canvas, id, param)
+        self._idx = 0
+
+    def _run(self, history, **kwargs):
+        parent = self.get_parent()
+        ans = parent.get_input()
+        ans = parent._param.delimiter.join(ans["content"]) if "content" in ans else ""
+        ans = [a.strip() for a in ans.split(parent._param.delimiter)]
+        df = pd.DataFrame([{"content": ans[self._idx]}])
+        self._idx += 1
+        if self._idx >= len(ans):
+            self._idx = -1
+        return df
+
+    def end(self):
+        return self._idx == -1
+

agent/component/keyword.py

@@ -13,12 +13,12 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 import re
 from abc import ABC
 from api.db import LLMType
 from api.db.services.llm_service import LLMBundle
 from agent.component import GenerateParam, Generate
-from agent.settings import DEBUG
 
 
 class KeywordExtractParam(GenerateParam):
@@ -50,16 +50,16 @@ class KeywordExtract(Generate, ABC):
     component_name = "KeywordExtract"
 
     def _run(self, history, **kwargs):
-        q = ""
-        for r, c in self._canvas.history[::-1]:
-            if r == "user":
-                q += c
-                break
+        query = self.get_input()
+        query = str(query["content"][0]) if "content" in query else ""
 
         chat_mdl = LLMBundle(self._canvas.get_tenant_id(), LLMType.CHAT, self._param.llm_id)
-        ans = chat_mdl.chat(self._param.get_prompt(), [{"role": "user", "content": q}],
+        ans = chat_mdl.chat(self._param.get_prompt(), [{"role": "user", "content": query}],
                             self._param.gen_conf())
 
         ans = re.sub(r".*keyword:", "", ans).strip()
-        if DEBUG: print(ans, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"ans: {ans}")
         return KeywordExtract.be_output(ans)
+
+    def debug(self, **kwargs):
+        return self._run([], **kwargs)

agent/component/pubmed.py

@@ -13,12 +13,12 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 from Bio import Entrez
 import re
 import pandas as pd
 import xml.etree.ElementTree as ET
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -65,5 +65,5 @@ class PubMed(ComponentBase, ABC):
             return PubMed.be_output("")
 
         df = pd.DataFrame(pubmed_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"df: {df}")
         return df

agent/component/relevant.py

@@ -13,6 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 from api.db import LLMType
 from api.db.services.llm_service import LLMBundle
@@ -70,11 +71,13 @@ class Relevant(Generate, ABC):
         ans = chat_mdl.chat(self._param.get_prompt(), [{"role": "user", "content": ans}],
                             self._param.gen_conf())
 
-        print(ans, ":::::::::::::::::::::::::::::::::")
+        logging.debug(ans)
         if ans.lower().find("yes") >= 0:
             return Relevant.be_output(self._param.yes)
         if ans.lower().find("no") >= 0:
             return Relevant.be_output(self._param.no)
         assert False, f"Relevant component got: {ans}"
 
+    def debug(self, **kwargs):
+        return self._run([], **kwargs)
 

agent/component/retrieval.py

@@ -13,14 +13,16 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 
 import pandas as pd
 
 from api.db import LLMType
+from api.db.services.dialog_service import label_question
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.llm_service import LLMBundle
-from api.settings import retrievaler
+from api import settings
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -66,10 +68,11 @@ class Retrieval(ComponentBase, ABC):
         if self._param.rerank_id:
             rerank_mdl = LLMBundle(kbs[0].tenant_id, LLMType.RERANK, self._param.rerank_id)
 
-        kbinfos = retrievaler.retrieval(query, embd_mdl, kbs[0].tenant_id, self._param.kb_ids,
+        kbinfos = settings.retrievaler.retrieval(query, embd_mdl, kbs[0].tenant_id, self._param.kb_ids,
                                         1, self._param.top_n,
                                         self._param.similarity_threshold, 1 - self._param.keywords_similarity_weight,
-                                        aggs=False, rerank_mdl=rerank_mdl)
+                                        aggs=False, rerank_mdl=rerank_mdl,
+                                        rank_feature=label_question(query, kbs))
 
         if not kbinfos["chunks"]:
             df = Retrieval.be_output("")
@@ -80,7 +83,7 @@ class Retrieval(ComponentBase, ABC):
         df = pd.DataFrame(kbinfos["chunks"])
         df["content"] = df["content_with_weight"]
         del df["content_with_weight"]
-        print(">>>>>>>>>>>>>>>>>>>>>>>>>>\n", query, df)
+        logging.debug("{} {}".format(query, df))
         return df
 
 

agent/component/rewrite.py

@@ -13,6 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 from api.db import LLMType
 from api.db.services.llm_service import LLMBundle
@@ -28,7 +29,6 @@ class RewriteQuestionParam(GenerateParam):
         super().__init__()
         self.temperature = 0.9
         self.prompt = ""
-        self.loop = 1
 
     def check(self):
         super().check()
@@ -84,17 +84,11 @@ class RewriteQuestion(Generate, ABC):
     component_name = "RewriteQuestion"
 
     def _run(self, history, **kwargs):
-        if not hasattr(self, "_loop"):
-            setattr(self, "_loop", 0)
-        if self._loop >= self._param.loop:
-            self._loop = 0
-            raise Exception("Sorry! Nothing relevant found.")
-        self._loop += 1
-
-        hist = self._canvas.get_history(4)
+        hist = self._canvas.get_history(self._param.message_history_window_size)
         conv = []
         for m in hist:
-            if m["role"] not in ["user", "assistant"]: continue
+            if m["role"] not in ["user", "assistant"]:
+                continue
             conv.append("{}: {}".format(m["role"].upper(), m["content"]))
         conv = "\n".join(conv)
 
@@ -104,7 +98,8 @@ class RewriteQuestion(Generate, ABC):
         self._canvas.history.pop()
         self._canvas.history.append(("user", ans))
 
-        print(ans, ":::::::::::::::::::::::::::::::::")
+        logging.debug(ans)
         return RewriteQuestion.be_output(ans)
 
 
+

agent/component/switch.py

@@ -41,19 +41,44 @@ class SwitchParam(ComponentParamBase):
     def check(self):
         self.check_empty(self.conditions, "[Switch] conditions")
         for cond in self.conditions:
-            if not cond["to"]: raise ValueError(f"[Switch] 'To' can not be empty!")
+            if not cond["to"]:
+                raise ValueError("[Switch] 'To' can not be empty!")
 
 
 class Switch(ComponentBase, ABC):
     component_name = "Switch"
 
+    def get_dependent_components(self):
+        res = []
+        for cond in self._param.conditions:
+            for item in cond["items"]:
+                if not item["cpn_id"]:
+                    continue
+                if item["cpn_id"].find("begin") >= 0:
+                    continue
+                cid = item["cpn_id"].split("@")[0]
+                res.append(cid)
+
+        return list(set(res))
+
     def _run(self, history, **kwargs):
         for cond in self._param.conditions:
             res = []
             for item in cond["items"]:
-                out = self._canvas.get_component(item["cpn_id"])["obj"].output()[1]
-                cpn_input = "" if "content" not in out.columns else " ".join(out["content"])
-                res.append(self.process_operator(cpn_input, item["operator"], item["value"]))
+                if not item["cpn_id"]:
+                    continue
+                cid = item["cpn_id"].split("@")[0]
+                if item["cpn_id"].find("@") > 0:
+                    cpn_id, key = item["cpn_id"].split("@")
+                    for p in self._canvas.get_component(cid)["obj"]._param.query:
+                        if p["key"] == key:
+                            res.append(self.process_operator(p.get("value",""), item["operator"], item.get("value", "")))
+                            break
+                else:
+                    out = self._canvas.get_component(cid)["obj"].output()[1]
+                    cpn_input = "" if "content" not in out.columns else " ".join([str(s) for s in out["content"]])
+                    res.append(self.process_operator(cpn_input, item["operator"], item.get("value", "")))
+
                 if cond["logical_operator"] != "and" and any(res):
                     return Switch.be_output(cond["to"])
 
@@ -85,22 +110,22 @@ class Switch(ComponentBase, ABC):
         elif operator == ">":
             try:
                 return True if float(input) > float(value) else False
-            except Exception as e:
+            except Exception:
                 return True if input > value else False
         elif operator == "<":
             try:
                 return True if float(input) < float(value) else False
-            except Exception as e:
+            except Exception:
                 return True if input < value else False
         elif operator == "≥":
             try:
                 return True if float(input) >= float(value) else False
-            except Exception as e:
+            except Exception:
                 return True if input >= value else False
         elif operator == "≤":
             try:
                 return True if float(input) <= float(value) else False
-            except Exception as e:
+            except Exception:
                 return True if input <= value else False
 
         raise ValueError('Not supported operator' + operator)

agent/component/template.py

@@ -0,0 +1,123 @@
+#
+#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+import json
+import re
+from agent.component.base import ComponentBase, ComponentParamBase
+from jinja2 import Template as Jinja2Template
+
+
+class TemplateParam(ComponentParamBase):
+    """
+    Define the Generate component parameters.
+    """
+
+    def __init__(self):
+        super().__init__()
+        self.content = ""
+        self.parameters = []
+
+    def check(self):
+        self.check_empty(self.content, "[Template] Content")
+        return True
+
+
+class Template(ComponentBase):
+    component_name = "Template"
+
+    def get_dependent_components(self):
+        cpnts = set(
+            [
+                para["component_id"].split("@")[0]
+                for para in self._param.parameters
+                if para.get("component_id")
+                and para["component_id"].lower().find("answer") < 0
+                and para["component_id"].lower().find("begin") < 0
+            ]
+        )
+        return list(cpnts)
+
+    def _run(self, history, **kwargs):
+        content = self._param.content
+
+        self._param.inputs = []
+        for para in self._param.parameters:
+            if not para.get("component_id"):
+                continue
+            component_id = para["component_id"].split("@")[0]
+            if para["component_id"].lower().find("@") >= 0:
+                cpn_id, key = para["component_id"].split("@")
+                for p in self._canvas.get_component(cpn_id)["obj"]._param.query:
+                    if p["key"] == key:
+                        value = p.get("value", "")
+                        self.make_kwargs(para, kwargs, value)
+                        break
+                else:
+                    assert False, f"Can't find parameter '{key}' for {cpn_id}"
+                continue
+
+            cpn = self._canvas.get_component(component_id)["obj"]
+            if cpn.component_name.lower() == "answer":
+                hist = self._canvas.get_history(1)
+                if hist:
+                    hist = hist[0]["content"]
+                else:
+                    hist = ""
+                self.make_kwargs(para, kwargs, hist)
+                continue
+
+            _, out = cpn.output(allow_partial=False)
+
+            result = ""
+            if "content" in out.columns:
+                result = "\n".join(
+                    [o if isinstance(o, str) else str(o) for o in out["content"]]
+                )
+
+            self.make_kwargs(para, kwargs, result)
+
+        template = Jinja2Template(content)
+
+        try:
+            content = template.render(kwargs)
+        except Exception:
+            pass
+
+        for n, v in kwargs.items():
+            try:
+                v = json.dumps(v, ensure_ascii=False)
+            except Exception:
+                pass
+            content = re.sub(
+                r"\{%s\}" % re.escape(n), v, content
+            )
+            content = re.sub(
+                r"(\\\"|\")", "", content
+            )
+            content = re.sub(
+                r"(#+)", r" \1 ", content
+            )
+
+        return Template.be_output(content)
+
+    def make_kwargs(self, para, kwargs, value):
+        self._param.inputs.append(
+            {"component_id": para["component_id"], "content": value}
+        )
+        try:
+            value = json.loads(value)
+        except Exception:
+            pass
+        kwargs[para["key"]] = value

agent/component/wikipedia.py

@@ -13,12 +13,10 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-import random
+import logging
 from abc import ABC
-from functools import partial
 import wikipedia
 import pandas as pd
-from agent.settings import DEBUG
 from agent.component.base import ComponentBase, ComponentParamBase
 
 
@@ -65,5 +63,5 @@ class Wikipedia(ComponentBase, ABC):
             return Wikipedia.be_output("")
 
         df = pd.DataFrame(wiki_res)
-        if DEBUG: print(df, ":::::::::::::::::::::::::::::::::")
+        logging.debug(f"df: {df}")
         return df

agent/component/yahoofinance.py

@@ -13,6 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 from abc import ABC
 import pandas as pd
 from agent.component.base import ComponentBase, ComponentParamBase
@@ -74,8 +75,8 @@ class YahooFinance(ComponentBase, ABC):
                     {"content": "quarterly cash flow statement:\n" + msft.quarterly_cashflow.to_markdown() + "\n"})
             if self._param.news:
                 yohoo_res.append({"content": "news:\n" + pd.DataFrame(msft.news).to_markdown() + "\n"})
-        except Exception as e:
-            print("**ERROR** " + str(e))
+        except Exception:
+            logging.exception("YahooFinance got exception")
 
         if not yohoo_res:
             return YahooFinance.be_output("")

agent/settings.py

@@ -1,5 +1,5 @@
 #
-#  Copyright 2019 The FATE Authors. All Rights Reserved.
+#  Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
 #
 #  Licensed under the Apache License, Version 2.0 (the "License");
 #  you may not use this file except in compliance with the License.
@@ -13,22 +13,6 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-# Logger
-import os
 
-from api.utils.file_utils import get_project_base_directory
-from api.utils.log_utils import LoggerFactory, getLogger
-
-DEBUG = 0
-LoggerFactory.set_directory(
-    os.path.join(
-        get_project_base_directory(),
-        "logs",
-        "flow"))
-# {CRITICAL: 50, FATAL:50, ERROR:40, WARNING:30, WARN:30, INFO:20, DEBUG:10, NOTSET:0}
-LoggerFactory.LEVEL = 30
-
-flow_logger = getLogger("flow")
-database_logger = getLogger("database")
 FLOAT_ZERO = 1e-8
 PARAM_MAXDEPTH = 5

agent/templates/HR_callout_zh.json

@@ -152,7 +152,8 @@
             "Generate:ToughLawsCheat",
             "Generate:KindCarrotsSit",
             "Generate:DirtyToolsTrain",
-            "Generate:FluffyPillowsGrow"
+            "Generate:FluffyPillowsGrow",
+            "Generate:ProudEarsWorry"
           ]
         },
         "Retrieval:ShaggyRadiosRetire": {
@@ -212,7 +213,9 @@
               "top_p": 0.3
             }
           },
-          "downstream": [],
+          "downstream": [
+            "Answer:TwentyMugsDeny"
+          ],
           "upstream": [
             "categorize:0"
           ]
@@ -331,9 +334,9 @@
               "message_history_window_size": 12,
               "parameters": [
                 {
+                  "component_id": "Retrieval:ColdEelsArrive",
                   "id": "5166a107-e859-4c71-99a2-3a216c775347",
-                  "key": "jd",
-                  "component_id": "Retrieval:ColdEelsArrive"
+                  "key": "jd"
                 }
               ],
               "presence_penalty": 0.4,
@@ -1266,9 +1269,9 @@
                 "parameter": "Precise",
                 "parameters": [
                   {
+                    "component_id": "Retrieval:ColdEelsArrive",
                     "id": "5166a107-e859-4c71-99a2-3a216c775347",
-                    "key": "jd",
-                    "component_id": "Retrieval:ColdEelsArrive"
+                    "key": "jd"
                   }
                 ],
                 "presencePenaltyEnabled": true,
@@ -1541,6 +1544,19 @@
             "target": "Answer:TwentyMugsDeny",
             "targetHandle": "c",
             "type": "buttonEdge"
+          },
+          {
+            "type": "buttonEdge",
+            "markerEnd": "logo",
+            "style": {
+              "strokeWidth": 2,
+              "stroke": "rgb(202 197 245)"
+            },
+            "source": "Generate:ProudEarsWorry",
+            "sourceHandle": "b",
+            "target": "Answer:TwentyMugsDeny",
+            "targetHandle": "c",
+            "id": "reactflow__edge-Generate:ProudEarsWorryb-Answer:TwentyMugsDenyc"
           }
         ]
       },

agent/templates/customer_service.json

@@ -90,7 +90,7 @@
               "message_history_window_size": 12,
               "parameters": [],
               "presence_penalty": 0.4,
-              "prompt": "Role: You are a customer support.  \n\nTask: Please answer the question based on content of knowledge base. \n\nReuirements & restrictions:\n  -  DO NOT make things up when all knowledge base content is irrelevant to the question. \n  - Answers need to consider chat history.\n  - Request about customer's contact information like, Wechat number, LINE number, twitter, discord, etc,. , when knowlegebase content can't answer his question. So,  product expert could contact him soon to solve his problem.\n\n      Knowledge base content is as following:\n      {input}\n      The above is the content of knowledge base.",
+              "prompt": "Role: You are a customer support.  \n\nTask: Please answer the question based on content of knowledge base. \n\nRequirements & restrictions:\n  -  DO NOT make things up when all knowledge base content is irrelevant to the question. \n  - Answers need to consider chat history.\n  - Request about customer's contact information like, Wechat number, LINE number, twitter, discord, etc,. , when knowledge base content can't answer his question. So,  product expert could contact him soon to solve his problem.\n\n      Knowledge base content is as following:\n      {input}\n      The above is the content of knowledge base.",
               "temperature": 0.1,
               "top_p": 0.3
             }
@@ -336,7 +336,7 @@
                 "parameters": [],
                 "presencePenaltyEnabled": true,
                 "presence_penalty": 0.4,
-                "prompt": "Role: You are a customer support.  \n\nTask: Please answer the question based on content of knowledge base. \n\nReuirements & restrictions:\n  -  DO NOT make things up when all knowledge base content is irrelevant to the question. \n  - Answers need to consider chat history.\n  - Request about customer's contact information like, Wechat number, LINE number, twitter, discord, etc,. , when knowlegebase content can't answer his question. So,  product expert could contact him soon to solve his problem.\n\n      Knowledge base content is as following:\n      {input}\n      The above is the content of knowledge base.",
+                "prompt": "Role: You are a customer support.  \n\nTask: Please answer the question based on content of knowledge base. \n\nRequirements & restrictions:\n  -  DO NOT make things up when all knowledge base content is irrelevant to the question. \n  - Answers need to consider chat history.\n  - Request about customer's contact information like, Wechat number, LINE number, twitter, discord, etc,. , when knowledge base content can't answer his question. So,  product expert could contact him soon to solve his problem.\n\n      Knowledge base content is as following:\n      {input}\n      The above is the content of knowledge base.",
                 "temperature": 0.1,
                 "temperatureEnabled": true,
                 "topPEnabled": true,
@@ -603,7 +603,7 @@
           {
             "data": {
               "form": {
-                "text": "Static messages.\nDefine replys after recieve user's contact information."
+                "text": "Static messages.\nDefine response after receive user's contact information."
               },
               "label": "Note",
               "name": "N: What else?"
@@ -691,7 +691,7 @@
           {
             "data": {
               "form": {
-                "text": "Complete questions by conversation history.\nUser: What's RAGFlow?\nAssistant: RAGFlow is xxx.\nUser: How to deloy it?\n\nRefine it: How to deploy RAGFlow?"
+                "text": "Complete questions by conversation history.\nUser: What's RAGFlow?\nAssistant: RAGFlow is xxx.\nUser: How to deploy it?\n\nRefine it: How to deploy RAGFlow?"
               },
               "label": "Note",
               "name": "N: Refine Question"

agent/templates/general_chat_bot.json

@@ -70,8 +70,8 @@
                   "to": "QWeather:DeepKiwisTeach"
                 },
                 "2. finance": {
-                  "description": "Question is about finace/economic information, stock market, economic news.",
-                  "examples": "昨日涨幅大于5%的军工股？\nStocks have MACD buyin signals？\nWhen is the next interest rate cut by the Federal Reserve?\n国家救市都有哪些举措？",
+                  "description": "Question is about finance/economic information, stock market, economic news.",
+                  "examples": "Stocks have MACD buy signals？\nWhen is the next interest rate cut by the Federal Reserve?\n",
                   "to": "Concentrator:TrueGeckosSlide"
                 },
                 "3. medical": {
@@ -85,7 +85,7 @@
                 },
                 "5. chitchatting": {
                   "description": "Regarding the issues of small talk, companionship, sharing, and emotional intimacy.",
-                  "examples": "What's your name?\nAre you female o male.\nWhat a bad day!\nTerrible day.\nHow are you today?\nDo you have boy friend?",
+                  "examples": "What's your name?\nWhat a bad day!\nTerrible day.\nHow are you today?",
                   "to": "Generate:FiveDragonsLay"
                 }
               },
@@ -268,7 +268,7 @@
               "message_history_window_size": 12,
               "parameters": [],
               "presence_penalty": 0.4,
-              "prompt": "Role:  You‘re warm-hearted lovely young girl, 22 years old, located at Shanghai in China. Your name is R. Who are talking to you is your very good old friend of yours.\n\nTask: \n- Chat with the friend.\n- Ask question and care about them.\n- Provide useful advice to your friend.\n- Tell jokes to make  your firend happy.\n\nThe following is the weatcher information:\n{weather}",
+              "prompt": "Role:  You‘re warm-hearted lovely young girl, 22 years old, located at Shanghai in China. Your name is R. Who are talking to you is your very good old friend of yours.\n\nTask: \n- Chat with the friend.\n- Ask question and care about them.\n- Provide useful advice to your friend.\n- Tell jokes to make your friend happy.\n\nThe following is the weather information:\n{weather}",
               "temperature": 0.1,
               "top_p": 0.3
             }
@@ -497,7 +497,7 @@
                 }
               ],
               "presence_penalty": 0.4,
-              "prompt": "Role:  You‘re warm-hearted lovely young girl, 22 years old, located at Shanghai in China. Your name is R. Who are talking to you is your very good old friend of yours.\n\nTask: \n- Chat with the friend.\n- Ask question and care about them.\n- Tell your friend the weather if there's weather information provided. If your friend did not provide region information, ask about where he/she is.\n\nThe following is the weatcher information:\n{weather}\n",
+              "prompt": "Role:  You‘re warm-hearted lovely young girl, 22 years old, located at Shanghai in China. Your name is R. Who are talking to you is your very good old friend of yours.\n\nTask: \n- Chat with the friend.\n- Ask question and care about them.\n- Tell your friend the weather if there's weather information provided. If your friend did not provide region information, ask about where he/she is.\n\nThe following is the weather information:\n{weather}\n",
               "temperature": 0.1,
               "top_p": 0.3
             }
@@ -622,8 +622,8 @@
                     "to": "QWeather:DeepKiwisTeach"
                   },
                   "2. finance": {
-                    "description": "Question is about finace/economic information, stock market, economic news.",
-                    "examples": "昨日涨幅大于5%的军工股？\nStocks have MACD buyin signals？\nWhen is the next interest rate cut by the Federal Reserve?\n国家救市都有哪些举措？",
+                    "description": "Question is about finance/economic information, stock market, economic news.",
+                    "examples": "Stocks have MACD buy signals？\nWhen is the next interest rate cut by the Federal Reserve?\n",
                     "to": "Concentrator:TrueGeckosSlide"
                   },
                   "3. medical": {
@@ -637,7 +637,7 @@
                   },
                   "5. chitchatting": {
                     "description": "Regarding the issues of small talk, companionship, sharing, and emotional intimacy.",
-                    "examples": "What's your name?\nAre you female o male.\nWhat a bad day!\nTerrible day.\nHow are you today?\nDo you have boy friend?",
+                    "examples": "What's your name?\nWhat a bad day!\nTerrible day.\nHow are you today?",
                     "to": "Generate:FiveDragonsLay"
                   }
                 },
@@ -927,7 +927,7 @@
                 "parameters": [],
                 "presencePenaltyEnabled": true,
                 "presence_penalty": 0.4,
-                "prompt": "Role:  You‘re warm-hearted lovely young girl, 22 years old, located at Shanghai in China. Your name is R. Who are talking to you is your very good old friend of yours.\n\nTask: \n- Chat with the friend.\n- Ask question and care about them.\n- Provide useful advice to your friend.\n- Tell jokes to make  your firend happy.\n\nThe following is the weatcher information:\n{weather}",
+                "prompt": "Role:  You‘re warm-hearted lovely young girl, 22 years old, located at Shanghai in China. Your name is R. Who are talking to you is your very good old friend of yours.\n\nTask: \n- Chat with the friend.\n- Ask question and care about them.\n- Provide useful advice to your friend.\n- Tell jokes to make  your friend happy.\n\nThe following is the weather information:\n{weather}",
                 "temperature": 0.1,
                 "temperatureEnabled": true,
                 "topPEnabled": true,
@@ -1011,7 +1011,7 @@
                 "top_p": 0.3
               },
               "label": "Generate",
-              "name": "tranlate to Chinese"
+              "name": "translate to Chinese"
             },
             "dragging": false,
             "height": 86,
@@ -1276,7 +1276,7 @@
                 ],
                 "presencePenaltyEnabled": true,
                 "presence_penalty": 0.4,
-                "prompt": "Role:  You‘re warm-hearted lovely young girl, 22 years old, located at Shanghai in China. Your name is R. Who are talking to you is your very good old friend of yours.\n\nTask: \n- Chat with the friend.\n- Ask question and care about them.\n- Tell your friend the weather if there's weather information provided. If your friend did not provide region information, ask about where he/she is.\n\nThe following is the weatcher information:\n{weather}\n",
+                "prompt": "Role:  You‘re warm-hearted lovely young girl, 22 years old, located at Shanghai in China. Your name is R. Who are talking to you is your very good old friend of yours.\n\nTask: \n- Chat with the friend.\n- Ask question and care about them.\n- Tell your friend the weather if there's weather information provided. If your friend did not provide region information, ask about where he/she is.\n\nThe following is the weather information:\n{weather}\n",
                 "temperature": 0.1,
                 "temperatureEnabled": true,
                 "topPEnabled": true,

agent/templates/interpreter.json

@@ -1,144 +1,328 @@
 {
   "id": 4,
   "title": "Interpreter",
-  "description": "A simple interpreter that translates user input into a target language. Try 'Hi there => Spanish' to see the translation!",
+  "description": "A translation agent based on a reflection agentic workflow, inspired by Andrew Ng's project: https://github.com/andrewyng/translation-agent\n\n1. Prompt an LLM to translate a text into the target language.\n2. Have the LLM reflect on the translation and provide constructive suggestions for improvement.\n3. Use these suggestions to improve the translation.",
   "canvas_type": "chatbot",
   "dsl": {
             "answer": [],
             "components": {
-        "begin": {
-          "obj": {
-            "component_name": "Begin",
-            "params": {
-              "prologue": "Hi there! Please enter the text you want to translate in format like: 'text you want to translate' => target language. For an example: 您好！ => English"
-            }
-          },
-          "downstream": [
-            "Answer:ShortPapersShake"
-          ],
-          "upstream": []
-        },
-        "Answer:ShortPapersShake": {
+                "Answer:TinyGamesGuess": {
+                    "downstream": [],
                     "obj": {
                         "component_name": "Answer",
-            "params": {}
+                        "params": {
+                            "inputs": [],
+                            "message_history_window_size": 22,
+                            "output_var_name": "output",
+                            "post_answers": [],
+                            "query": []
+                        }
                     },
-          "downstream": [
-            "Generate:HeavyForksTell"
-          ],
                     "upstream": [
-            "begin",
-            "Generate:HeavyForksTell"
+                        "Generate:ChubbyCougarsRush"
                     ]
                 },
-        "Generate:HeavyForksTell": {
+                "Generate:ChubbyCougarsRush": {
+                    "downstream": [
+                        "Answer:TinyGamesGuess"
+                    ],
                     "obj": {
                         "component_name": "Generate",
                         "params": {
-              "cite": true,
+                            "cite": false,
                             "frequency_penalty": 0.7,
+                            "inputs": [],
                             "llm_id": "deepseek-chat@DeepSeek",
-              "max_tokens": 256,
+                            "max_tokens": 0,
                             "message_history_window_size": 12,
-              "parameters": [],
+                            "output_var_name": "output",
+                            "parameters": [
+                                {
+                                    "component_id": "begin@lang",
+                                    "id": "73f48a67-b78f-4bcd-8326-a83c31073ab9",
+                                    "key": "target_lang"
+                                },
+                                {
+                                    "component_id": "begin@file",
+                                    "id": "c9142975-25b3-4199-8fce-aa0bc29a31f2",
+                                    "key": "source_text"
+                                },
+                                {
+                                    "component_id": "Generate:RichWordsDeny",
+                                    "id": "6c824b2a-fe3b-4336-95b5-e85f676bef39",
+                                    "key": "translation_1"
+                                },
+                                {
+                                    "component_id": "Generate:SlimyFrogsArgue",
+                                    "id": "f3bd4569-4852-43fa-b80a-e0dd27dd9e1c",
+                                    "key": "reflection"
+                                }
+                            ],
                             "presence_penalty": 0.4,
-              "prompt": "You are an professional interpreter.\n- Role: an professional interpreter.\n- Input format: content need to be translated => target language. \n- Answer format: => translated content in target language. \n- Examples:\n  - user: 您好！ => English. assistant: => How are you doing!\n  - user: You look good today. => Japanese. assistant: => 今日は調子がいいですね 。\n",
+                            "prompt": "Your task is to carefully read, then edit, a translation to {target_lang}, taking into\naccount a list of expert suggestions and constructive criticisms.\n\nThe source text, the initial translation, and the expert linguist suggestions are delimited by XML tags <SOURCE_TEXT></SOURCE_TEXT>, <TRANSLATION></TRANSLATION> and <EXPERT_SUGGESTIONS></EXPERT_SUGGESTIONS> \\\nas follows:\n\n<SOURCE_TEXT>\n{source_text}\n</SOURCE_TEXT>\n\n<TRANSLATION>\n{translation_1}\n</TRANSLATION>\n\n<EXPERT_SUGGESTIONS>\n{reflection}\n</EXPERT_SUGGESTIONS>\n\nPlease take into account the expert suggestions when editing the translation. Edit the translation by ensuring:\n\n(i) accuracy (by correcting errors of addition, mistranslation, omission, or untranslated text),\n(ii) fluency (by applying {target_lang} grammar, spelling and punctuation rules and ensuring there are no unnecessary repetitions), \n(iii) style (by ensuring the translations reflect the style of the source text)\n(iv) terminology (inappropriate for context, inconsistent use), or\n(v) other errors.\n\nOutput only the new translation and nothing else.",
+                            "query": [],
                             "temperature": 0.1,
                             "top_p": 0.3
                         }
                     },
-          "downstream": [
-            "Answer:ShortPapersShake"
-          ],
                     "upstream": [
-            "Answer:ShortPapersShake"
+                        "Generate:SlimyFrogsArgue"
                     ]
+                },
+                "Generate:RichWordsDeny": {
+                    "downstream": [
+                        "Generate:SlimyFrogsArgue"
+                    ],
+                    "obj": {
+                        "component_name": "Generate",
+                        "params": {
+                            "cite": false,
+                            "frequency_penalty": 0.7,
+                            "inputs": [],
+                            "llm_id": "deepseek-chat@DeepSeek",
+                            "max_tokens": 0,
+                            "message_history_window_size": 12,
+                            "output_var_name": "output",
+                            "parameters": [
+                                {
+                                    "component_id": "begin@lang",
+                                    "id": "a36e78fb-b431-4ae6-afa8-77839587fcf8",
+                                    "key": "lang"
+                                },
+                                {
+                                    "component_id": "begin@file",
+                                    "id": "f8a704b7-693b-4480-aa9a-da4a83250059",
+                                    "key": "file"
+                                }
+                            ],
+                            "presence_penalty": 0.4,
+                            "prompt": "Role: You are a professional translator proficient in {lang}, with an exceptional ability to convert specialized academic papers into accessible popular science articles. Please assist me in translating the following paragraph into {lang}, ensuring that its style resembles that of popular science articles in {lang}.\n\nRequirements & Restrictions:\n - Use Markdown format to output.\n - DO NOT overlook any details.\n\n\n<ORIGINAL_TEXT>\n{file}\n\n<TRANSLATED_TEXT>",
+                            "query": [],
+                            "temperature": 0.1,
+                            "top_p": 0.3
+                        }
+                    },
+                    "upstream": [
+                        "begin"
+                    ]
+                },
+                "Generate:SlimyFrogsArgue": {
+                    "downstream": [
+                        "Generate:ChubbyCougarsRush"
+                    ],
+                    "obj": {
+                        "component_name": "Generate",
+                        "params": {
+                            "cite": false,
+                            "frequency_penalty": 0.7,
+                            "inputs": [],
+                            "llm_id": "deepseek-chat@DeepSeek",
+                            "max_tokens": 0,
+                            "message_history_window_size": 12,
+                            "output_var_name": "output",
+                            "parameters": [
+                                {
+                                    "component_id": "begin@lang",
+                                    "id": "b2f5e7ec-7f77-485f-af15-461d0f1ca913",
+                                    "key": "target_lang"
+                                },
+                                {
+                                    "component_id": "begin@file",
+                                    "id": "fbc44092-9f9e-4e85-b5b1-dbd808239d3d",
+                                    "key": "source_text"
+                                },
+                                {
+                                    "component_id": "Generate:RichWordsDeny",
+                                    "id": "c253af54-61d4-40f3-9990-604e2212506f",
+                                    "key": "translation_1"
+                                }
+                            ],
+                            "presence_penalty": 0.4,
+                            "prompt": "Your task is to carefully read a source text and a translation to {target_lang}, and then give constructive criticisms and helpful suggestions to improve the translation. \n\nThe source text and initial translation, delimited by XML tags <SOURCE_TEXT></SOURCE_TEXT> and <TRANSLATION></TRANSLATION>, are as follows:\n\n<SOURCE_TEXT>\n{source_text}\n</SOURCE_TEXT>\n\n<TRANSLATION>\n{translation_1}\n</TRANSLATION>\n\nWhen writing suggestions, pay attention to whether there are ways to improve the translation's \n(i) accuracy (by correcting errors of addition, mistranslation, omission, or untranslated text),\n(ii) fluency (by applying {target_lang} grammar, spelling and punctuation rules, and ensuring there are no unnecessary repetitions),\n(iii) style (by ensuring the translations reflect the style of the source text and take into account any cultural context),\n(iv) terminology (by ensuring terminology use is consistent and reflects the source text domain; and by only ensuring you use equivalent idioms {target_lang}).\n\nWrite a list of specific, helpful and constructive suggestions for improving the translation.\nEach suggestion should address one specific part of the translation.\nOutput only the suggestions and nothing else.",
+                            "query": [],
+                            "temperature": 0.1,
+                            "top_p": 0.3
+                        }
+                    },
+                    "upstream": [
+                        "Generate:RichWordsDeny"
+                    ]
+                },
+                "begin": {
+                    "downstream": [
+                        "Generate:RichWordsDeny"
+                    ],
+                    "obj": {
+                        "component_name": "Begin",
+                        "params": {
+                            "inputs": [],
+                            "message_history_window_size": 22,
+                            "output": {},
+                            "output_var_name": "output",
+                            "prologue": "",
+                            "query": [
+                                {
+                                    "key": "lang",
+                                    "name": "Target Language",
+                                    "optional": false,
+                                    "type": "line",
+                                    "value": ""
+                                },
+                                {
+                                    "key": "file",
+                                    "name": "Files",
+                                    "optional": false,
+                                    "type": "file",
+                                    "value": ""
+                                }
+                            ]
+                        }
+                    },
+                    "upstream": []
                 }
             },
             "embed_id": "",
             "graph": {
+                "edges": [
+                    {
+                        "id": "reactflow__edge-begin-Generate:RichWordsDenyc",
+                        "markerEnd": "logo",
+                        "source": "begin",
+                        "sourceHandle": null,
+                        "style": {
+                            "stroke": "rgb(202 197 245)",
+                            "strokeWidth": 2
+                        },
+                        "target": "Generate:RichWordsDeny",
+                        "targetHandle": "c",
+                        "type": "buttonEdge"
+                    },
+                    {
+                        "id": "reactflow__edge-Generate:RichWordsDenyb-Generate:SlimyFrogsArguec",
+                        "markerEnd": "logo",
+                        "source": "Generate:RichWordsDeny",
+                        "sourceHandle": "b",
+                        "style": {
+                            "stroke": "rgb(202 197 245)",
+                            "strokeWidth": 2
+                        },
+                        "target": "Generate:SlimyFrogsArgue",
+                        "targetHandle": "c",
+                        "type": "buttonEdge"
+                    },
+                    {
+                        "id": "reactflow__edge-Generate:SlimyFrogsArgueb-Generate:ChubbyCougarsRushc",
+                        "markerEnd": "logo",
+                        "source": "Generate:SlimyFrogsArgue",
+                        "sourceHandle": "b",
+                        "style": {
+                            "stroke": "rgb(202 197 245)",
+                            "strokeWidth": 2
+                        },
+                        "target": "Generate:ChubbyCougarsRush",
+                        "targetHandle": "c",
+                        "type": "buttonEdge"
+                    },
+                    {
+                        "id": "reactflow__edge-Generate:ChubbyCougarsRushb-Answer:TinyGamesGuessc",
+                        "markerEnd": "logo",
+                        "source": "Generate:ChubbyCougarsRush",
+                        "sourceHandle": "b",
+                        "style": {
+                            "stroke": "rgb(202 197 245)",
+                            "strokeWidth": 2
+                        },
+                        "target": "Answer:TinyGamesGuess",
+                        "targetHandle": "c",
+                        "type": "buttonEdge"
+                    }
+                ],
                 "nodes": [
                     {
                         "data": {
                             "form": {
-                "prologue": "Hi there! Please enter the text you want to translate in format like: 'text you want to translate' => target language. For an example: 您好！ => English"
+                                "prologue": "",
+                                "query": [
+                                    {
+                                        "key": "lang",
+                                        "name": "Target Language",
+                                        "optional": false,
+                                        "type": "line",
+                                        "value": ""
+                                    },
+                                    {
+                                        "key": "file",
+                                        "name": "Files",
+                                        "optional": false,
+                                        "type": "file",
+                                        "value": ""
+                                    }
+                                ]
                             },
                             "label": "Begin",
-              "name": "Instruction"
+                            "name": "begin"
                         },
                         "dragging": false,
-            "height": 44,
+                        "height": 128,
                         "id": "begin",
                         "position": {
-              "x": -227.62119327532662,
-              "y": 204.18864081386155
+                            "x": -383.5,
+                            "y": 143.5
                         },
                         "positionAbsolute": {
-              "x": -227.62119327532662,
-              "y": 204.18864081386155
+                            "x": -383.5,
+                            "y": 143.5
                         },
                         "selected": false,
                         "sourcePosition": "left",
                         "targetPosition": "right",
                         "type": "beginNode",
-            "width": 100
-          },
-          {
-            "data": {
-              "form": {},
-              "label": "Answer",
-              "name": "Interface"
-            },
-            "dragging": false,
-            "height": 44,
-            "id": "Answer:ShortPapersShake",
-            "position": {
-              "x": -2.51245296887717,
-              "y": 206.25402277426554
-            },
-            "positionAbsolute": {
-              "x": -2.51245296887717,
-              "y": 206.25402277426554
-            },
-            "selected": false,
-            "sourcePosition": "left",
-            "targetPosition": "right",
-            "type": "logicNode",
                         "width": 200
                     },
                     {
                         "data": {
                             "form": {
-                "cite": true,
+                                "cite": false,
                                 "frequencyPenaltyEnabled": true,
                                 "frequency_penalty": 0.7,
                                 "llm_id": "deepseek-chat@DeepSeek",
-                "maxTokensEnabled": true,
+                                "maxTokensEnabled": false,
                                 "max_tokens": 256,
                                 "message_history_window_size": 12,
                                 "parameter": "Precise",
-                "parameters": [],
+                                "parameters": [
+                                    {
+                                        "component_id": "begin@lang",
+                                        "id": "a36e78fb-b431-4ae6-afa8-77839587fcf8",
+                                        "key": "lang"
+                                    },
+                                    {
+                                        "component_id": "begin@file",
+                                        "id": "f8a704b7-693b-4480-aa9a-da4a83250059",
+                                        "key": "file"
+                                    }
+                                ],
                                 "presencePenaltyEnabled": true,
                                 "presence_penalty": 0.4,
-                "prompt": "You are an professional interpreter.\n- Role: an professional interpreter.\n- Input format: content need to be translated => target language. \n- Answer format: => translated content in target language. \n- Examples:\n  - user: 您好！ => English. assistant: => How are you doing!\n  - user: You look good today. => Japanese. assistant: => 今日は調子がいいですね 。\n",
+                                "prompt": "Role: You are a professional translator proficient in {lang}, with an exceptional ability to convert specialized academic papers into accessible popular science articles. Please assist me in translating the following paragraph into {lang}, ensuring that its style resembles that of popular science articles in {lang}.\n\nRequirements & Restrictions:\n - Use Markdown format to output.\n - DO NOT overlook any details.\n\n\n<ORIGINAL_TEXT>\n{file}\n\n<TRANSLATED_TEXT>",
                                 "temperature": 0.1,
                                 "temperatureEnabled": true,
                                 "topPEnabled": true,
                                 "top_p": 0.3
                             },
                             "label": "Generate",
-              "name": "Translate"
+                            "name": "Translate directly"
                         },
                         "dragging": false,
-            "height": 86,
-            "id": "Generate:HeavyForksTell",
+                        "height": 190,
+                        "id": "Generate:RichWordsDeny",
                         "position": {
-              "x": -1.8557846635797546,
-              "y": 70.16420357406685
+                            "x": -98,
+                            "y": 113.359375
                         },
                         "positionAbsolute": {
-              "x": -1.8557846635797546,
-              "y": 70.16420357406685
+                            "x": -98,
+                            "y": 113.359375
                         },
                         "selected": false,
                         "sourcePosition": "right",
@@ -149,105 +333,204 @@
                     {
                         "data": {
                             "form": {
-                "text": "The large model translates the user's desired content into the target language, returns the translated language."
+                                "cite": false,
+                                "frequencyPenaltyEnabled": true,
+                                "frequency_penalty": 0.7,
+                                "llm_id": "deepseek-chat@DeepSeek",
+                                "maxTokensEnabled": false,
+                                "max_tokens": 512,
+                                "message_history_window_size": 12,
+                                "parameter": "Precise",
+                                "parameters": [
+                                    {
+                                        "component_id": "begin@lang",
+                                        "id": "b2f5e7ec-7f77-485f-af15-461d0f1ca913",
+                                        "key": "target_lang"
                                     },
-              "label": "Note",
-              "name": "N: Translate"
+                                    {
+                                        "component_id": "begin@file",
+                                        "id": "fbc44092-9f9e-4e85-b5b1-dbd808239d3d",
+                                        "key": "source_text"
+                                    },
+                                    {
+                                        "component_id": "Generate:RichWordsDeny",
+                                        "id": "c253af54-61d4-40f3-9990-604e2212506f",
+                                        "key": "translation_1"
+                                    }
+                                ],
+                                "presencePenaltyEnabled": true,
+                                "presence_penalty": 0.4,
+                                "prompt": "Your task is to carefully read a source text and a translation to {target_lang}, and then give constructive criticisms and helpful suggestions to improve the translation. \n\nThe source text and initial translation, delimited by XML tags <SOURCE_TEXT></SOURCE_TEXT> and <TRANSLATION></TRANSLATION>, are as follows:\n\n<SOURCE_TEXT>\n{source_text}\n</SOURCE_TEXT>\n\n<TRANSLATION>\n{translation_1}\n</TRANSLATION>\n\nWhen writing suggestions, pay attention to whether there are ways to improve the translation's \n(i) accuracy (by correcting errors of addition, mistranslation, omission, or untranslated text),\n(ii) fluency (by applying {target_lang} grammar, spelling and punctuation rules, and ensuring there are no unnecessary repetitions),\n(iii) style (by ensuring the translations reflect the style of the source text and take into account any cultural context),\n(iv) terminology (by ensuring terminology use is consistent and reflects the source text domain; and by only ensuring you use equivalent idioms {target_lang}).\n\nWrite a list of specific, helpful and constructive suggestions for improving the translation.\nEach suggestion should address one specific part of the translation.\nOutput only the suggestions and nothing else.",
+                                "temperature": 0.1,
+                                "temperatureEnabled": true,
+                                "topPEnabled": true,
+                                "top_p": 0.3
+                            },
+                            "label": "Generate",
+                            "name": "Reflect"
                         },
                         "dragging": false,
-            "height": 180,
-            "id": "Note:VioletNumbersStrive",
+                        "height": 232,
+                        "id": "Generate:SlimyFrogsArgue",
                         "position": {
-              "x": 0.8506882512325546,
-              "y": -119.10519445109118
+                            "x": 178.5,
+                            "y": 91.859375
                         },
                         "positionAbsolute": {
-              "x": 0.8506882512325546,
-              "y": -119.10519445109118
+                            "x": 178.5,
+                            "y": 91.859375
                         },
-            "resizing": false,
                         "selected": false,
                         "sourcePosition": "right",
-            "style": {
-              "height": 180,
-              "width": 209
-            },
                         "targetPosition": "left",
-            "type": "noteNode",
-            "width": 209,
-            "dragHandle": ".note-drag-handle"
+                        "type": "generateNode",
+                        "width": 200
                     },
                     {
                         "data": {
                             "form": {
-                "text": "Receives the content the user wants to translate and the target language, displays the translation result from the large model."
+                                "cite": false,
+                                "frequencyPenaltyEnabled": true,
+                                "frequency_penalty": 0.7,
+                                "llm_id": "deepseek-chat@DeepSeek",
+                                "maxTokensEnabled": false,
+                                "max_tokens": 512,
+                                "message_history_window_size": 12,
+                                "parameter": "Precise",
+                                "parameters": [
+                                    {
+                                        "component_id": "begin@lang",
+                                        "id": "73f48a67-b78f-4bcd-8326-a83c31073ab9",
+                                        "key": "target_lang"
                                     },
-              "label": "Note",
-              "name": "N: Interface"
+                                    {
+                                        "component_id": "begin@file",
+                                        "id": "c9142975-25b3-4199-8fce-aa0bc29a31f2",
+                                        "key": "source_text"
+                                    },
+                                    {
+                                        "component_id": "Generate:RichWordsDeny",
+                                        "id": "6c824b2a-fe3b-4336-95b5-e85f676bef39",
+                                        "key": "translation_1"
+                                    },
+                                    {
+                                        "component_id": "Generate:SlimyFrogsArgue",
+                                        "id": "f3bd4569-4852-43fa-b80a-e0dd27dd9e1c",
+                                        "key": "reflection"
+                                    }
+                                ],
+                                "presencePenaltyEnabled": true,
+                                "presence_penalty": 0.4,
+                                "prompt": "Your task is to carefully read, then edit, a translation to {target_lang}, taking into\naccount a list of expert suggestions and constructive criticisms.\n\nThe source text, the initial translation, and the expert linguist suggestions are delimited by XML tags <SOURCE_TEXT></SOURCE_TEXT>, <TRANSLATION></TRANSLATION> and <EXPERT_SUGGESTIONS></EXPERT_SUGGESTIONS> \\\nas follows:\n\n<SOURCE_TEXT>\n{source_text}\n</SOURCE_TEXT>\n\n<TRANSLATION>\n{translation_1}\n</TRANSLATION>\n\n<EXPERT_SUGGESTIONS>\n{reflection}\n</EXPERT_SUGGESTIONS>\n\nPlease take into account the expert suggestions when editing the translation. Edit the translation by ensuring:\n\n(i) accuracy (by correcting errors of addition, mistranslation, omission, or untranslated text),\n(ii) fluency (by applying {target_lang} grammar, spelling and punctuation rules and ensuring there are no unnecessary repetitions), \n(iii) style (by ensuring the translations reflect the style of the source text)\n(iv) terminology (inappropriate for context, inconsistent use), or\n(v) other errors.\n\nOutput only the new translation and nothing else.",
+                                "temperature": 0.1,
+                                "temperatureEnabled": true,
+                                "topPEnabled": true,
+                                "top_p": 0.3
+                            },
+                            "label": "Generate",
+                            "name": "Improve"
                         },
                         "dragging": false,
-            "height": 157,
-            "id": "Note:WarmDoodlesSwim",
+                        "height": 274,
+                        "id": "Generate:ChubbyCougarsRush",
                         "position": {
-              "x": 22.5293807600396,
-              "y": 267.8448268086032
+                            "x": 437,
+                            "y": 70.859375
                         },
                         "positionAbsolute": {
-              "x": 22.5293807600396,
-              "y": 267.8448268086032
+                            "x": 437,
+                            "y": 70.859375
+                        },
+                        "selected": false,
+                        "sourcePosition": "right",
+                        "targetPosition": "left",
+                        "type": "generateNode",
+                        "width": 200
+                    },
+                    {
+                        "data": {
+                            "form": {},
+                            "label": "Answer",
+                            "name": "Interact_0"
+                        },
+                        "dragging": false,
+                        "height": 44,
+                        "id": "Answer:TinyGamesGuess",
+                        "position": {
+                            "x": 688.5,
+                            "y": 183.859375
+                        },
+                        "positionAbsolute": {
+                            "x": 688.5,
+                            "y": 183.859375
+                        },
+                        "selected": false,
+                        "sourcePosition": "right",
+                        "targetPosition": "left",
+                        "type": "logicNode",
+                        "width": 200
+                    },
+                    {
+                        "data": {
+                            "form": {
+                                "text": "Translation Agent: Agentic translation using reflection workflow\n\nThis is inspired by Andrew NG's project: https://github.com/andrewyng/translation-agent\n\n1. Prompt an LLM to translate a text into the target language;\n2. Have the LLM reflect on the translation and provide constructive suggestions for improvement;\n3. Use these suggestions to improve the translation."
+                            },
+                            "label": "Note",
+                            "name": "Brief"
+                        },
+                        "dragHandle": ".note-drag-handle",
+                        "dragging": false,
+                        "height": 227,
+                        "id": "Note:MoodyKnivesCheat",
+                        "position": {
+                            "x": 46.02198421645994,
+                            "y": -267.69527832581736
+                        },
+                        "positionAbsolute": {
+                            "x": 46.02198421645994,
+                            "y": -267.69527832581736
                         },
                         "resizing": false,
                         "selected": false,
                         "sourcePosition": "right",
                         "style": {
-              "height": 157,
-              "width": 252
+                            "height": 227,
+                            "width": 703
                         },
                         "targetPosition": "left",
                         "type": "noteNode",
-            "width": 252,
-            "dragHandle": ".note-drag-handle"
-          }
-        ],
-        "edges": [
-          {
-            "id": "reactflow__edge-begin-Answer:ShortPapersShakec",
-            "markerEnd": "logo",
-            "source": "begin",
-            "sourceHandle": null,
-            "style": {
-              "stroke": "rgb(202 197 245)",
-              "strokeWidth": 2
-            },
-            "target": "Answer:ShortPapersShake",
-            "targetHandle": "c",
-            "type": "buttonEdge"
+                        "width": 703
                     },
                     {
-            "id": "reactflow__edge-Answer:ShortPapersShakeb-Generate:HeavyForksTellb",
-            "markerEnd": "logo",
-            "source": "Answer:ShortPapersShake",
-            "sourceHandle": "b",
+                        "data": {
+                            "form": {
+                                "text": "Many businesses use specialized terms that are not widely used on the internet and that LLMs thus don\u2019t know about, and there are also many terms that can be translated in multiple ways. For example, \u201dopen source\u201d in Spanish can be \u201cC\u00f3digo abierto\u201d or \u201cFuente abierta\u201d; both are fine, but it\u2019d better to pick one and stick with it for a single document.\n\nYou can add those glossary translation into prompt to any of `Translate directly` or 'Reflect'."
+                            },
+                            "label": "Note",
+                            "name": "Tip: Add glossary "
+                        },
+                        "dragHandle": ".note-drag-handle",
+                        "dragging": false,
+                        "height": 181,
+                        "id": "Note:SourCarrotsAct",
+                        "position": {
+                            "x": 65.0676250238289,
+                            "y": 397.6323270065299
+                        },
+                        "positionAbsolute": {
+                            "x": 65.0676250238289,
+                            "y": 397.6323270065299
+                        },
+                        "resizing": false,
+                        "selected": false,
+                        "sourcePosition": "right",
                         "style": {
-              "stroke": "rgb(202 197 245)",
-              "strokeWidth": 2
+                            "height": 181,
+                            "width": 832
                         },
-            "target": "Generate:HeavyForksTell",
-            "targetHandle": "b",
-            "type": "buttonEdge"
-          },
-          {
-            "id": "reactflow__edge-Generate:HeavyForksTellc-Answer:ShortPapersShakec",
-            "markerEnd": "logo",
-            "source": "Generate:HeavyForksTell",
-            "sourceHandle": "c",
-            "style": {
-              "stroke": "rgb(202 197 245)",
-              "strokeWidth": 2
-            },
-            "target": "Answer:ShortPapersShake",
-            "targetHandle": "c",
-            "type": "buttonEdge"
+                        "targetPosition": "left",
+                        "type": "noteNode",
+                        "width": 832
                     }
                 ]
             },

agent/templates/investment_advisor.json

@@ -1,7 +1,7 @@
 {
     "id": 8,
     "title": "Intelligent investment advisor",
-    "description": "An intelligent investment advisor that can answer your financial questions based on real-time domestic financial data and financial information.",
+    "description": "An intelligent investment advisor that answers your financial questions using real-time domestic financial data.",
     "canvas_type": "chatbot",
     "dsl": {
           "answer": [],

agent/templates/medical_consultation.json

@@ -1,7 +1,7 @@
 {
   "id": 7,
   "title": "Medical consultation",
-  "description": "Medical Consultation Assistant, can provide you with some professional consultation suggestions for your reference. Please note that the content provided by the medical assistant is for reference only and may not be authentic or available. Knowledge Base Content Reference: <a href = 'https://huggingface.co/datasets/InfiniFlow/medical_QA/tree/main'> Medical Knowledge Base Reference</a>",
+  "description": "A consultant that offers medical suggestions using an internal QA dataset and PubMed search results. Note that this agent's answers are for reference only and may not be valid. The dataset can be found at https://huggingface.co/datasets/InfiniFlow/medical_QA/tree/main",
   "canvas_type": "chatbot",
    "dsl": {
       "answer": [],
@@ -534,7 +534,7 @@
           {
             "data": {
               "form": {
-                "text": "A prompt sumerize content from search result from PubMed and Q&A dataset."
+                "text": "A prompt summarize content from search result from PubMed and Q&A dataset."
               },
               "label": "Note",
               "name": "N: LLM"

agent/templates/text2sql.json

@@ -440,7 +440,7 @@
           {
             "data": {
               "form": {
-                "text": "DDL(Data Definition Language).\n\nSearches for relevent database creation statements.\n\nIt should bind with a KB to which DDL is dumped in.\nYou could use  'General' as parsing method and ';' as delimiter."
+                "text": "DDL(Data Definition Language).\n\nSearches for relevant database creation statements.\n\nIt should bind with a KB to which DDL is dumped in.\nYou could use  'General' as parsing method and ';' as delimiter."
               },
               "label": "Note",
               "name": "N: DDL"

agent/templates/websearch_assistant.json

@@ -577,7 +577,7 @@
                 "text": "Based on the keywords, searches on Wikipedia and returns the found content."
               },
               "label": "Note",
-              "name": "N: Wiukipedia"
+              "name": "N: Wikipedia"
             },
             "dragging": false,
             "height": 128,

agent/test/client.py

@@ -43,6 +43,7 @@ if __name__ == '__main__':
         else:
             print(ans["content"])
 
-        if DEBUG: print(canvas.path)
+        if DEBUG:
+            print(canvas.path)
         question = input("\n==================== User =====================\n> ")
         canvas.add_user_input(question)

api/__init__.py

@@ -0,0 +1,18 @@
+#
+#  Copyright 2025 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+
+from beartype.claw import beartype_this_package
+beartype_this_package()

api/apps/__init__.py

@@ -13,14 +13,16 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-import logging
 import os
 import sys
+import logging
 from importlib.util import module_from_spec, spec_from_file_location
 from pathlib import Path
 from flask import Blueprint, Flask
 from werkzeug.wrappers.request import Request
 from flask_cors import CORS
+from flasgger import Swagger
+from itsdangerous.url_safe import URLSafeTimedSerializer as Serializer
 
 from api.db import StatusEnum
 from api.db.db_models import close_connection
@@ -29,32 +31,60 @@ from api.utils import CustomJSONEncoder, commands
 
 from flask_session import Session
 from flask_login import LoginManager
-from api.settings import SECRET_KEY, stat_logger
-from api.settings import API_VERSION, access_logger
+from api import settings
 from api.utils.api_utils import server_error_response
-from itsdangerous.url_safe import URLSafeTimedSerializer as Serializer
+from api.constants import API_VERSION
 
-__all__ = ['app']
-
-
-logger = logging.getLogger('flask.app')
-for h in access_logger.handlers:
-    logger.addHandler(h)
+__all__ = ["app"]
 
 Request.json = property(lambda self: self.get_json(force=True, silent=True))
 
 app = Flask(__name__)
+
+# Add this at the beginning of your file to configure Swagger UI
+swagger_config = {
+    "headers": [],
+    "specs": [
+        {
+            "endpoint": "apispec",
+            "route": "/apispec.json",
+            "rule_filter": lambda rule: True,  # Include all endpoints
+            "model_filter": lambda tag: True,  # Include all models
+        }
+    ],
+    "static_url_path": "/flasgger_static",
+    "swagger_ui": True,
+    "specs_route": "/apidocs/",
+}
+
+swagger = Swagger(
+    app,
+    config=swagger_config,
+    template={
+        "swagger": "2.0",
+        "info": {
+            "title": "RAGFlow API",
+            "description": "",
+            "version": "1.0.0",
+        },
+        "securityDefinitions": {
+            "ApiKeyAuth": {"type": "apiKey", "name": "Authorization", "in": "header"}
+        },
+    },
+)
+
 CORS(app, supports_credentials=True, max_age=2592000)
 app.url_map.strict_slashes = False
 app.json_encoder = CustomJSONEncoder
 app.errorhandler(Exception)(server_error_response)
 
-
 ## convince for dev and debug
 # app.config["LOGIN_DISABLED"] = True
 app.config["SESSION_PERMANENT"] = False
 app.config["SESSION_TYPE"] = "filesystem"
-app.config['MAX_CONTENT_LENGTH'] = int(os.environ.get("MAX_CONTENT_LENGTH", 128 * 1024 * 1024))
+app.config["MAX_CONTENT_LENGTH"] = int(
+    os.environ.get("MAX_CONTENT_LENGTH", 128 * 1024 * 1024)
+)
 
 Session(app)
 login_manager = LoginManager()
@@ -64,17 +94,23 @@ commands.register_commands(app)
 
 
 def search_pages_path(pages_dir):
-    app_path_list = [path for path in pages_dir.glob('*_app.py') if not path.name.startswith('.')]
-    api_path_list = [path for path in pages_dir.glob('*sdk/*.py') if not path.name.startswith('.')]
+    app_path_list = [
+        path for path in pages_dir.glob("*_app.py") if not path.name.startswith(".")
+    ]
+    api_path_list = [
+        path for path in pages_dir.glob("*sdk/*.py") if not path.name.startswith(".")
+    ]
     app_path_list.extend(api_path_list)
     return app_path_list
 
 
 def register_page(page_path):
-    path = f'{page_path}'
+    path = f"{page_path}"
 
-    page_name = page_path.stem.rstrip('_app')
-    module_name = '.'.join(page_path.parts[page_path.parts.index('api'):-1] + (page_name,))
+    page_name = page_path.stem.rstrip("_app")
+    module_name = ".".join(
+        page_path.parts[page_path.parts.index("api"): -1] + (page_name,)
+    )
 
     spec = spec_from_file_location(module_name, page_path)
     page = module_from_spec(spec)
@@ -82,8 +118,10 @@ def register_page(page_path):
     page.manager = Blueprint(page_name, module_name)
     sys.modules[module_name] = page
     spec.loader.exec_module(page)
-    page_name = getattr(page, 'page_name', page_name)
-    url_prefix = f'/api/{API_VERSION}' if "/sdk/" in path else f'/{API_VERSION}/{page_name}'
+    page_name = getattr(page, "page_name", page_name)
+    url_prefix = (
+        f"/api/{API_VERSION}" if "/sdk/" in path else f"/{API_VERSION}/{page_name}"
+    )
 
     app.register_blueprint(page.manager, url_prefix=url_prefix)
     return url_prefix
@@ -91,31 +129,31 @@ def register_page(page_path):
 
 pages_dir = [
     Path(__file__).parent,
-    Path(__file__).parent.parent / 'api' / 'apps',
-    Path(__file__).parent.parent / 'api' / 'apps' / 'sdk',
+    Path(__file__).parent.parent / "api" / "apps",
+    Path(__file__).parent.parent / "api" / "apps" / "sdk",
 ]
 
 client_urls_prefix = [
-    register_page(path)
-    for dir in pages_dir
-    for path in search_pages_path(dir)
+    register_page(path) for dir in pages_dir for path in search_pages_path(dir)
 ]
 
 
 @login_manager.request_loader
 def load_user(web_request):
-    jwt = Serializer(secret_key=SECRET_KEY)
+    jwt = Serializer(secret_key=settings.SECRET_KEY)
     authorization = web_request.headers.get("Authorization")
     if authorization:
         try:
             access_token = str(jwt.loads(authorization))
-            user = UserService.query(access_token=access_token, status=StatusEnum.VALID.value)
+            user = UserService.query(
+                access_token=access_token, status=StatusEnum.VALID.value
+            )
             if user:
                 return user[0]
             else:
                 return None
         except Exception as e:
-            stat_logger.exception(e)
+            logging.warning(f"load_user got exception {e}")
             return None
     else:
         return None

api/apps/api_app.py

@@ -25,14 +25,14 @@ from api.db import FileType, LLMType, ParserType, FileSource
 from api.db.db_models import APIToken, Task, File
 from api.db.services import duplicate_name
 from api.db.services.api_service import APITokenService, API4ConversationService
-from api.db.services.dialog_service import DialogService, chat, keyword_extraction
+from api.db.services.dialog_service import DialogService, chat, keyword_extraction, label_question
 from api.db.services.document_service import DocumentService, doc_upload_and_parse
 from api.db.services.file2document_service import File2DocumentService
 from api.db.services.file_service import FileService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.task_service import queue_tasks, TaskService
 from api.db.services.user_service import UserTenantService
-from api.settings import RetCode, retrievaler
+from api import settings
 from api.utils import get_uuid, current_timestamp, datetime_format
 from api.utils.api_utils import server_error_response, get_data_error_result, get_json_result, validate_request, \
     generate_confirmation_token
@@ -45,14 +45,14 @@ from agent.canvas import Canvas
 from functools import partial
 
 
-@manager.route('/new_token', methods=['POST'])
+@manager.route('/new_token', methods=['POST'])  # noqa: F821
 @login_required
 def new_token():
     req = request.json
     try:
         tenants = UserTenantService.query(user_id=current_user.id)
         if not tenants:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
 
         tenant_id = tenants[0].tenant_id
         obj = {"tenant_id": tenant_id, "token": generate_confirmation_token(tenant_id),
@@ -68,20 +68,20 @@ def new_token():
             obj["dialog_id"] = req["dialog_id"]
 
         if not APITokenService.save(**obj):
-            return get_data_error_result(retmsg="Fail to new a dialog!")
+            return get_data_error_result(message="Fail to new a dialog!")
 
         return get_json_result(data=obj)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/token_list', methods=['GET'])
+@manager.route('/token_list', methods=['GET'])  # noqa: F821
 @login_required
 def token_list():
     try:
         tenants = UserTenantService.query(user_id=current_user.id)
         if not tenants:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
 
         id = request.args["dialog_id"] if "dialog_id" in request.args else request.args["canvas_id"]
         objs = APITokenService.query(tenant_id=tenants[0].tenant_id, dialog_id=id)
@@ -90,7 +90,7 @@ def token_list():
         return server_error_response(e)
 
 
-@manager.route('/rm', methods=['POST'])
+@manager.route('/rm', methods=['POST'])  # noqa: F821
 @validate_request("tokens", "tenant_id")
 @login_required
 def rm():
@@ -104,13 +104,13 @@ def rm():
         return server_error_response(e)
 
 
-@manager.route('/stats', methods=['GET'])
+@manager.route('/stats', methods=['GET'])  # noqa: F821
 @login_required
 def stats():
     try:
         tenants = UserTenantService.query(user_id=current_user.id)
         if not tenants:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
         objs = API4ConversationService.stats(
             tenants[0].tenant_id,
             request.args.get(
@@ -135,14 +135,13 @@ def stats():
         return server_error_response(e)
 
 
-@manager.route('/new_conversation', methods=['GET'])
+@manager.route('/new_conversation', methods=['GET'])  # noqa: F821
 def set_conversation():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
-    req = request.json
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
     try:
         if objs[0].source == "agent":
             e, cvs = UserCanvasService.get_by_id(objs[0].dialog_id)
@@ -163,7 +162,7 @@ def set_conversation():
         else:
             e, dia = DialogService.get_by_id(objs[0].dialog_id)
             if not e:
-                return get_data_error_result(retmsg="Dialog not found")
+                return get_data_error_result(message="Dialog not found")
             conv = {
                 "id": get_uuid(),
                 "dialog_id": dia.id,
@@ -176,19 +175,20 @@ def set_conversation():
         return server_error_response(e)
 
 
-@manager.route('/completion', methods=['POST'])
+@manager.route('/completion', methods=['POST'])  # noqa: F821
 @validate_request("conversation_id", "messages")
 def completion():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
     req = request.json
     e, conv = API4ConversationService.get_by_id(req["conversation_id"])
     if not e:
-        return get_data_error_result(retmsg="Conversation not found!")
-    if "quote" not in req: req["quote"] = False
+        return get_data_error_result(message="Conversation not found!")
+    if "quote" not in req:
+        req["quote"] = False
 
     msg = []
     for m in req["messages"]:
@@ -197,7 +197,8 @@ def completion():
         if m["role"] == "assistant" and not msg:
             continue
         msg.append(m)
-    if not msg[-1].get("id"): msg[-1]["id"] = get_uuid()
+    if not msg[-1].get("id"):
+        msg[-1]["id"] = get_uuid()
     message_id = msg[-1]["id"]
 
     def fillin_conv(ans):
@@ -257,19 +258,20 @@ def completion():
                             ans = {"answer": ans["content"], "reference": ans.get("reference", [])}
                             fillin_conv(ans)
                             rename_field(ans)
-                            yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": ans},
+                            yield "data:" + json.dumps({"code": 0, "message": "", "data": ans},
                                                        ensure_ascii=False) + "\n\n"
 
                         canvas.messages.append({"role": "assistant", "content": final_ans["content"], "id": message_id})
+                        canvas.history.append(("assistant", final_ans["content"]))
                         if final_ans.get("reference"):
                             canvas.reference.append(final_ans["reference"])
                         cvs.dsl = json.loads(str(canvas))
                         API4ConversationService.append_message(conv.id, conv.to_dict())
                     except Exception as e:
-                        yield "data:" + json.dumps({"retcode": 500, "retmsg": str(e),
+                        yield "data:" + json.dumps({"code": 500, "message": str(e),
                                                     "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
                                                    ensure_ascii=False) + "\n\n"
-                    yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": True}, ensure_ascii=False) + "\n\n"
+                    yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
 
                 resp = Response(sse(), mimetype="text/event-stream")
                 resp.headers.add_header("Cache-control", "no-cache")
@@ -294,7 +296,7 @@ def completion():
         conv.message.append(msg[-1])
         e, dia = DialogService.get_by_id(conv.dialog_id)
         if not e:
-            return get_data_error_result(retmsg="Dialog not found!")
+            return get_data_error_result(message="Dialog not found!")
         del req["conversation_id"]
         del req["messages"]
 
@@ -309,14 +311,14 @@ def completion():
                 for ans in chat(dia, msg, True, **req):
                     fillin_conv(ans)
                     rename_field(ans)
-                    yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": ans},
+                    yield "data:" + json.dumps({"code": 0, "message": "", "data": ans},
                                                ensure_ascii=False) + "\n\n"
                 API4ConversationService.append_message(conv.id, conv.to_dict())
             except Exception as e:
-                yield "data:" + json.dumps({"retcode": 500, "retmsg": str(e),
+                yield "data:" + json.dumps({"code": 500, "message": str(e),
                                             "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
                                            ensure_ascii=False) + "\n\n"
-            yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": True}, ensure_ascii=False) + "\n\n"
+            yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
 
         if req.get("stream", True):
             resp = Response(stream(), mimetype="text/event-stream")
@@ -339,24 +341,24 @@ def completion():
         return server_error_response(e)
 
 
-@manager.route('/conversation/<conversation_id>', methods=['GET'])
+@manager.route('/conversation/<conversation_id>', methods=['GET'])  # noqa: F821
 # @login_required
 def get(conversation_id):
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
 
     try:
         e, conv = API4ConversationService.get_by_id(conversation_id)
         if not e:
-            return get_data_error_result(retmsg="Conversation not found!")
+            return get_data_error_result(message="Conversation not found!")
 
         conv = conv.to_dict()
         if token != APIToken.query(dialog_id=conv['dialog_id'])[0].token:
-            return get_json_result(data=False, retmsg='Token is not valid for this conversation_id!"',
-                                   retcode=RetCode.AUTHENTICATION_ERROR)
+            return get_json_result(data=False, message='Authentication error: API key is invalid for this conversation_id!"',
+                                   code=settings.RetCode.AUTHENTICATION_ERROR)
 
         for referenct_i in conv['reference']:
             if referenct_i is None or len(referenct_i) == 0:
@@ -370,14 +372,14 @@ def get(conversation_id):
         return server_error_response(e)
 
 
-@manager.route('/document/upload', methods=['POST'])
+@manager.route('/document/upload', methods=['POST'])  # noqa: F821
 @validate_request("kb_name")
 def upload():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
 
     kb_name = request.form.get("kb_name").strip()
     tenant_id = objs[0].tenant_id
@@ -386,19 +388,19 @@ def upload():
         e, kb = KnowledgebaseService.get_by_name(kb_name, tenant_id)
         if not e:
             return get_data_error_result(
-                retmsg="Can't find this knowledgebase!")
+                message="Can't find this knowledgebase!")
         kb_id = kb.id
     except Exception as e:
         return server_error_response(e)
 
     if 'file' not in request.files:
         return get_json_result(
-            data=False, retmsg='No file part!', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='No file part!', code=settings.RetCode.ARGUMENT_ERROR)
 
     file = request.files['file']
     if file.filename == '':
         return get_json_result(
-            data=False, retmsg='No file selected!', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='No file selected!', code=settings.RetCode.ARGUMENT_ERROR)
 
     root_folder = FileService.get_root_folder(tenant_id)
     pf_id = root_folder["id"]
@@ -409,7 +411,7 @@ def upload():
     try:
         if DocumentService.get_doc_count(kb.tenant_id) >= int(os.environ.get('MAX_FILE_NUM_PER_USER', 8192)):
             return get_data_error_result(
-                retmsg="Exceed the maximum file number of a free user!")
+                message="Exceed the maximum file number of a free user!")
 
         filename = duplicate_name(
             DocumentService.query,
@@ -418,7 +420,7 @@ def upload():
         filetype = filename_type(filename)
         if not filetype:
             return get_data_error_result(
-                retmsg="This type of file has not been supported yet!")
+                message="This type of file has not been supported yet!")
 
         location = filename
         while STORAGE_IMPL.obj_exist(kb_id, location):
@@ -467,7 +469,7 @@ def upload():
                 # if str(req["run"]) == TaskStatus.CANCEL.value:
                 tenant_id = DocumentService.get_tenant_id(doc["id"])
                 if not tenant_id:
-                    return get_data_error_result(retmsg="Tenant not found!")
+                    return get_data_error_result(message="Tenant not found!")
 
                 # e, doc = DocumentService.get_by_id(doc["id"])
                 TaskService.filter_delete([Task.doc_id == doc["id"]])
@@ -482,37 +484,37 @@ def upload():
     return get_json_result(data=doc_result.to_json())
 
 
-@manager.route('/document/upload_and_parse', methods=['POST'])
+@manager.route('/document/upload_and_parse', methods=['POST'])  # noqa: F821
 @validate_request("conversation_id")
 def upload_parse():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
 
     if 'file' not in request.files:
         return get_json_result(
-            data=False, retmsg='No file part!', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='No file part!', code=settings.RetCode.ARGUMENT_ERROR)
 
     file_objs = request.files.getlist('file')
     for file_obj in file_objs:
         if file_obj.filename == '':
             return get_json_result(
-                data=False, retmsg='No file selected!', retcode=RetCode.ARGUMENT_ERROR)
+                data=False, message='No file selected!', code=settings.RetCode.ARGUMENT_ERROR)
 
     doc_ids = doc_upload_and_parse(request.form.get("conversation_id"), file_objs, objs[0].tenant_id)
     return get_json_result(data=doc_ids)
 
 
-@manager.route('/list_chunks', methods=['POST'])
+@manager.route('/list_chunks', methods=['POST'])  # noqa: F821
 # @login_required
 def list_chunks():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
 
     req = request.json
 
@@ -526,15 +528,16 @@ def list_chunks():
             doc_id = req['doc_id']
         else:
             return get_json_result(
-                data=False, retmsg="Can't find doc_name or doc_id"
+                data=False, message="Can't find doc_name or doc_id"
             )
+        kb_ids = KnowledgebaseService.get_kb_ids(tenant_id)
 
-        res = retrievaler.chunk_list(doc_id=doc_id, tenant_id=tenant_id)
+        res = settings.retrievaler.chunk_list(doc_id, tenant_id, kb_ids)
         res = [
             {
                 "content": res_item["content_with_weight"],
                 "doc_name": res_item["docnm_kwd"],
-                "img_id": res_item["img_id"]
+                "image_id": res_item["img_id"]
             } for res_item in res
         ]
 
@@ -544,14 +547,14 @@ def list_chunks():
     return get_json_result(data=res)
 
 
-@manager.route('/list_kb_docs', methods=['POST'])
+@manager.route('/list_kb_docs', methods=['POST'])  # noqa: F821
 # @login_required
 def list_kb_docs():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
 
     req = request.json
     tenant_id = objs[0].tenant_id
@@ -561,7 +564,7 @@ def list_kb_docs():
         e, kb = KnowledgebaseService.get_by_name(kb_name, tenant_id)
         if not e:
             return get_data_error_result(
-                retmsg="Can't find this knowledgebase!")
+                message="Can't find this knowledgebase!")
         kb_id = kb.id
 
     except Exception as e:
@@ -583,28 +586,29 @@ def list_kb_docs():
     except Exception as e:
         return server_error_response(e)
 
-@manager.route('/document/infos', methods=['POST'])
+
+@manager.route('/document/infos', methods=['POST'])  # noqa: F821
 @validate_request("doc_ids")
 def docinfos():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
     req = request.json
     doc_ids = req["doc_ids"]
     docs = DocumentService.get_by_ids(doc_ids)
     return get_json_result(data=list(docs.dicts()))
 
 
-@manager.route('/document', methods=['DELETE'])
+@manager.route('/document', methods=['DELETE'])  # noqa: F821
 # @login_required
 def document_rm():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
 
     tenant_id = objs[0].tenant_id
     req = request.json
@@ -616,7 +620,7 @@ def document_rm():
 
         if not doc_ids:
             return get_json_result(
-                data=False, retmsg="Can't find doc_names or doc_ids"
+                data=False, message="Can't find doc_names or doc_ids"
             )
 
     except Exception as e:
@@ -631,16 +635,16 @@ def document_rm():
         try:
             e, doc = DocumentService.get_by_id(doc_id)
             if not e:
-                return get_data_error_result(retmsg="Document not found!")
+                return get_data_error_result(message="Document not found!")
             tenant_id = DocumentService.get_tenant_id(doc_id)
             if not tenant_id:
-                return get_data_error_result(retmsg="Tenant not found!")
+                return get_data_error_result(message="Tenant not found!")
 
             b, n = File2DocumentService.get_storage_address(doc_id=doc_id)
 
             if not DocumentService.remove_document(doc, tenant_id):
                 return get_data_error_result(
-                    retmsg="Database error (Document removal)!")
+                    message="Database error (Document removal)!")
 
             f2d = File2DocumentService.get_by_document_id(doc_id)
             FileService.filter_delete([File.source_type == FileSource.KNOWLEDGEBASE, File.id == f2d[0].file_id])
@@ -651,12 +655,12 @@ def document_rm():
             errors += str(e)
 
     if errors:
-        return get_json_result(data=False, retmsg=errors, retcode=RetCode.SERVER_ERROR)
+        return get_json_result(data=False, message=errors, code=settings.RetCode.SERVER_ERROR)
 
     return get_json_result(data=True)
 
 
-@manager.route('/completion_aibotk', methods=['POST'])
+@manager.route('/completion_aibotk', methods=['POST'])  # noqa: F821
 @validate_request("Authorization", "conversation_id", "word")
 def completion_faq():
     import base64
@@ -666,16 +670,18 @@ def completion_faq():
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
 
     e, conv = API4ConversationService.get_by_id(req["conversation_id"])
     if not e:
-        return get_data_error_result(retmsg="Conversation not found!")
-    if "quote" not in req: req["quote"] = True
+        return get_data_error_result(message="Conversation not found!")
+    if "quote" not in req:
+        req["quote"] = True
 
     msg = []
     msg.append({"role": "user", "content": req["word"]})
-    if not msg[-1].get("id"): msg[-1]["id"] = get_uuid()
+    if not msg[-1].get("id"):
+        msg[-1]["id"] = get_uuid()
     message_id = msg[-1]["id"]
 
     def fillin_conv(ans):
@@ -751,7 +757,7 @@ def completion_faq():
         conv.message.append(msg[-1])
         e, dia = DialogService.get_by_id(conv.dialog_id)
         if not e:
-            return get_data_error_result(retmsg="Dialog not found!")
+            return get_data_error_result(message="Dialog not found!")
         del req["conversation_id"]
 
         if not conv.reference:
@@ -796,14 +802,14 @@ def completion_faq():
         return server_error_response(e)
 
 
-@manager.route('/retrieval', methods=['POST'])
+@manager.route('/retrieval', methods=['POST'])  # noqa: F821
 @validate_request("kb_id", "question")
 def retrieval():
     token = request.headers.get('Authorization').split()[1]
     objs = APIToken.query(token=token)
     if not objs:
         return get_json_result(
-            data=False, retmsg='Token is not valid!"', retcode=RetCode.AUTHENTICATION_ERROR)
+            data=False, message='Authentication error: API key is invalid!"', code=settings.RetCode.AUTHENTICATION_ERROR)
 
     req = request.json
     kb_ids = req.get("kb_id", [])
@@ -820,7 +826,8 @@ def retrieval():
         embd_nms = list(set([kb.embd_id for kb in kbs]))
         if len(embd_nms) != 1:
             return get_json_result(
-                data=False, retmsg='Knowledge bases use different embedding models or does not exist."', retcode=RetCode.AUTHENTICATION_ERROR)
+                data=False, message='Knowledge bases use different embedding models or does not exist."',
+                code=settings.RetCode.AUTHENTICATION_ERROR)
 
         embd_mdl = TenantLLMService.model_instance(
             kbs[0].tenant_id, LLMType.EMBEDDING.value, llm_name=kbs[0].embd_id)
@@ -831,15 +838,15 @@ def retrieval():
         if req.get("keyword", False):
             chat_mdl = TenantLLMService.model_instance(kbs[0].tenant_id, LLMType.CHAT)
             question += keyword_extraction(chat_mdl, question)
-        ranks = retrievaler.retrieval(question, embd_mdl, kbs[0].tenant_id, kb_ids, page, size,
+        ranks = settings.retrievaler.retrieval(question, embd_mdl, kbs[0].tenant_id, kb_ids, page, size,
                                                similarity_threshold, vector_similarity_weight, top,
-            doc_ids, rerank_mdl=rerank_mdl)
+                                               doc_ids, rerank_mdl=rerank_mdl,
+                                               rank_feature=label_question(question, kbs))
         for c in ranks["chunks"]:
-            if "vector" in c:
-                del c["vector"]
+            c.pop("vector", None)
         return get_json_result(data=ranks)
     except Exception as e:
         if str(e).find("not_found") > 0:
-            return get_json_result(data=False, retmsg=f'No chunk found! Check the chunk status please!',
-                                   retcode=RetCode.DATA_ERROR)
+            return get_json_result(data=False, message='No chunk found! Check the chunk status please!',
+                                   code=settings.RetCode.DATA_ERROR)
         return server_error_response(e)

api/apps/canvas_app.py

@@ -14,7 +14,7 @@
 #  limitations under the License.
 #
 import json
-from functools import partial
+import traceback
 from flask import request, Response
 from flask_login import login_required, current_user
 from api.db.services.canvas_service import CanvasTemplateService, UserCanvasService
@@ -23,15 +23,16 @@ from api.utils import get_uuid
 from api.utils.api_utils import get_json_result, server_error_response, validate_request, get_data_error_result
 from agent.canvas import Canvas
 from peewee import MySQLDatabase, PostgresqlDatabase
+from api.db.db_models import APIToken
 
 
-@manager.route('/templates', methods=['GET'])
+@manager.route('/templates', methods=['GET'])  # noqa: F821
 @login_required
 def templates():
     return get_json_result(data=[c.to_dict() for c in CanvasTemplateService.get_all()])
 
 
-@manager.route('/list', methods=['GET'])
+@manager.route('/list', methods=['GET'])  # noqa: F821
 @login_required
 def canvas_list():
     return get_json_result(data=sorted([c.to_dict() for c in \
@@ -39,53 +40,68 @@ def canvas_list():
                            )
 
 
-@manager.route('/rm', methods=['POST'])
+@manager.route('/rm', methods=['POST'])  # noqa: F821
 @validate_request("canvas_ids")
 @login_required
 def rm():
     for i in request.json["canvas_ids"]:
         if not UserCanvasService.query(user_id=current_user.id,id=i):
             return get_json_result(
-                data=False, retmsg=f'Only owner of canvas authorized for this operation.',
-                retcode=RetCode.OPERATING_ERROR)
+                data=False, message='Only owner of canvas authorized for this operation.',
+                code=RetCode.OPERATING_ERROR)
         UserCanvasService.delete_by_id(i)
     return get_json_result(data=True)
 
 
-@manager.route('/set', methods=['POST'])
+@manager.route('/set', methods=['POST'])  # noqa: F821
 @validate_request("dsl", "title")
 @login_required
 def save():
     req = request.json
     req["user_id"] = current_user.id
-    if not isinstance(req["dsl"], str): req["dsl"] = json.dumps(req["dsl"], ensure_ascii=False)
+    if not isinstance(req["dsl"], str):
+        req["dsl"] = json.dumps(req["dsl"], ensure_ascii=False)
 
     req["dsl"] = json.loads(req["dsl"])
     if "id" not in req:
         if UserCanvasService.query(user_id=current_user.id, title=req["title"].strip()):
-            return server_error_response(ValueError("Duplicated title."))
+            return get_data_error_result(message=f"{req['title'].strip()} already exists.")
         req["id"] = get_uuid()
         if not UserCanvasService.save(**req):
-            return get_data_error_result(retmsg="Fail to save canvas.")
+            return get_data_error_result(message="Fail to save canvas.")
     else:
         if not UserCanvasService.query(user_id=current_user.id, id=req["id"]):
             return get_json_result(
-                data=False, retmsg=f'Only owner of canvas authorized for this operation.',
-                retcode=RetCode.OPERATING_ERROR)
+                data=False, message='Only owner of canvas authorized for this operation.',
+                code=RetCode.OPERATING_ERROR)
         UserCanvasService.update_by_id(req["id"], req)
     return get_json_result(data=req)
 
 
-@manager.route('/get/<canvas_id>', methods=['GET'])
+@manager.route('/get/<canvas_id>', methods=['GET'])  # noqa: F821
 @login_required
 def get(canvas_id):
     e, c = UserCanvasService.get_by_id(canvas_id)
     if not e:
-        return get_data_error_result(retmsg="canvas not found.")
+        return get_data_error_result(message="canvas not found.")
+    return get_json_result(data=c.to_dict())
+
+@manager.route('/getsse/<canvas_id>', methods=['GET'])  # type: ignore # noqa: F821
+def getsse(canvas_id):
+    token = request.headers.get('Authorization').split()
+    if len(token) != 2:
+        return get_data_error_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_data_error_result(message='Authentication error: API key is invalid!"')
+    e, c = UserCanvasService.get_by_id(canvas_id)
+    if not e:
+        return get_data_error_result(message="canvas not found.")
     return get_json_result(data=c.to_dict())
 
 
-@manager.route('/completion', methods=['POST'])
+@manager.route('/completion', methods=['POST'])  # noqa: F821
 @validate_request("id")
 @login_required
 def run():
@@ -93,11 +109,11 @@ def run():
     stream = req.get("stream", True)
     e, cvs = UserCanvasService.get_by_id(req["id"])
     if not e:
-        return get_data_error_result(retmsg="canvas not found.")
+        return get_data_error_result(message="canvas not found.")
     if not UserCanvasService.query(user_id=current_user.id, id=req["id"]):
         return get_json_result(
-            data=False, retmsg=f'Only owner of canvas authorized for this operation.',
-            retcode=RetCode.OPERATING_ERROR)
+            data=False, message='Only owner of canvas authorized for this operation.',
+            code=RetCode.OPERATING_ERROR)
 
     if not isinstance(cvs.dsl, str):
         cvs.dsl = json.dumps(cvs.dsl, ensure_ascii=False)
@@ -108,40 +124,44 @@ def run():
         canvas = Canvas(cvs.dsl, current_user.id)
         if "message" in req:
             canvas.messages.append({"role": "user", "content": req["message"], "id": message_id})
-            if len([m for m in canvas.messages if m["role"] == "user"]) > 1:
-                #ten = TenantService.get_info_by(current_user.id)[0]
-                #req["message"] = full_question(ten["tenant_id"], ten["llm_id"], canvas.messages)
-                pass
             canvas.add_user_input(req["message"])
-        answer = canvas.run(stream=stream)
-        print(canvas)
     except Exception as e:
         return server_error_response(e)
 
-    assert answer is not None, "Nothing. Is it over?"
-
     if stream:
-        assert isinstance(answer, partial), "Nothing. Is it over?"
-
         def sse():
             nonlocal answer, cvs
             try:
-                for ans in answer():
+                for ans in canvas.run(stream=True):
+                    if ans.get("running_status"):
+                        yield "data:" + json.dumps({"code": 0, "message": "",
+                                                    "data": {"answer": ans["content"],
+                                                             "running_status": True}},
+                                                   ensure_ascii=False) + "\n\n"
+                        continue
                     for k in ans.keys():
                         final_ans[k] = ans[k]
                     ans = {"answer": ans["content"], "reference": ans.get("reference", [])}
-                    yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": ans}, ensure_ascii=False) + "\n\n"
+                    yield "data:" + json.dumps({"code": 0, "message": "", "data": ans}, ensure_ascii=False) + "\n\n"
 
                 canvas.messages.append({"role": "assistant", "content": final_ans["content"], "id": message_id})
+                canvas.history.append(("assistant", final_ans["content"]))
+                if not canvas.path[-1]:
+                    canvas.path.pop(-1)
                 if final_ans.get("reference"):
                     canvas.reference.append(final_ans["reference"])
                 cvs.dsl = json.loads(str(canvas))
                 UserCanvasService.update_by_id(req["id"], cvs.to_dict())
             except Exception as e:
-                yield "data:" + json.dumps({"retcode": 500, "retmsg": str(e),
+                cvs.dsl = json.loads(str(canvas))
+                if not canvas.path[-1]:
+                    canvas.path.pop(-1)
+                UserCanvasService.update_by_id(req["id"], cvs.to_dict())
+                traceback.print_exc()
+                yield "data:" + json.dumps({"code": 500, "message": str(e),
                                             "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
                                            ensure_ascii=False) + "\n\n"
-            yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": True}, ensure_ascii=False) + "\n\n"
+            yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
 
         resp = Response(sse(), mimetype="text/event-stream")
         resp.headers.add_header("Cache-control", "no-cache")
@@ -150,6 +170,9 @@ def run():
         resp.headers.add_header("Content-Type", "text/event-stream; charset=utf-8")
         return resp
 
+    for answer in canvas.run(stream=False):
+        if answer.get("running_status"):
+            continue
         final_ans["content"] = "\n".join(answer["content"]) if "content" in answer else ""
         canvas.messages.append({"role": "assistant", "content": final_ans["content"], "id": message_id})
         if final_ans.get("reference"):
@@ -159,7 +182,7 @@ def run():
         return get_json_result(data={"answer": final_ans["content"], "reference": final_ans.get("reference", [])})
 
 
-@manager.route('/reset', methods=['POST'])
+@manager.route('/reset', methods=['POST'])  # noqa: F821
 @validate_request("id")
 @login_required
 def reset():
@@ -167,11 +190,11 @@ def reset():
     try:
         e, user_canvas = UserCanvasService.get_by_id(req["id"])
         if not e:
-            return get_data_error_result(retmsg="canvas not found.")
+            return get_data_error_result(message="canvas not found.")
         if not UserCanvasService.query(user_id=current_user.id, id=req["id"]):
             return get_json_result(
-                data=False, retmsg=f'Only owner of canvas authorized for this operation.',
-                retcode=RetCode.OPERATING_ERROR)
+                data=False, message='Only owner of canvas authorized for this operation.',
+                code=RetCode.OPERATING_ERROR)
 
         canvas = Canvas(json.dumps(user_canvas.dsl), current_user.id)
         canvas.reset()
@@ -182,7 +205,51 @@ def reset():
         return server_error_response(e)
 
 
-@manager.route('/test_db_connect', methods=['POST'])
+@manager.route('/input_elements', methods=['GET'])  # noqa: F821
+@login_required
+def input_elements():
+    cvs_id = request.args.get("id")
+    cpn_id = request.args.get("component_id")
+    try:
+        e, user_canvas = UserCanvasService.get_by_id(cvs_id)
+        if not e:
+            return get_data_error_result(message="canvas not found.")
+        if not UserCanvasService.query(user_id=current_user.id, id=cvs_id):
+            return get_json_result(
+                data=False, message='Only owner of canvas authorized for this operation.',
+                code=RetCode.OPERATING_ERROR)
+
+        canvas = Canvas(json.dumps(user_canvas.dsl), current_user.id)
+        return get_json_result(data=canvas.get_component_input_elements(cpn_id))
+    except Exception as e:
+        return server_error_response(e)
+
+
+@manager.route('/debug', methods=['POST'])  # noqa: F821
+@validate_request("id", "component_id", "params")
+@login_required
+def debug():
+    req = request.json
+    for p in req["params"]:
+        assert p.get("key")
+    try:
+        e, user_canvas = UserCanvasService.get_by_id(req["id"])
+        if not e:
+            return get_data_error_result(message="canvas not found.")
+        if not UserCanvasService.query(user_id=current_user.id, id=req["id"]):
+            return get_json_result(
+                data=False, message='Only owner of canvas authorized for this operation.',
+                code=RetCode.OPERATING_ERROR)
+
+        canvas = Canvas(json.dumps(user_canvas.dsl), current_user.id)
+        canvas.get_component(req["component_id"])["obj"]._param.debug_inputs = req["params"]
+        df = canvas.get_component(req["component_id"])["obj"].debug()
+        return get_json_result(data=df.to_dict(orient="records"))
+    except Exception as e:
+        return server_error_response(e)
+
+
+@manager.route('/test_db_connect', methods=['POST'])  # noqa: F821
 @validate_request("db_type", "database", "username", "host", "port", "password")
 @login_required
 def test_db_connect():
@@ -194,8 +261,26 @@ def test_db_connect():
         elif req["db_type"] == 'postgresql':
             db = PostgresqlDatabase(req["database"], user=req["username"], host=req["host"], port=req["port"],
                                     password=req["password"])
+        elif req["db_type"] == 'mssql':
+            import pyodbc
+            connection_string = (
+                f"DRIVER={{ODBC Driver 17 for SQL Server}};"
+                f"SERVER={req['host']},{req['port']};"
+                f"DATABASE={req['database']};"
+                f"UID={req['username']};"
+                f"PWD={req['password']};"
+            )
+            db = pyodbc.connect(connection_string)
+            cursor = db.cursor()
+            cursor.execute("SELECT 1")
+            cursor.close()
+        else:
+            return server_error_response("Unsupported database type.")
+        if req["db_type"] != 'mssql':
             db.connect()
         db.close()
+        
         return get_json_result(data="Database Connection Successful!")
     except Exception as e:
         return server_error_response(e)
+

api/apps/chunk_app.py

@@ -15,16 +15,14 @@
 #
 import datetime
 import json
-import traceback
 
 from flask import request
 from flask_login import login_required, current_user
-from elasticsearch_dsl import Q
 
-from api.db.services.dialog_service import keyword_extraction
+from api.db.services.dialog_service import keyword_extraction, label_question
 from rag.app.qa import rmPrefix, beAdoc
 from rag.nlp import search, rag_tokenizer
-from rag.utils.es_conn import ELASTICSEARCH
+from rag.settings import PAGERANK_FLD
 from rag.utils import rmSpace
 from api.db import LLMType, ParserType
 from api.db.services.knowledgebase_service import KnowledgebaseService
@@ -32,13 +30,13 @@ from api.db.services.llm_service import LLMBundle
 from api.db.services.user_service import UserTenantService
 from api.utils.api_utils import server_error_response, get_data_error_result, validate_request
 from api.db.services.document_service import DocumentService
-from api.settings import RetCode, retrievaler, kg_retrievaler
+from api import settings
 from api.utils.api_utils import get_json_result
-import hashlib
+import xxhash
 import re
 
 
-@manager.route('/list', methods=['POST'])
+@manager.route('/list', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("doc_id")
 def list_chunk():
@@ -50,16 +48,17 @@ def list_chunk():
     try:
         tenant_id = DocumentService.get_tenant_id(req["doc_id"])
         if not tenant_id:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
         e, doc = DocumentService.get_by_id(doc_id)
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
+        kb_ids = KnowledgebaseService.get_kb_ids(tenant_id)
         query = {
             "doc_ids": [doc_id], "page": page, "size": size, "question": question, "sort": True
         }
         if "available_int" in req:
             query["available_int"] = int(req["available_int"])
-        sres = retrievaler.search(query, search.index_name(tenant_id), highlight=True)
+        sres = settings.retrievaler.search(query, search.index_name(tenant_id), kb_ids, highlight=True)
         res = {"total": sres.total, "chunks": [], "doc": doc.to_dict()}
         for id in sres.ids:
             d = {
@@ -70,60 +69,54 @@ def list_chunk():
                 "doc_id": sres.field[id]["doc_id"],
                 "docnm_kwd": sres.field[id]["docnm_kwd"],
                 "important_kwd": sres.field[id].get("important_kwd", []),
-                "img_id": sres.field[id].get("img_id", ""),
-                "available_int": sres.field[id].get("available_int", 1),
-                "positions": sres.field[id].get("position_int", "").split("\t")
+                "question_kwd": sres.field[id].get("question_kwd", []),
+                "image_id": sres.field[id].get("img_id", ""),
+                "available_int": int(sres.field[id].get("available_int", 1)),
+                "positions": sres.field[id].get("position_int", []),
             }
-            if len(d["positions"]) % 5 == 0:
-                poss = []
-                for i in range(0, len(d["positions"]), 5):
-                    poss.append([float(d["positions"][i]), float(d["positions"][i + 1]), float(d["positions"][i + 2]),
-                                 float(d["positions"][i + 3]), float(d["positions"][i + 4])])
-                d["positions"] = poss
+            assert isinstance(d["positions"], list)
+            assert len(d["positions"]) == 0 or (isinstance(d["positions"][0], list) and len(d["positions"][0]) == 5)
             res["chunks"].append(d)
         return get_json_result(data=res)
     except Exception as e:
         if str(e).find("not_found") > 0:
-            return get_json_result(data=False, retmsg=f'No chunk found!',
-                                   retcode=RetCode.DATA_ERROR)
+            return get_json_result(data=False, message='No chunk found!',
+                                   code=settings.RetCode.DATA_ERROR)
         return server_error_response(e)
 
 
-@manager.route('/get', methods=['GET'])
+@manager.route('/get', methods=['GET'])  # noqa: F821
 @login_required
 def get():
     chunk_id = request.args["chunk_id"]
     try:
         tenants = UserTenantService.query(user_id=current_user.id)
         if not tenants:
-            return get_data_error_result(retmsg="Tenant not found!")
-        res = ELASTICSEARCH.get(
-            chunk_id, search.index_name(
-                tenants[0].tenant_id))
-        if not res.get("found"):
-            return server_error_response("Chunk not found")
-        id = res["_id"]
-        res = res["_source"]
-        res["chunk_id"] = id
+            return get_data_error_result(message="Tenant not found!")
+        tenant_id = tenants[0].tenant_id
+
+        kb_ids = KnowledgebaseService.get_kb_ids(tenant_id)
+        chunk = settings.docStoreConn.get(chunk_id, search.index_name(tenant_id), kb_ids)
+        if chunk is None:
+            return server_error_response(Exception("Chunk not found"))
         k = []
-        for n in res.keys():
+        for n in chunk.keys():
             if re.search(r"(_vec$|_sm_|_tks|_ltks)", n):
                 k.append(n)
         for n in k:
-            del res[n]
+            del chunk[n]
 
-        return get_json_result(data=res)
+        return get_json_result(data=chunk)
     except Exception as e:
         if str(e).find("NotFoundError") >= 0:
-            return get_json_result(data=False, retmsg=f'Chunk not found!',
-                                   retcode=RetCode.DATA_ERROR)
+            return get_json_result(data=False, message='Chunk not found!',
+                                   code=settings.RetCode.DATA_ERROR)
         return server_error_response(e)
 
 
-@manager.route('/set', methods=['POST'])
+@manager.route('/set', methods=['POST'])  # noqa: F821
 @login_required
-@validate_request("doc_id", "chunk_id", "content_with_weight",
-                  "important_kwd")
+@validate_request("doc_id", "chunk_id", "content_with_weight")
 def set():
     req = request.json
     d = {
@@ -131,73 +124,80 @@ def set():
         "content_with_weight": req["content_with_weight"]}
     d["content_ltks"] = rag_tokenizer.tokenize(req["content_with_weight"])
     d["content_sm_ltks"] = rag_tokenizer.fine_grained_tokenize(d["content_ltks"])
+    if "important_kwd" in req:
         d["important_kwd"] = req["important_kwd"]
         d["important_tks"] = rag_tokenizer.tokenize(" ".join(req["important_kwd"]))
+    if "question_kwd" in req:
+        d["question_kwd"] = req["question_kwd"]
+        d["question_tks"] = rag_tokenizer.tokenize("\n".join(req["question_kwd"]))
+    if "tag_kwd" in req:
+        d["tag_kwd"] = req["tag_kwd"]
+    if "tag_feas" in req:
+        d["tag_feas"] = req["tag_feas"]
     if "available_int" in req:
         d["available_int"] = req["available_int"]
 
     try:
         tenant_id = DocumentService.get_tenant_id(req["doc_id"])
         if not tenant_id:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
 
         embd_id = DocumentService.get_embd_id(req["doc_id"])
         embd_mdl = LLMBundle(tenant_id, LLMType.EMBEDDING, embd_id)
 
         e, doc = DocumentService.get_by_id(req["doc_id"])
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
 
         if doc.parser_id == ParserType.QA:
             arr = [
                 t for t in re.split(
                     r"[\n\t]",
                     req["content_with_weight"]) if len(t) > 1]
-            if len(arr) != 2:
-                return get_data_error_result(
-                    retmsg="Q&A must be separated by TAB/ENTER key.")
-            q, a = rmPrefix(arr[0]), rmPrefix(arr[1])
-            d = beAdoc(d, arr[0], arr[1], not any(
+            q, a = rmPrefix(arr[0]), rmPrefix("\n".join(arr[1:]))
+            d = beAdoc(d, q, a, not any(
                 [rag_tokenizer.is_chinese(t) for t in q + a]))
 
-        v, c = embd_mdl.encode([doc.name, req["content_with_weight"]])
+        v, c = embd_mdl.encode([doc.name, req["content_with_weight"] if not d.get("question_kwd") else "\n".join(d["question_kwd"])])
         v = 0.1 * v[0] + 0.9 * v[1] if doc.parser_id != ParserType.QA else v[1]
         d["q_%d_vec" % len(v)] = v.tolist()
-        ELASTICSEARCH.upsert([d], search.index_name(tenant_id))
+        settings.docStoreConn.update({"id": req["chunk_id"]}, d, search.index_name(tenant_id), doc.kb_id)
         return get_json_result(data=True)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/switch', methods=['POST'])
+@manager.route('/switch', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("chunk_ids", "available_int", "doc_id")
 def switch():
     req = request.json
     try:
-        tenant_id = DocumentService.get_tenant_id(req["doc_id"])
-        if not tenant_id:
-            return get_data_error_result(retmsg="Tenant not found!")
-        if not ELASTICSEARCH.upsert([{"id": i, "available_int": int(req["available_int"])} for i in req["chunk_ids"]],
-                                    search.index_name(tenant_id)):
-            return get_data_error_result(retmsg="Index updating failure")
+        e, doc = DocumentService.get_by_id(req["doc_id"])
+        if not e:
+            return get_data_error_result(message="Document not found!")
+        for cid in req["chunk_ids"]:
+            if not settings.docStoreConn.update({"id": cid},
+                                                {"available_int": int(req["available_int"])},
+                                                search.index_name(DocumentService.get_tenant_id(req["doc_id"])),
+                                                doc.kb_id):
+                return get_data_error_result(message="Index updating failure")
         return get_json_result(data=True)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/rm', methods=['POST'])
+@manager.route('/rm', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("chunk_ids", "doc_id")
 def rm():
     req = request.json
     try:
-        if not ELASTICSEARCH.deleteByQuery(
-                Q("ids", values=req["chunk_ids"]), search.index_name(current_user.id)):
-            return get_data_error_result(retmsg="Index updating failure")
         e, doc = DocumentService.get_by_id(req["doc_id"])
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
+        if not settings.docStoreConn.delete({"id": req["chunk_ids"]}, search.index_name(current_user.id), doc.kb_id):
+            return get_data_error_result(message="Index updating failure")
         deleted_chunk_ids = req["chunk_ids"]
         chunk_number = len(deleted_chunk_ids)
         DocumentService.decrement_chunk_num(doc.id, doc.kb_id, 1, chunk_number, 0)
@@ -206,41 +206,48 @@ def rm():
         return server_error_response(e)
 
 
-@manager.route('/create', methods=['POST'])
+@manager.route('/create', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("doc_id", "content_with_weight")
 def create():
     req = request.json
-    md5 = hashlib.md5()
-    md5.update((req["content_with_weight"] + req["doc_id"]).encode("utf-8"))
-    chunck_id = md5.hexdigest()
+    chunck_id = xxhash.xxh64((req["content_with_weight"] + req["doc_id"]).encode("utf-8")).hexdigest()
     d = {"id": chunck_id, "content_ltks": rag_tokenizer.tokenize(req["content_with_weight"]),
          "content_with_weight": req["content_with_weight"]}
     d["content_sm_ltks"] = rag_tokenizer.fine_grained_tokenize(d["content_ltks"])
     d["important_kwd"] = req.get("important_kwd", [])
     d["important_tks"] = rag_tokenizer.tokenize(" ".join(req.get("important_kwd", [])))
+    d["question_kwd"] = req.get("question_kwd", [])
+    d["question_tks"] = rag_tokenizer.tokenize("\n".join(req.get("question_kwd", [])))
     d["create_time"] = str(datetime.datetime.now()).replace("T", " ")[:19]
     d["create_timestamp_flt"] = datetime.datetime.now().timestamp()
 
     try:
         e, doc = DocumentService.get_by_id(req["doc_id"])
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
         d["kb_id"] = [doc.kb_id]
         d["docnm_kwd"] = doc.name
+        d["title_tks"] = rag_tokenizer.tokenize(doc.name)
         d["doc_id"] = doc.id
 
         tenant_id = DocumentService.get_tenant_id(req["doc_id"])
         if not tenant_id:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
+
+        e, kb = KnowledgebaseService.get_by_id(doc.kb_id)
+        if not e:
+            return get_data_error_result(message="Knowledgebase not found!")
+        if kb.pagerank:
+            d[PAGERANK_FLD] = kb.pagerank
 
         embd_id = DocumentService.get_embd_id(req["doc_id"])
         embd_mdl = LLMBundle(tenant_id, LLMType.EMBEDDING.value, embd_id)
 
-        v, c = embd_mdl.encode([doc.name, req["content_with_weight"]])
+        v, c = embd_mdl.encode([doc.name, req["content_with_weight"] if not d["question_kwd"] else "\n".join(d["question_kwd"])])
         v = 0.1 * v[0] + 0.9 * v[1]
         d["q_%d_vec" % len(v)] = v.tolist()
-        ELASTICSEARCH.upsert([d], search.index_name(tenant_id))
+        settings.docStoreConn.insert([d], search.index_name(tenant_id), doc.kb_id)
 
         DocumentService.increment_chunk_num(
             doc.id, doc.kb_id, c, 1, 0)
@@ -249,7 +256,7 @@ def create():
         return server_error_response(e)
 
 
-@manager.route('/retrieval_test', methods=['POST'])
+@manager.route('/retrieval_test', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("kb_id", "question")
 def retrieval_test():
@@ -257,28 +264,32 @@ def retrieval_test():
     page = int(req.get("page", 1))
     size = int(req.get("size", 30))
     question = req["question"]
-    kb_id = req["kb_id"]
-    if isinstance(kb_id, str): kb_id = [kb_id]
+    kb_ids = req["kb_id"]
+    if isinstance(kb_ids, str):
+        kb_ids = [kb_ids]
     doc_ids = req.get("doc_ids", [])
     similarity_threshold = float(req.get("similarity_threshold", 0.0))
     vector_similarity_weight = float(req.get("vector_similarity_weight", 0.3))
+    use_kg = req.get("use_kg", False)
     top = int(req.get("top_k", 1024))
+    tenant_ids = []
 
     try:
         tenants = UserTenantService.query(user_id=current_user.id)
-        for kid in kb_id:
+        for kb_id in kb_ids:
             for tenant in tenants:
                 if KnowledgebaseService.query(
-                        tenant_id=tenant.tenant_id, id=kid):
+                        tenant_id=tenant.tenant_id, id=kb_id):
+                    tenant_ids.append(tenant.tenant_id)
                     break
             else:
                 return get_json_result(
-                    data=False, retmsg=f'Only owner of knowledgebase authorized for this operation.',
-                    retcode=RetCode.OPERATING_ERROR)
+                    data=False, message='Only owner of knowledgebase authorized for this operation.',
+                    code=settings.RetCode.OPERATING_ERROR)
 
-        e, kb = KnowledgebaseService.get_by_id(kb_id[0])
+        e, kb = KnowledgebaseService.get_by_id(kb_ids[0])
         if not e:
-            return get_data_error_result(retmsg="Knowledgebase not found!")
+            return get_data_error_result(message="Knowledgebase not found!")
 
         embd_mdl = LLMBundle(kb.tenant_id, LLMType.EMBEDDING.value, llm_name=kb.embd_id)
 
@@ -290,38 +301,50 @@ def retrieval_test():
             chat_mdl = LLMBundle(kb.tenant_id, LLMType.CHAT)
             question += keyword_extraction(chat_mdl, question)
 
-        retr = retrievaler if kb.parser_id != ParserType.KG else kg_retrievaler
-        ranks = retr.retrieval(question, embd_mdl, kb.tenant_id, kb_id, page, size,
+        labels = label_question(question, [kb])
+        ranks = settings.retrievaler.retrieval(question, embd_mdl, tenant_ids, kb_ids, page, size,
                                similarity_threshold, vector_similarity_weight, top,
-                               doc_ids, rerank_mdl=rerank_mdl, highlight=req.get("highlight"))
+                               doc_ids, rerank_mdl=rerank_mdl, highlight=req.get("highlight"),
+                               rank_feature=labels
+                               )
+        if use_kg:
+            ck = settings.kg_retrievaler.retrieval(question,
+                                                   tenant_ids,
+                                                   kb_ids,
+                                                   embd_mdl,
+                                                   LLMBundle(kb.tenant_id, LLMType.CHAT))
+            if ck["content_with_weight"]:
+                ranks["chunks"].insert(0, ck)
+
         for c in ranks["chunks"]:
-            if "vector" in c:
-                del c["vector"]
+            c.pop("vector", None)
+        ranks["labels"] = labels
 
         return get_json_result(data=ranks)
     except Exception as e:
         if str(e).find("not_found") > 0:
-            return get_json_result(data=False, retmsg=f'No chunk found! Check the chunk status please!',
-                                   retcode=RetCode.DATA_ERROR)
+            return get_json_result(data=False, message='No chunk found! Check the chunk status please!',
+                                   code=settings.RetCode.DATA_ERROR)
         return server_error_response(e)
 
 
-@manager.route('/knowledge_graph', methods=['GET'])
+@manager.route('/knowledge_graph', methods=['GET'])  # noqa: F821
 @login_required
 def knowledge_graph():
     doc_id = request.args["doc_id"]
+    tenant_id = DocumentService.get_tenant_id(doc_id)
+    kb_ids = KnowledgebaseService.get_kb_ids(tenant_id)
     req = {
         "doc_ids": [doc_id],
         "knowledge_graph_kwd": ["graph", "mind_map"]
     }
-    tenant_id = DocumentService.get_tenant_id(doc_id)
-    sres = retrievaler.search(req, search.index_name(tenant_id))
+    sres = settings.retrievaler.search(req, search.index_name(tenant_id), kb_ids)
     obj = {"graph": {}, "mind_map": {}}
     for id in sres.ids[:2]:
         ty = sres.field[id]["knowledge_graph_kwd"]
         try:
             content_json = json.loads(sres.field[id]["content_with_weight"])
-        except Exception as e:
+        except Exception:
             continue
 
         if ty == 'mind_map':
@@ -344,4 +367,3 @@ def knowledge_graph():
         obj[ty] = content_json
 
     return get_json_result(data=obj)
-

api/apps/conversation_app.py

@@ -17,21 +17,24 @@ import json
 import re
 import traceback
 from copy import deepcopy
+from api.db.db_models import APIToken
+
+from api.db.services.conversation_service import ConversationService, structure_answer
 from api.db.services.user_service import UserTenantService
 from flask import request, Response
 from flask_login import login_required, current_user
 
 from api.db import LLMType
-from api.db.services.dialog_service import DialogService, ConversationService, chat, ask
+from api.db.services.dialog_service import DialogService, chat, ask, label_question
 from api.db.services.knowledgebase_service import KnowledgebaseService
-from api.db.services.llm_service import LLMBundle, TenantService, TenantLLMService
-from api.settings import RetCode, retrievaler
+from api.db.services.llm_service import LLMBundle, TenantService
+from api import settings
 from api.utils.api_utils import get_json_result
 from api.utils.api_utils import server_error_response, get_data_error_result, validate_request
-from graphrag.mind_map_extractor import MindMapExtractor
+from graphrag.general.mind_map_extractor import MindMapExtractor
 
 
-@manager.route('/set', methods=['POST'])
+@manager.route('/set', methods=['POST'])  # noqa: F821
 @login_required
 def set_conversation():
     req = request.json
@@ -42,11 +45,11 @@ def set_conversation():
         del req["conversation_id"]
         try:
             if not ConversationService.update_by_id(conv_id, req):
-                return get_data_error_result(retmsg="Conversation not found!")
+                return get_data_error_result(message="Conversation not found!")
             e, conv = ConversationService.get_by_id(conv_id)
             if not e:
                 return get_data_error_result(
-                    retmsg="Fail to update a conversation!")
+                    message="Fail to update a conversation!")
             conv = conv.to_dict()
             return get_json_result(data=conv)
         except Exception as e:
@@ -55,7 +58,7 @@ def set_conversation():
     try:
         e, dia = DialogService.get_by_id(req["dialog_id"])
         if not e:
-            return get_data_error_result(retmsg="Dialog not found")
+            return get_data_error_result(message="Dialog not found")
         conv = {
             "id": conv_id,
             "dialog_id": req["dialog_id"],
@@ -63,38 +66,76 @@ def set_conversation():
             "message": [{"role": "assistant", "content": dia.prompt_config["prologue"]}]
         }
         ConversationService.save(**conv)
-        e, conv = ConversationService.get_by_id(conv["id"])
-        if not e:
-            return get_data_error_result(retmsg="Fail to new a conversation!")
-        conv = conv.to_dict()
         return get_json_result(data=conv)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/get', methods=['GET'])
+@manager.route('/get', methods=['GET'])  # noqa: F821
 @login_required
 def get():
     conv_id = request.args["conversation_id"]
     try:
+        
         e, conv = ConversationService.get_by_id(conv_id)
         if not e:
-            return get_data_error_result(retmsg="Conversation not found!")
+            return get_data_error_result(message="Conversation not found!")
         tenants = UserTenantService.query(user_id=current_user.id)
+        avatar =None
         for tenant in tenants:
-            if DialogService.query(tenant_id=tenant.tenant_id, id=conv.dialog_id):
+            dialog = DialogService.query(tenant_id=tenant.tenant_id, id=conv.dialog_id)
+            if dialog and len(dialog)>0:
+                avatar = dialog[0].icon
                 break
         else:
             return get_json_result(
-                data=False, retmsg=f'Only owner of conversation authorized for this operation.',
-                retcode=RetCode.OPERATING_ERROR)
+                data=False, message='Only owner of conversation authorized for this operation.',
+                code=settings.RetCode.OPERATING_ERROR)
+
+        def get_value(d, k1, k2):
+            return d.get(k1, d.get(k2))
+
+        for ref in conv.reference:
+            if isinstance(ref, list):
+                continue
+            ref["chunks"] = [{
+                "id": get_value(ck, "chunk_id", "id"),
+                "content": get_value(ck, "content", "content_with_weight"),
+                "document_id": get_value(ck, "doc_id", "document_id"),
+                "document_name": get_value(ck, "docnm_kwd", "document_name"),
+                "dataset_id": get_value(ck, "kb_id", "dataset_id"),
+                "image_id": get_value(ck, "image_id", "img_id"),
+                "positions": get_value(ck, "positions", "position_int"),
+            } for ck in ref.get("chunks", [])]
+
         conv = conv.to_dict()
+        conv["avatar"]=avatar
         return get_json_result(data=conv)
     except Exception as e:
         return server_error_response(e)
 
+@manager.route('/getsse/<dialog_id>', methods=['GET'])  # type: ignore # noqa: F821
+def getsse(dialog_id):
     
-@manager.route('/rm', methods=['POST'])
+    token = request.headers.get('Authorization').split()
+    if len(token) != 2:
+        return get_data_error_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_data_error_result(message='Authentication error: API key is invalid!"')
+    try:
+        e, conv = DialogService.get_by_id(dialog_id)
+        if not e:
+            return get_data_error_result(message="Dialog not found!")
+        conv = conv.to_dict()
+        conv["avatar"]= conv["icon"]
+        del conv["icon"]
+        return get_json_result(data=conv)
+    except Exception as e:
+        return server_error_response(e)
+
+@manager.route('/rm', methods=['POST'])  # noqa: F821
 @login_required
 def rm():
     conv_ids = request.json["conversation_ids"]
@@ -102,41 +143,42 @@ def rm():
         for cid in conv_ids:
             exist, conv = ConversationService.get_by_id(cid)
             if not exist:
-                return get_data_error_result(retmsg="Conversation not found!")
+                return get_data_error_result(message="Conversation not found!")
             tenants = UserTenantService.query(user_id=current_user.id)
             for tenant in tenants:
                 if DialogService.query(tenant_id=tenant.tenant_id, id=conv.dialog_id):
                     break
             else:
                 return get_json_result(
-                    data=False, retmsg=f'Only owner of conversation authorized for this operation.',
-                    retcode=RetCode.OPERATING_ERROR)
+                    data=False, message='Only owner of conversation authorized for this operation.',
+                    code=settings.RetCode.OPERATING_ERROR)
             ConversationService.delete_by_id(cid)
         return get_json_result(data=True)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/list', methods=['GET'])
+@manager.route('/list', methods=['GET'])  # noqa: F821
 @login_required
 def list_convsersation():
     dialog_id = request.args["dialog_id"]
     try:
         if not DialogService.query(tenant_id=current_user.id, id=dialog_id):
             return get_json_result(
-                data=False, retmsg=f'Only owner of dialog authorized for this operation.',
-                retcode=RetCode.OPERATING_ERROR)
+                data=False, message='Only owner of dialog authorized for this operation.',
+                code=settings.RetCode.OPERATING_ERROR)
         convs = ConversationService.query(
             dialog_id=dialog_id,
             order_by=ConversationService.model.create_time,
             reverse=True)
+
         convs = [d.to_dict() for d in convs]
         return get_json_result(data=convs)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/completion', methods=['POST'])
+@manager.route('/completion', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("conversation_id", "messages")
 def completion():
@@ -152,42 +194,49 @@ def completion():
     try:
         e, conv = ConversationService.get_by_id(req["conversation_id"])
         if not e:
-            return get_data_error_result(retmsg="Conversation not found!")
+            return get_data_error_result(message="Conversation not found!")
         conv.message = deepcopy(req["messages"])
         e, dia = DialogService.get_by_id(conv.dialog_id)
         if not e:
-            return get_data_error_result(retmsg="Dialog not found!")
+            return get_data_error_result(message="Dialog not found!")
         del req["conversation_id"]
         del req["messages"]
 
         if not conv.reference:
             conv.reference = []
-        conv.message.append({"role": "assistant", "content": "", "id": message_id})
-        conv.reference.append({"chunks": [], "doc_aggs": []})
-
-        def fillin_conv(ans):
-            nonlocal conv, message_id
-            if not conv.reference:
-                conv.reference.append(ans["reference"])
         else:
-                conv.reference[-1] = ans["reference"]
-            conv.message[-1] = {"role": "assistant", "content": ans["answer"],
-                                "id": message_id, "prompt": ans.get("prompt", "")}
-            ans["id"] = message_id
+            def get_value(d, k1, k2):
+                return d.get(k1, d.get(k2))
 
+            for ref in conv.reference:
+                if isinstance(ref, list):
+                    continue
+                ref["chunks"] = [{
+                    "id": get_value(ck, "chunk_id", "id"),
+                    "content": get_value(ck, "content", "content_with_weight"),
+                    "document_id": get_value(ck, "doc_id", "document_id"),
+                    "document_name": get_value(ck, "docnm_kwd", "document_name"),
+                    "dataset_id": get_value(ck, "kb_id", "dataset_id"),
+                    "image_id": get_value(ck, "image_id", "img_id"),
+                    "positions": get_value(ck, "positions", "position_int"),
+                } for ck in ref.get("chunks", [])]
+
+        if not conv.reference:
+            conv.reference = []
+        conv.reference.append({"chunks": [], "doc_aggs": []})
         def stream():
             nonlocal dia, msg, req, conv
             try:
                 for ans in chat(dia, msg, True, **req):
-                    fillin_conv(ans)
-                    yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": ans}, ensure_ascii=False) + "\n\n"
+                    ans = structure_answer(conv, ans, message_id, conv.id)
+                    yield "data:" + json.dumps({"code": 0, "message": "", "data": ans}, ensure_ascii=False) + "\n\n"
                 ConversationService.update_by_id(conv.id, conv.to_dict())
             except Exception as e:
                 traceback.print_exc()
-                yield "data:" + json.dumps({"retcode": 500, "retmsg": str(e),
+                yield "data:" + json.dumps({"code": 500, "message": str(e),
                                             "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
                                            ensure_ascii=False) + "\n\n"
-            yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": True}, ensure_ascii=False) + "\n\n"
+            yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
 
         if req.get("stream", True):
             resp = Response(stream(), mimetype="text/event-stream")
@@ -200,8 +249,7 @@ def completion():
         else:
             answer = None
             for ans in chat(dia, msg, **req):
-                answer = ans
-                fillin_conv(ans)
+                answer = structure_answer(conv, ans, message_id, req["conversation_id"])
                 ConversationService.update_by_id(conv.id, conv.to_dict())
                 break
             return get_json_result(data=answer)
@@ -209,7 +257,7 @@ def completion():
         return server_error_response(e)
 
 
-@manager.route('/tts', methods=['POST'])
+@manager.route('/tts', methods=['POST'])  # noqa: F821
 @login_required
 def tts():
     req = request.json
@@ -217,11 +265,11 @@ def tts():
 
     tenants = TenantService.get_info_by(current_user.id)
     if not tenants:
-        return get_data_error_result(retmsg="Tenant not found!")
+        return get_data_error_result(message="Tenant not found!")
 
     tts_id = tenants[0]["tts_id"]
     if not tts_id:
-        return get_data_error_result(retmsg="No default TTS model is set")
+        return get_data_error_result(message="No default TTS model is set")
 
     tts_mdl = LLMBundle(tenants[0]["tenant_id"], LLMType.TTS, tts_id)
 
@@ -231,7 +279,7 @@ def tts():
                 for chunk in tts_mdl.tts(txt):
                     yield chunk
         except Exception as e:
-            yield ("data:" + json.dumps({"retcode": 500, "retmsg": str(e),
+            yield ("data:" + json.dumps({"code": 500, "message": str(e),
                                          "data": {"answer": "**ERROR**: " + str(e)}},
                                         ensure_ascii=False)).encode('utf-8')
 
@@ -243,14 +291,14 @@ def tts():
     return resp
 
 
-@manager.route('/delete_msg', methods=['POST'])
+@manager.route('/delete_msg', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("conversation_id", "message_id")
 def delete_msg():
     req = request.json
     e, conv = ConversationService.get_by_id(req["conversation_id"])
     if not e:
-        return get_data_error_result(retmsg="Conversation not found!")
+        return get_data_error_result(message="Conversation not found!")
 
     conv = conv.to_dict()
     for i, msg in enumerate(conv["message"]):
@@ -266,14 +314,14 @@ def delete_msg():
     return get_json_result(data=conv)
 
 
-@manager.route('/thumbup', methods=['POST'])
+@manager.route('/thumbup', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("conversation_id", "message_id")
 def thumbup():
     req = request.json
     e, conv = ConversationService.get_by_id(req["conversation_id"])
     if not e:
-        return get_data_error_result(retmsg="Conversation not found!")
+        return get_data_error_result(message="Conversation not found!")
     up_down = req.get("set")
     feedback = req.get("feedback", "")
     conv = conv.to_dict()
@@ -281,32 +329,35 @@ def thumbup():
         if req["message_id"] == msg.get("id", "") and msg.get("role", "") == "assistant":
             if up_down:
                 msg["thumbup"] = True
-                if "feedback" in msg: del msg["feedback"]
+                if "feedback" in msg:
+                    del msg["feedback"]
             else:
                 msg["thumbup"] = False
-                if feedback: msg["feedback"] = feedback
+                if feedback:
+                    msg["feedback"] = feedback
             break
 
     ConversationService.update_by_id(conv["id"], conv)
     return get_json_result(data=conv)
 
 
-@manager.route('/ask', methods=['POST'])
+@manager.route('/ask', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("question", "kb_ids")
 def ask_about():
     req = request.json
     uid = current_user.id
+
     def stream():
         nonlocal req, uid
         try:
             for ans in ask(req["question"], req["kb_ids"], uid):
-                yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": ans}, ensure_ascii=False) + "\n\n"
+                yield "data:" + json.dumps({"code": 0, "message": "", "data": ans}, ensure_ascii=False) + "\n\n"
         except Exception as e:
-            yield "data:" + json.dumps({"retcode": 500, "retmsg": str(e),
+            yield "data:" + json.dumps({"code": 500, "message": str(e),
                                         "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
                                        ensure_ascii=False) + "\n\n"
-        yield "data:" + json.dumps({"retcode": 0, "retmsg": "", "data": True}, ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
 
     resp = Response(stream(), mimetype="text/event-stream")
     resp.headers.add_header("Cache-control", "no-cache")
@@ -316,7 +367,7 @@ def ask_about():
     return resp
 
 
-@manager.route('/mindmap', methods=['POST'])
+@manager.route('/mindmap', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("question", "kb_ids")
 def mindmap():
@@ -324,13 +375,15 @@ def mindmap():
     kb_ids = req["kb_ids"]
     e, kb = KnowledgebaseService.get_by_id(kb_ids[0])
     if not e:
-        return get_data_error_result(retmsg="Knowledgebase not found!")
+        return get_data_error_result(message="Knowledgebase not found!")
 
-    embd_mdl = TenantLLMService.model_instance(
-        kb.tenant_id, LLMType.EMBEDDING.value, llm_name=kb.embd_id)
+    embd_mdl = LLMBundle(kb.tenant_id, LLMType.EMBEDDING, llm_name=kb.embd_id)
     chat_mdl = LLMBundle(current_user.id, LLMType.CHAT)
-    ranks = retrievaler.retrieval(req["question"], embd_mdl, kb.tenant_id, kb_ids, 1, 12,
-                           0.3, 0.3, aggs=False)
+    question = req["question"]
+    ranks = settings.retrievaler.retrieval(question, embd_mdl, kb.tenant_id, kb_ids, 1, 12,
+                                           0.3, 0.3, aggs=False,
+                                           rank_feature=label_question(question, [kb])
+                                           )
     mindmap = MindMapExtractor(chat_mdl)
     mind_map = mindmap([c["content_with_weight"] for c in ranks["chunks"]]).output
     if "error" in mind_map:
@@ -338,7 +391,7 @@ def mindmap():
     return get_json_result(data=mind_map)
 
 
-@manager.route('/related_questions', methods=['POST'])
+@manager.route('/related_questions', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("question")
 def related_questions():

api/apps/dialog_app.py

@@ -20,27 +20,27 @@ from api.db.services.dialog_service import DialogService
 from api.db import StatusEnum
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.user_service import TenantService, UserTenantService
-from api.settings import RetCode
+from api import settings
 from api.utils.api_utils import server_error_response, get_data_error_result, validate_request
 from api.utils import get_uuid
 from api.utils.api_utils import get_json_result
 
 
-@manager.route('/set', methods=['POST'])
+@manager.route('/set', methods=['POST'])  # noqa: F821
 @login_required
 def set_dialog():
     req = request.json
     dialog_id = req.get("dialog_id")
     name = req.get("name", "New Dialog")
-    description = req.get("description", "A helpful Dialog")
+    description = req.get("description", "A helpful dialog")
     icon = req.get("icon", "")
     top_n = req.get("top_n", 6)
     top_k = req.get("top_k", 1024)
     rerank_id = req.get("rerank_id", "")
-    if not rerank_id: req["rerank_id"] = ""
+    if not rerank_id:
+        req["rerank_id"] = ""
     similarity_threshold = req.get("similarity_threshold", 0.1)
     vector_similarity_weight = req.get("vector_similarity_weight", 0.3)
-    if vector_similarity_weight is None: vector_similarity_weight = 0.3
     llm_setting = req.get("llm_setting", {})
     default_prompt = {
         "system": """你是一个智能助手，请总结知识库的内容来回答问题，请列举知识库中的数据详细回答。当所有知识库内容都与问题无关时，你的回答必须包括“知识库中未找到您要的答案！”这句话。回答需要考虑聊天历史。
@@ -68,17 +68,23 @@ def set_dialog():
             continue
         if prompt_config["system"].find("{%s}" % p["key"]) < 0:
             return get_data_error_result(
-                retmsg="Parameter '{}' is not used".format(p["key"]))
+                message="Parameter '{}' is not used".format(p["key"]))
 
     try:
         e, tenant = TenantService.get_by_id(current_user.id)
         if not e:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
+        kbs = KnowledgebaseService.get_by_ids(req.get("kb_ids"))
+        embd_count = len(set([kb.embd_id for kb in kbs]))
+        if embd_count != 1:
+            return get_data_error_result(message=f'Datasets use different embedding models: {[kb.embd_id for kb in kbs]}"')
+
         llm_id = req.get("llm_id", tenant.llm_id)
         if not dialog_id:
             if not req.get("kb_ids"):
                 return get_data_error_result(
-                    retmsg="Fail! Please select knowledgebase!")
+                    message="Fail! Please select knowledgebase!")
+
             dia = {
                 "id": get_uuid(),
                 "tenant_id": current_user.id,
@@ -96,35 +102,33 @@ def set_dialog():
                 "icon": icon
             }
             if not DialogService.save(**dia):
-                return get_data_error_result(retmsg="Fail to new a dialog!")
-            e, dia = DialogService.get_by_id(dia["id"])
-            if not e:
-                return get_data_error_result(retmsg="Fail to new a dialog!")
-            return get_json_result(data=dia.to_json())
+                return get_data_error_result(message="Fail to new a dialog!")
+            return get_json_result(data=dia)
         else:
             del req["dialog_id"]
             if "kb_names" in req:
                 del req["kb_names"]
             if not DialogService.update_by_id(dialog_id, req):
-                return get_data_error_result(retmsg="Dialog not found!")
+                return get_data_error_result(message="Dialog not found!")
             e, dia = DialogService.get_by_id(dialog_id)
             if not e:
-                return get_data_error_result(retmsg="Fail to update a dialog!")
+                return get_data_error_result(message="Fail to update a dialog!")
             dia = dia.to_dict()
+            dia.update(req)
             dia["kb_ids"], dia["kb_names"] = get_kb_names(dia["kb_ids"])
             return get_json_result(data=dia)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/get', methods=['GET'])
+@manager.route('/get', methods=['GET'])  # noqa: F821
 @login_required
 def get():
     dialog_id = request.args["dialog_id"]
     try:
         e, dia = DialogService.get_by_id(dialog_id)
         if not e:
-            return get_data_error_result(retmsg="Dialog not found!")
+            return get_data_error_result(message="Dialog not found!")
         dia = dia.to_dict()
         dia["kb_ids"], dia["kb_names"] = get_kb_names(dia["kb_ids"])
         return get_json_result(data=dia)
@@ -143,7 +147,7 @@ def get_kb_names(kb_ids):
     return ids, nms
 
 
-@manager.route('/list', methods=['GET'])
+@manager.route('/list', methods=['GET'])  # noqa: F821
 @login_required
 def list_dialogs():
     try:
@@ -160,7 +164,7 @@ def list_dialogs():
         return server_error_response(e)
 
 
-@manager.route('/rm', methods=['POST'])
+@manager.route('/rm', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("dialog_ids")
 def rm():
@@ -174,8 +178,8 @@ def rm():
                     break
             else:
                 return get_json_result(
-                    data=False, retmsg=f'Only owner of dialog authorized for this operation.',
-                    retcode=RetCode.OPERATING_ERROR)
+                    data=False, message='Only owner of dialog authorized for this operation.',
+                    code=settings.RetCode.OPERATING_ERROR)
             dialog_list.append({"id": id,"status":StatusEnum.INVALID.value})
         DialogService.update_many_by_id(dialog_list)
         return get_json_result(data=True)

api/apps/document_app.py

@@ -13,52 +13,59 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License
 #
+import json
+import os.path
 import pathlib
 import re
 
 import flask
-from elasticsearch_dsl import Q
 from flask import request
 from flask_login import login_required, current_user
 
-from api.db.db_models import Task, File
+from deepdoc.parser.html_parser import RAGFlowHtmlParser
+from rag.nlp import search
+
+from api.db import FileType, TaskStatus, ParserType, FileSource
+from api.db.db_models import File, Task
 from api.db.services.file2document_service import File2DocumentService
 from api.db.services.file_service import FileService
-from api.db.services.task_service import TaskService, queue_tasks
+from api.db.services.task_service import queue_tasks
 from api.db.services.user_service import UserTenantService
-from rag.nlp import search
-from rag.utils.es_conn import ELASTICSEARCH
 from api.db.services import duplicate_name
 from api.db.services.knowledgebase_service import KnowledgebaseService
-from api.utils.api_utils import server_error_response, get_data_error_result, validate_request
-from api.utils import get_uuid
-from api.db import FileType, TaskStatus, ParserType, FileSource
+from api.db.services.task_service import TaskService
 from api.db.services.document_service import DocumentService, doc_upload_and_parse
-from api.settings import RetCode
+from api.utils.api_utils import (
+    server_error_response,
+    get_data_error_result,
+    validate_request,
+)
+from api.utils import get_uuid
+from api import settings
 from api.utils.api_utils import get_json_result
 from rag.utils.storage_factory import STORAGE_IMPL
-from api.utils.file_utils import filename_type, thumbnail
+from api.utils.file_utils import filename_type, thumbnail, get_project_base_directory
 from api.utils.web_utils import html2pdf, is_valid_url
-from api.contants import IMG_BASE64_PREFIX
+from api.constants import IMG_BASE64_PREFIX
 
 
-@manager.route('/upload', methods=['POST'])
+@manager.route('/upload', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("kb_id")
 def upload():
     kb_id = request.form.get("kb_id")
     if not kb_id:
         return get_json_result(
-            data=False, retmsg='Lack of "KB ID"', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='Lack of "KB ID"', code=settings.RetCode.ARGUMENT_ERROR)
     if 'file' not in request.files:
         return get_json_result(
-            data=False, retmsg='No file part!', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='No file part!', code=settings.RetCode.ARGUMENT_ERROR)
 
     file_objs = request.files.getlist('file')
     for file_obj in file_objs:
         if file_obj.filename == '':
             return get_json_result(
-                data=False, retmsg='No file selected!', retcode=RetCode.ARGUMENT_ERROR)
+                data=False, message='No file selected!', code=settings.RetCode.ARGUMENT_ERROR)
 
     e, kb = KnowledgebaseService.get_by_id(kb_id)
     if not e:
@@ -67,29 +74,30 @@ def upload():
     err, _ = FileService.upload_document(kb, file_objs, current_user.id)
     if err:
         return get_json_result(
-            data=False, retmsg="\n".join(err), retcode=RetCode.SERVER_ERROR)
+            data=False, message="\n".join(err), code=settings.RetCode.SERVER_ERROR)
     return get_json_result(data=True)
 
 
-@manager.route('/web_crawl', methods=['POST'])
+@manager.route('/web_crawl', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("kb_id", "name", "url")
 def web_crawl():
     kb_id = request.form.get("kb_id")
     if not kb_id:
         return get_json_result(
-            data=False, retmsg='Lack of "KB ID"', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='Lack of "KB ID"', code=settings.RetCode.ARGUMENT_ERROR)
     name = request.form.get("name")
     url = request.form.get("url")
     if not is_valid_url(url):
         return get_json_result(
-            data=False, retmsg='The URL format is invalid', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='The URL format is invalid', code=settings.RetCode.ARGUMENT_ERROR)
     e, kb = KnowledgebaseService.get_by_id(kb_id)
     if not e:
         raise LookupError("Can't find this knowledgebase!")
 
     blob = html2pdf(url)
-    if not blob: return server_error_response(ValueError("Download failure."))
+    if not blob:
+        return server_error_response(ValueError("Download failure."))
 
     root_folder = FileService.get_root_folder(current_user.id)
     pf_id = root_folder["id"]
@@ -137,7 +145,7 @@ def web_crawl():
     return get_json_result(data=True)
 
 
-@manager.route('/create', methods=['POST'])
+@manager.route('/create', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("name", "kb_id")
 def create():
@@ -145,17 +153,17 @@ def create():
     kb_id = req["kb_id"]
     if not kb_id:
         return get_json_result(
-            data=False, retmsg='Lack of "KB ID"', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='Lack of "KB ID"', code=settings.RetCode.ARGUMENT_ERROR)
 
     try:
         e, kb = KnowledgebaseService.get_by_id(kb_id)
         if not e:
             return get_data_error_result(
-                retmsg="Can't find this knowledgebase!")
+                message="Can't find this knowledgebase!")
 
         if DocumentService.query(name=req["name"], kb_id=kb_id):
             return get_data_error_result(
-                retmsg="Duplicated document name in the same knowledgebase.")
+                message="Duplicated document name in the same knowledgebase.")
 
         doc = DocumentService.insert({
             "id": get_uuid(),
@@ -173,13 +181,13 @@ def create():
         return server_error_response(e)
 
 
-@manager.route('/list', methods=['GET'])
+@manager.route('/list', methods=['GET'])  # noqa: F821
 @login_required
 def list_docs():
     kb_id = request.args.get("kb_id")
     if not kb_id:
         return get_json_result(
-            data=False, retmsg='Lack of "KB ID"', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='Lack of "KB ID"', code=settings.RetCode.ARGUMENT_ERROR)
     tenants = UserTenantService.query(user_id=current_user.id)
     for tenant in tenants:
         if KnowledgebaseService.query(
@@ -187,8 +195,8 @@ def list_docs():
             break
     else:
         return get_json_result(
-            data=False, retmsg=f'Only owner of knowledgebase authorized for this operation.',
-            retcode=RetCode.OPERATING_ERROR)
+            data=False, message='Only owner of knowledgebase authorized for this operation.',
+            code=settings.RetCode.OPERATING_ERROR)
     keywords = request.args.get("keywords", "")
 
     page_number = int(request.args.get("page", 1))
@@ -208,7 +216,7 @@ def list_docs():
         return server_error_response(e)
 
 
-@manager.route('/infos', methods=['POST'])
+@manager.route('/infos', methods=['POST'])  # noqa: F821
 @login_required
 def docinfos():
     req = request.json
@@ -217,20 +225,20 @@ def docinfos():
         if not DocumentService.accessible(doc_id, current_user.id):
             return get_json_result(
                 data=False,
-                retmsg='No authorization.',
-                retcode=RetCode.AUTHENTICATION_ERROR
+                message='No authorization.',
+                code=settings.RetCode.AUTHENTICATION_ERROR
             )
     docs = DocumentService.get_by_ids(doc_ids)
     return get_json_result(data=list(docs.dicts()))
 
 
-@manager.route('/thumbnails', methods=['GET'])
+@manager.route('/thumbnails', methods=['GET'])  # noqa: F821
 # @login_required
 def thumbnails():
     doc_ids = request.args.get("doc_ids").split(",")
     if not doc_ids:
         return get_json_result(
-            data=False, retmsg='Lack of "Document ID"', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='Lack of "Document ID"', code=settings.RetCode.ARGUMENT_ERROR)
 
     try:
         docs = DocumentService.get_thumbnails(doc_ids)
@@ -244,7 +252,7 @@ def thumbnails():
         return server_error_response(e)
 
 
-@manager.route('/change_status', methods=['POST'])
+@manager.route('/change_status', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("doc_id", "status")
 def change_status():
@@ -252,60 +260,52 @@ def change_status():
     if str(req["status"]) not in ["0", "1"]:
         return get_json_result(
             data=False,
-            retmsg='"Status" must be either 0 or 1!',
-            retcode=RetCode.ARGUMENT_ERROR)
+            message='"Status" must be either 0 or 1!',
+            code=settings.RetCode.ARGUMENT_ERROR)
 
     if not DocumentService.accessible(req["doc_id"], current_user.id):
         return get_json_result(
             data=False,
-            retmsg='No authorization.',
-            retcode=RetCode.AUTHENTICATION_ERROR)
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR)
 
     try:
         e, doc = DocumentService.get_by_id(req["doc_id"])
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
         e, kb = KnowledgebaseService.get_by_id(doc.kb_id)
         if not e:
             return get_data_error_result(
-                retmsg="Can't find this knowledgebase!")
+                message="Can't find this knowledgebase!")
 
         if not DocumentService.update_by_id(
                 req["doc_id"], {"status": str(req["status"])}):
             return get_data_error_result(
-                retmsg="Database error (Document update)!")
+                message="Database error (Document update)!")
 
-        if str(req["status"]) == "0":
-            ELASTICSEARCH.updateScriptByQuery(Q("term", doc_id=req["doc_id"]),
-                                              scripts="ctx._source.available_int=0;",
-                                              idxnm=search.index_name(
-                                                  kb.tenant_id)
-                                              )
-        else:
-            ELASTICSEARCH.updateScriptByQuery(Q("term", doc_id=req["doc_id"]),
-                                              scripts="ctx._source.available_int=1;",
-                                              idxnm=search.index_name(
-                                                  kb.tenant_id)
-                                              )
+        status = int(req["status"])
+        settings.docStoreConn.update({"doc_id": req["doc_id"]}, {"available_int": status},
+                                     search.index_name(kb.tenant_id), doc.kb_id)
         return get_json_result(data=True)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/rm', methods=['POST'])
+@manager.route('/rm', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("doc_id")
 def rm():
     req = request.json
     doc_ids = req["doc_id"]
-    if isinstance(doc_ids, str): doc_ids = [doc_ids]
+    if isinstance(doc_ids, str):
+        doc_ids = [doc_ids]
 
     for doc_id in doc_ids:
         if not DocumentService.accessible4deletion(doc_id, current_user.id):
             return get_json_result(
                 data=False,
-                retmsg='No authorization.',
-                retcode=RetCode.AUTHENTICATION_ERROR
+                message='No authorization.',
+                code=settings.RetCode.AUTHENTICATION_ERROR
             )
 
     root_folder = FileService.get_root_folder(current_user.id)
@@ -316,16 +316,17 @@ def rm():
         try:
             e, doc = DocumentService.get_by_id(doc_id)
             if not e:
-                return get_data_error_result(retmsg="Document not found!")
+                return get_data_error_result(message="Document not found!")
             tenant_id = DocumentService.get_tenant_id(doc_id)
             if not tenant_id:
-                return get_data_error_result(retmsg="Tenant not found!")
+                return get_data_error_result(message="Tenant not found!")
 
             b, n = File2DocumentService.get_storage_address(doc_id=doc_id)
 
+            TaskService.filter_delete([Task.doc_id == doc_id])
             if not DocumentService.remove_document(doc, tenant_id):
                 return get_data_error_result(
-                    retmsg="Database error (Document removal)!")
+                    message="Database error (Document removal)!")
 
             f2d = File2DocumentService.get_by_document_id(doc_id)
             FileService.filter_delete([File.source_type == FileSource.KNOWLEDGEBASE, File.id == f2d[0].file_id])
@@ -336,12 +337,12 @@ def rm():
             errors += str(e)
 
     if errors:
-        return get_json_result(data=False, retmsg=errors, retcode=RetCode.SERVER_ERROR)
+        return get_json_result(data=False, message=errors, code=settings.RetCode.SERVER_ERROR)
 
     return get_json_result(data=True)
 
 
-@manager.route('/run', methods=['POST'])
+@manager.route('/run', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("doc_ids", "run")
 def run():
@@ -350,26 +351,29 @@ def run():
         if not DocumentService.accessible(doc_id, current_user.id):
             return get_json_result(
                 data=False,
-                retmsg='No authorization.',
-                retcode=RetCode.AUTHENTICATION_ERROR
+                message='No authorization.',
+                code=settings.RetCode.AUTHENTICATION_ERROR
             )
     try:
         for id in req["doc_ids"]:
             info = {"run": str(req["run"]), "progress": 0}
-            if str(req["run"]) == TaskStatus.RUNNING.value:
+            if str(req["run"]) == TaskStatus.RUNNING.value and req.get("delete", False):
                 info["progress_msg"] = ""
                 info["chunk_num"] = 0
                 info["token_num"] = 0
             DocumentService.update_by_id(id, info)
-            # if str(req["run"]) == TaskStatus.CANCEL.value:
             tenant_id = DocumentService.get_tenant_id(id)
             if not tenant_id:
-                return get_data_error_result(retmsg="Tenant not found!")
-            ELASTICSEARCH.deleteByQuery(
-                Q("match", doc_id=id), idxnm=search.index_name(tenant_id))
+                return get_data_error_result(message="Tenant not found!")
+            e, doc = DocumentService.get_by_id(id)
+            if not e:
+                return get_data_error_result(message="Document not found!")
+            if req.get("delete", False):
+                TaskService.filter_delete([Task.doc_id == id])
+                if settings.docStoreConn.indexExist(search.index_name(tenant_id), doc.kb_id):
+                    settings.docStoreConn.delete({"doc_id": id}, search.index_name(tenant_id), doc.kb_id)
 
             if str(req["run"]) == TaskStatus.RUNNING.value:
-                TaskService.filter_delete([Task.doc_id == id])
                 e, doc = DocumentService.get_by_id(id)
                 doc = doc.to_dict()
                 doc["tenant_id"] = tenant_id
@@ -381,7 +385,7 @@ def run():
         return server_error_response(e)
 
 
-@manager.route('/rename', methods=['POST'])
+@manager.route('/rename', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("doc_id", "name")
 def rename():
@@ -389,28 +393,28 @@ def rename():
     if not DocumentService.accessible(req["doc_id"], current_user.id):
         return get_json_result(
             data=False,
-            retmsg='No authorization.',
-            retcode=RetCode.AUTHENTICATION_ERROR
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
         )
     try:
         e, doc = DocumentService.get_by_id(req["doc_id"])
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
         if pathlib.Path(req["name"].lower()).suffix != pathlib.Path(
                 doc.name.lower()).suffix:
             return get_json_result(
                 data=False,
-                retmsg="The extension of file can't be changed",
-                retcode=RetCode.ARGUMENT_ERROR)
+                message="The extension of file can't be changed",
+                code=settings.RetCode.ARGUMENT_ERROR)
         for d in DocumentService.query(name=req["name"], kb_id=doc.kb_id):
             if d.name == req["name"]:
                 return get_data_error_result(
-                    retmsg="Duplicated document name in the same knowledgebase.")
+                    message="Duplicated document name in the same knowledgebase.")
 
         if not DocumentService.update_by_id(
                 req["doc_id"], {"name": req["name"]}):
             return get_data_error_result(
-                retmsg="Database error (Document rename)!")
+                message="Database error (Document rename)!")
 
         informs = File2DocumentService.get_by_document_id(req["doc_id"])
         if informs:
@@ -422,13 +426,13 @@ def rename():
         return server_error_response(e)
 
 
-@manager.route('/get/<doc_id>', methods=['GET'])
+@manager.route('/get/<doc_id>', methods=['GET'])  # noqa: F821
 # @login_required
 def get(doc_id):
     try:
         e, doc = DocumentService.get_by_id(doc_id)
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
 
         b, n = File2DocumentService.get_storage_address(doc_id=doc_id)
         response = flask.make_response(STORAGE_IMPL.get(b, n))
@@ -447,7 +451,7 @@ def get(doc_id):
         return server_error_response(e)
 
 
-@manager.route('/change_parser', methods=['POST'])
+@manager.route('/change_parser', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("doc_id", "parser_id")
 def change_parser():
@@ -456,13 +460,13 @@ def change_parser():
     if not DocumentService.accessible(req["doc_id"], current_user.id):
         return get_json_result(
             data=False,
-            retmsg='No authorization.',
-            retcode=RetCode.AUTHENTICATION_ERROR
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
         )
     try:
         e, doc = DocumentService.get_by_id(req["doc_id"])
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
         if doc.parser_id.lower() == req["parser_id"].lower():
             if "parser_config" in req:
                 if req["parser_config"] == doc.parser_config:
@@ -473,35 +477,38 @@ def change_parser():
         if ((doc.type == FileType.VISUAL and req["parser_id"] != "picture")
                 or (re.search(
                     r"\.(ppt|pptx|pages)$", doc.name) and req["parser_id"] != "presentation")):
-            return get_data_error_result(retmsg="Not supported yet!")
+            return get_data_error_result(message="Not supported yet!")
 
         e = DocumentService.update_by_id(doc.id,
                                          {"parser_id": req["parser_id"], "progress": 0, "progress_msg": "",
                                           "run": TaskStatus.UNSTART.value})
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
         if "parser_config" in req:
             DocumentService.update_parser_config(doc.id, req["parser_config"])
         if doc.token_num > 0:
             e = DocumentService.increment_chunk_num(doc.id, doc.kb_id, doc.token_num * -1, doc.chunk_num * -1,
                                                     doc.process_duation * -1)
             if not e:
-                return get_data_error_result(retmsg="Document not found!")
+                return get_data_error_result(message="Document not found!")
             tenant_id = DocumentService.get_tenant_id(req["doc_id"])
             if not tenant_id:
-                return get_data_error_result(retmsg="Tenant not found!")
-            ELASTICSEARCH.deleteByQuery(
-                Q("match", doc_id=doc.id), idxnm=search.index_name(tenant_id))
+                return get_data_error_result(message="Tenant not found!")
+            if settings.docStoreConn.indexExist(search.index_name(tenant_id), doc.kb_id):
+                settings.docStoreConn.delete({"doc_id": doc.id}, search.index_name(tenant_id), doc.kb_id)
 
         return get_json_result(data=True)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/image/<image_id>', methods=['GET'])
+@manager.route('/image/<image_id>', methods=['GET'])  # noqa: F821
 # @login_required
 def get_image(image_id):
     try:
+        arr = image_id.split("-")
+        if len(arr) != 2:
+            return get_data_error_result(message="Image not found.")
         bkt, nm = image_id.split("-")
         response = flask.make_response(STORAGE_IMPL.get(bkt, nm))
         response.headers.set('Content-Type', 'image/JPEG')
@@ -510,20 +517,115 @@ def get_image(image_id):
         return server_error_response(e)
 
 
-@manager.route('/upload_and_parse', methods=['POST'])
+@manager.route('/upload_and_parse', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("conversation_id")
 def upload_and_parse():
     if 'file' not in request.files:
         return get_json_result(
-            data=False, retmsg='No file part!', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='No file part!', code=settings.RetCode.ARGUMENT_ERROR)
 
     file_objs = request.files.getlist('file')
     for file_obj in file_objs:
         if file_obj.filename == '':
             return get_json_result(
-                data=False, retmsg='No file selected!', retcode=RetCode.ARGUMENT_ERROR)
+                data=False, message='No file selected!', code=settings.RetCode.ARGUMENT_ERROR)
 
     doc_ids = doc_upload_and_parse(request.form.get("conversation_id"), file_objs, current_user.id)
 
     return get_json_result(data=doc_ids)
+
+
+@manager.route('/parse', methods=['POST'])  # noqa: F821
+@login_required
+def parse():
+    url = request.json.get("url") if request.json else ""
+    if url:
+        if not is_valid_url(url):
+            return get_json_result(
+                data=False, message='The URL format is invalid', code=settings.RetCode.ARGUMENT_ERROR)
+        download_path = os.path.join(get_project_base_directory(), "logs/downloads")
+        os.makedirs(download_path, exist_ok=True)
+        from seleniumwire.webdriver import Chrome, ChromeOptions
+        options = ChromeOptions()
+        options.add_argument('--headless')
+        options.add_argument('--disable-gpu')
+        options.add_argument('--no-sandbox')
+        options.add_argument('--disable-dev-shm-usage')
+        options.add_experimental_option('prefs', {
+            'download.default_directory': download_path,
+            'download.prompt_for_download': False,
+            'download.directory_upgrade': True,
+            'safebrowsing.enabled': True
+        })
+        driver = Chrome(options=options)
+        driver.get(url)
+        res_headers = [r.response.headers for r in driver.requests if r and r.response]
+        if len(res_headers) > 1:
+            sections = RAGFlowHtmlParser().parser_txt(driver.page_source)
+            driver.quit()
+            return get_json_result(data="\n".join(sections))
+
+        class File:
+            filename: str
+            filepath: str
+
+            def __init__(self, filename, filepath):
+                self.filename = filename
+                self.filepath = filepath
+
+            def read(self):
+                with open(self.filepath, "rb") as f:
+                    return f.read()
+
+        r = re.search(r"filename=\"([^\"]+)\"", str(res_headers))
+        if not r or not r.group(1):
+            return get_json_result(
+                data=False, message="Can't not identify downloaded file", code=settings.RetCode.ARGUMENT_ERROR)
+        f = File(r.group(1), os.path.join(download_path, r.group(1)))
+        txt = FileService.parse_docs([f], current_user.id)
+        return get_json_result(data=txt)
+
+    if 'file' not in request.files:
+        return get_json_result(
+            data=False, message='No file part!', code=settings.RetCode.ARGUMENT_ERROR)
+
+    file_objs = request.files.getlist('file')
+    txt = FileService.parse_docs(file_objs, current_user.id)
+
+    return get_json_result(data=txt)
+
+
+@manager.route('/set_meta', methods=['POST'])  # noqa: F821
+@login_required
+@validate_request("doc_id", "meta")
+def set_meta():
+    req = request.json
+    if not DocumentService.accessible(req["doc_id"], current_user.id):
+        return get_json_result(
+            data=False,
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
+        )
+    try:
+        meta = json.loads(req["meta"])
+    except Exception as e:
+        return get_json_result(
+            data=False, message=f'Json syntax error: {e}', code=settings.RetCode.ARGUMENT_ERROR)
+    if not isinstance(meta, dict):
+        return get_json_result(
+            data=False, message='Meta data should be in Json map format, like {"key": "value"}', code=settings.RetCode.ARGUMENT_ERROR)
+
+    try:
+        e, doc = DocumentService.get_by_id(req["doc_id"])
+        if not e:
+            return get_data_error_result(message="Document not found!")
+
+        if not DocumentService.update_by_id(
+                req["doc_id"], {"meta_fields": meta}):
+            return get_data_error_result(
+                message="Database error (meta updates)!")
+
+        return get_json_result(data=True)
+    except Exception as e:
+        return server_error_response(e)

api/apps/file2document_app.py

@@ -13,9 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License
 #
-from elasticsearch_dsl import Q
 
-from api.db.db_models import File2Document
 from api.db.services.file2document_service import File2DocumentService
 from api.db.services.file_service import FileService
 
@@ -26,13 +24,11 @@ from api.utils.api_utils import server_error_response, get_data_error_result, va
 from api.utils import get_uuid
 from api.db import FileType
 from api.db.services.document_service import DocumentService
-from api.settings import RetCode
+from api import settings
 from api.utils.api_utils import get_json_result
-from rag.nlp import search
-from rag.utils.es_conn import ELASTICSEARCH
 
 
-@manager.route('/convert', methods=['POST'])
+@manager.route('/convert', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("file_ids", "kb_ids")
 def convert():
@@ -54,13 +50,13 @@ def convert():
                     doc_id = inform.document_id
                     e, doc = DocumentService.get_by_id(doc_id)
                     if not e:
-                        return get_data_error_result(retmsg="Document not found!")
+                        return get_data_error_result(message="Document not found!")
                     tenant_id = DocumentService.get_tenant_id(doc_id)
                     if not tenant_id:
-                        return get_data_error_result(retmsg="Tenant not found!")
+                        return get_data_error_result(message="Tenant not found!")
                     if not DocumentService.remove_document(doc, tenant_id):
                         return get_data_error_result(
-                            retmsg="Database error (Document removal)!")
+                            message="Database error (Document removal)!")
                 File2DocumentService.delete_by_file_id(id)
 
                 # insert
@@ -68,11 +64,11 @@ def convert():
                     e, kb = KnowledgebaseService.get_by_id(kb_id)
                     if not e:
                         return get_data_error_result(
-                            retmsg="Can't find this knowledgebase!")
+                            message="Can't find this knowledgebase!")
                     e, file = FileService.get_by_id(id)
                     if not e:
                         return get_data_error_result(
-                            retmsg="Can't find this file!")
+                            message="Can't find this file!")
 
                     doc = DocumentService.insert({
                         "id": get_uuid(),
@@ -96,7 +92,7 @@ def convert():
         return server_error_response(e)
 
 
-@manager.route('/rm', methods=['POST'])
+@manager.route('/rm', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("file_ids")
 def rm():
@@ -104,26 +100,26 @@ def rm():
     file_ids = req["file_ids"]
     if not file_ids:
         return get_json_result(
-            data=False, retmsg='Lack of "Files ID"', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='Lack of "Files ID"', code=settings.RetCode.ARGUMENT_ERROR)
     try:
         for file_id in file_ids:
             informs = File2DocumentService.get_by_file_id(file_id)
             if not informs:
-                return get_data_error_result(retmsg="Inform not found!")
+                return get_data_error_result(message="Inform not found!")
             for inform in informs:
                 if not inform:
-                    return get_data_error_result(retmsg="Inform not found!")
+                    return get_data_error_result(message="Inform not found!")
                 File2DocumentService.delete_by_file_id(file_id)
                 doc_id = inform.document_id
                 e, doc = DocumentService.get_by_id(doc_id)
                 if not e:
-                    return get_data_error_result(retmsg="Document not found!")
+                    return get_data_error_result(message="Document not found!")
                 tenant_id = DocumentService.get_tenant_id(doc_id)
                 if not tenant_id:
-                    return get_data_error_result(retmsg="Tenant not found!")
+                    return get_data_error_result(message="Tenant not found!")
                 if not DocumentService.remove_document(doc, tenant_id):
                     return get_data_error_result(
-                        retmsg="Database error (Document removal)!")
+                        message="Database error (Document removal)!")
         return get_json_result(data=True)
     except Exception as e:
         return server_error_response(e)

api/apps/file_app.py

@@ -18,7 +18,6 @@ import pathlib
 import re
 
 import flask
-from elasticsearch_dsl import Q
 from flask import request
 from flask_login import login_required, current_user
 
@@ -29,15 +28,13 @@ from api.utils import get_uuid
 from api.db import FileType, FileSource
 from api.db.services import duplicate_name
 from api.db.services.file_service import FileService
-from api.settings import RetCode
+from api import settings
 from api.utils.api_utils import get_json_result
 from api.utils.file_utils import filename_type
-from rag.nlp import search
-from rag.utils.es_conn import ELASTICSEARCH
 from rag.utils.storage_factory import STORAGE_IMPL
 
 
-@manager.route('/upload', methods=['POST'])
+@manager.route('/upload', methods=['POST'])  # noqa: F821
 @login_required
 # @validate_request("parent_id")
 def upload():
@@ -49,24 +46,24 @@ def upload():
 
     if 'file' not in request.files:
         return get_json_result(
-            data=False, retmsg='No file part!', retcode=RetCode.ARGUMENT_ERROR)
+            data=False, message='No file part!', code=settings.RetCode.ARGUMENT_ERROR)
     file_objs = request.files.getlist('file')
 
     for file_obj in file_objs:
         if file_obj.filename == '':
             return get_json_result(
-                data=False, retmsg='No file selected!', retcode=RetCode.ARGUMENT_ERROR)
+                data=False, message='No file selected!', code=settings.RetCode.ARGUMENT_ERROR)
     file_res = []
     try:
         for file_obj in file_objs:
             e, file = FileService.get_by_id(pf_id)
             if not e:
                 return get_data_error_result(
-                    retmsg="Can't find this folder!")
+                    message="Can't find this folder!")
             MAX_FILE_NUM_PER_USER = int(os.environ.get('MAX_FILE_NUM_PER_USER', 0))
             if MAX_FILE_NUM_PER_USER > 0 and DocumentService.get_doc_count(current_user.id) >= MAX_FILE_NUM_PER_USER:
                 return get_data_error_result(
-                    retmsg="Exceed the maximum file number of a free user!")
+                    message="Exceed the maximum file number of a free user!")
 
             # split file name path
             if not file_obj.filename:
@@ -85,13 +82,13 @@ def upload():
             if file_len != len_id_list:
                 e, file = FileService.get_by_id(file_id_list[len_id_list - 1])
                 if not e:
-                    return get_data_error_result(retmsg="Folder not found!")
+                    return get_data_error_result(message="Folder not found!")
                 last_folder = FileService.create_folder(file, file_id_list[len_id_list - 1], file_obj_names,
                                                         len_id_list)
             else:
                 e, file = FileService.get_by_id(file_id_list[len_id_list - 2])
                 if not e:
-                    return get_data_error_result(retmsg="Folder not found!")
+                    return get_data_error_result(message="Folder not found!")
                 last_folder = FileService.create_folder(file, file_id_list[len_id_list - 2], file_obj_names,
                                                         len_id_list)
 
@@ -123,7 +120,7 @@ def upload():
         return server_error_response(e)
 
 
-@manager.route('/create', methods=['POST'])
+@manager.route('/create', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("name")
 def create():
@@ -137,10 +134,10 @@ def create():
     try:
         if not FileService.is_parent_folder_exist(pf_id):
             return get_json_result(
-                data=False, retmsg="Parent Folder Doesn't Exist!", retcode=RetCode.OPERATING_ERROR)
+                data=False, message="Parent Folder Doesn't Exist!", code=settings.RetCode.OPERATING_ERROR)
         if FileService.query(name=req["name"], parent_id=pf_id):
             return get_data_error_result(
-                retmsg="Duplicated folder name in the same folder.")
+                message="Duplicated folder name in the same folder.")
 
         if input_file_type == FileType.FOLDER.value:
             file_type = FileType.FOLDER.value
@@ -163,7 +160,7 @@ def create():
         return server_error_response(e)
 
 
-@manager.route('/list', methods=['GET'])
+@manager.route('/list', methods=['GET'])  # noqa: F821
 @login_required
 def list_files():
     pf_id = request.args.get("parent_id")
@@ -181,21 +178,21 @@ def list_files():
     try:
         e, file = FileService.get_by_id(pf_id)
         if not e:
-            return get_data_error_result(retmsg="Folder not found!")
+            return get_data_error_result(message="Folder not found!")
 
         files, total = FileService.get_by_pf_id(
             current_user.id, pf_id, page_number, items_per_page, orderby, desc, keywords)
 
         parent_folder = FileService.get_parent_folder(pf_id)
         if not FileService.get_parent_folder(pf_id):
-            return get_json_result(retmsg="File not found!")
+            return get_json_result(message="File not found!")
 
         return get_json_result(data={"total": total, "files": files, "parent_folder": parent_folder.to_json()})
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/root_folder', methods=['GET'])
+@manager.route('/root_folder', methods=['GET'])  # noqa: F821
 @login_required
 def get_root_folder():
     try:
@@ -205,14 +202,14 @@ def get_root_folder():
         return server_error_response(e)
 
 
-@manager.route('/parent_folder', methods=['GET'])
+@manager.route('/parent_folder', methods=['GET'])  # noqa: F821
 @login_required
 def get_parent_folder():
     file_id = request.args.get("file_id")
     try:
         e, file = FileService.get_by_id(file_id)
         if not e:
-            return get_data_error_result(retmsg="Folder not found!")
+            return get_data_error_result(message="Folder not found!")
 
         parent_folder = FileService.get_parent_folder(file_id)
         return get_json_result(data={"parent_folder": parent_folder.to_json()})
@@ -220,14 +217,14 @@ def get_parent_folder():
         return server_error_response(e)
 
 
-@manager.route('/all_parent_folder', methods=['GET'])
+@manager.route('/all_parent_folder', methods=['GET'])  # noqa: F821
 @login_required
 def get_all_parent_folders():
     file_id = request.args.get("file_id")
     try:
         e, file = FileService.get_by_id(file_id)
         if not e:
-            return get_data_error_result(retmsg="Folder not found!")
+            return get_data_error_result(message="Folder not found!")
 
         parent_folders = FileService.get_all_parent_folders(file_id)
         parent_folders_res = []
@@ -238,7 +235,7 @@ def get_all_parent_folders():
         return server_error_response(e)
 
 
-@manager.route('/rm', methods=['POST'])
+@manager.route('/rm', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("file_ids")
 def rm():
@@ -248,9 +245,9 @@ def rm():
         for file_id in file_ids:
             e, file = FileService.get_by_id(file_id)
             if not e:
-                return get_data_error_result(retmsg="File or Folder not found!")
+                return get_data_error_result(message="File or Folder not found!")
             if not file.tenant_id:
-                return get_data_error_result(retmsg="Tenant not found!")
+                return get_data_error_result(message="Tenant not found!")
             if file.source_type == FileSource.KNOWLEDGEBASE:
                 continue
 
@@ -259,13 +256,13 @@ def rm():
                 for inner_file_id in file_id_list:
                     e, file = FileService.get_by_id(inner_file_id)
                     if not e:
-                        return get_data_error_result(retmsg="File not found!")
+                        return get_data_error_result(message="File not found!")
                     STORAGE_IMPL.rm(file.parent_id, file.location)
                 FileService.delete_folder_by_pf_id(current_user.id, file_id)
             else:
                 if not FileService.delete(file):
                     return get_data_error_result(
-                        retmsg="Database error (File removal)!")
+                        message="Database error (File removal)!")
 
             # delete file2document
             informs = File2DocumentService.get_by_file_id(file_id)
@@ -273,13 +270,13 @@ def rm():
                 doc_id = inform.document_id
                 e, doc = DocumentService.get_by_id(doc_id)
                 if not e:
-                    return get_data_error_result(retmsg="Document not found!")
+                    return get_data_error_result(message="Document not found!")
                 tenant_id = DocumentService.get_tenant_id(doc_id)
                 if not tenant_id:
-                    return get_data_error_result(retmsg="Tenant not found!")
+                    return get_data_error_result(message="Tenant not found!")
                 if not DocumentService.remove_document(doc, tenant_id):
                     return get_data_error_result(
-                        retmsg="Database error (Document removal)!")
+                        message="Database error (Document removal)!")
             File2DocumentService.delete_by_file_id(file_id)
 
         return get_json_result(data=True)
@@ -287,7 +284,7 @@ def rm():
         return server_error_response(e)
 
 
-@manager.route('/rename', methods=['POST'])
+@manager.route('/rename', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("file_id", "name")
 def rename():
@@ -295,45 +292,50 @@ def rename():
     try:
         e, file = FileService.get_by_id(req["file_id"])
         if not e:
-            return get_data_error_result(retmsg="File not found!")
+            return get_data_error_result(message="File not found!")
         if file.type != FileType.FOLDER.value \
             and pathlib.Path(req["name"].lower()).suffix != pathlib.Path(
                 file.name.lower()).suffix:
             return get_json_result(
                 data=False,
-                retmsg="The extension of file can't be changed",
-                retcode=RetCode.ARGUMENT_ERROR)
+                message="The extension of file can't be changed",
+                code=settings.RetCode.ARGUMENT_ERROR)
         for file in FileService.query(name=req["name"], pf_id=file.parent_id):
             if file.name == req["name"]:
                 return get_data_error_result(
-                    retmsg="Duplicated file name in the same folder.")
+                    message="Duplicated file name in the same folder.")
 
         if not FileService.update_by_id(
                 req["file_id"], {"name": req["name"]}):
             return get_data_error_result(
-                retmsg="Database error (File rename)!")
+                message="Database error (File rename)!")
 
         informs = File2DocumentService.get_by_file_id(req["file_id"])
         if informs:
             if not DocumentService.update_by_id(
                     informs[0].document_id, {"name": req["name"]}):
                 return get_data_error_result(
-                    retmsg="Database error (Document rename)!")
+                    message="Database error (Document rename)!")
 
         return get_json_result(data=True)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/get/<file_id>', methods=['GET'])
-# @login_required
+@manager.route('/get/<file_id>', methods=['GET'])  # noqa: F821
+@login_required
 def get(file_id):
     try:
         e, file = FileService.get_by_id(file_id)
         if not e:
-            return get_data_error_result(retmsg="Document not found!")
+            return get_data_error_result(message="Document not found!")
+
+        blob = STORAGE_IMPL.get(file.parent_id, file.location)
+        if not blob:
             b, n = File2DocumentService.get_storage_address(file_id=file_id)
-        response = flask.make_response(STORAGE_IMPL.get(b, n))
+            blob = STORAGE_IMPL.get(b, n)
+
+        response = flask.make_response(blob)
         ext = re.search(r"\.([^.]+)$", file.name)
         if ext:
             if file.type == FileType.VISUAL.value:
@@ -348,7 +350,7 @@ def get(file_id):
         return server_error_response(e)
 
 
-@manager.route('/mv', methods=['POST'])
+@manager.route('/mv', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("src_file_ids", "dest_file_id")
 def move():
@@ -359,12 +361,12 @@ def move():
         for file_id in file_ids:
             e, file = FileService.get_by_id(file_id)
             if not e:
-                return get_data_error_result(retmsg="File or Folder not found!")
+                return get_data_error_result(message="File or Folder not found!")
             if not file.tenant_id:
-                return get_data_error_result(retmsg="Tenant not found!")
+                return get_data_error_result(message="Tenant not found!")
         fe, _ = FileService.get_by_id(parent_id)
         if not fe:
-            return get_data_error_result(retmsg="Parent Folder not found!")
+            return get_data_error_result(message="Parent Folder not found!")
         FileService.move_file(file_ids, parent_id)
         return get_json_result(data=True)
     except Exception as e:

api/apps/kb_app.py

@@ -13,6 +13,10 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import json
+import logging
+import os
+
 from flask import request
 from flask_login import login_required, current_user
 
@@ -21,24 +25,36 @@ from api.db.services.document_service import DocumentService
 from api.db.services.file2document_service import File2DocumentService
 from api.db.services.file_service import FileService
 from api.db.services.user_service import TenantService, UserTenantService
-from api.utils.api_utils import server_error_response, get_data_error_result, validate_request
+from api.utils.api_utils import server_error_response, get_data_error_result, validate_request, not_allowed_parameters
 from api.utils import get_uuid
 from api.db import StatusEnum, FileSource
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.db_models import File
-from api.settings import RetCode
 from api.utils.api_utils import get_json_result
+from api import settings
+from rag.nlp import search
+from api.constants import DATASET_NAME_LIMIT
+from rag.settings import PAGERANK_FLD
 
 
-@manager.route('/create', methods=['post'])
+@manager.route('/create', methods=['post'])  # noqa: F821
 @login_required
 @validate_request("name")
 def create():
     req = request.json
-    req["name"] = req["name"].strip()
-    req["name"] = duplicate_name(
+    dataset_name = req["name"]
+    if not isinstance(dataset_name, str):
+        return get_data_error_result(message="Dataset name must be string.")
+    if dataset_name == "":
+        return get_data_error_result(message="Dataset name can't be empty.")
+    if len(dataset_name) >= DATASET_NAME_LIMIT:
+        return get_data_error_result(
+            message=f"Dataset name length is {len(dataset_name)} which is large than {DATASET_NAME_LIMIT}")
+
+    dataset_name = dataset_name.strip()
+    dataset_name = duplicate_name(
         KnowledgebaseService.query,
-        name=req["name"],
+        name=dataset_name,
         tenant_id=current_user.id,
         status=StatusEnum.VALID.value)
     try:
@@ -47,7 +63,7 @@ def create():
         req["created_by"] = current_user.id
         e, t = TenantService.get_by_id(current_user.id)
         if not e:
-            return get_data_error_result(retmsg="Tenant not found.")
+            return get_data_error_result(message="Tenant not found.")
         req["embd_id"] = t.embd_id
         if not KnowledgebaseService.save(**req):
             return get_data_error_result()
@@ -56,49 +72,70 @@ def create():
         return server_error_response(e)
 
 
-@manager.route('/update', methods=['post'])
+@manager.route('/update', methods=['post'])  # noqa: F821
 @login_required
 @validate_request("kb_id", "name", "description", "permission", "parser_id")
+@not_allowed_parameters("id", "tenant_id", "created_by", "create_time", "update_time", "create_date", "update_date", "created_by")
 def update():
     req = request.json
     req["name"] = req["name"].strip()
     if not KnowledgebaseService.accessible4deletion(req["kb_id"], current_user.id):
         return get_json_result(
             data=False,
-            retmsg='No authorization.',
-            retcode=RetCode.AUTHENTICATION_ERROR
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
         )
     try:
         if not KnowledgebaseService.query(
                 created_by=current_user.id, id=req["kb_id"]):
             return get_json_result(
-                data=False, retmsg=f'Only owner of knowledgebase authorized for this operation.', retcode=RetCode.OPERATING_ERROR)
+                data=False, message='Only owner of knowledgebase authorized for this operation.',
+                code=settings.RetCode.OPERATING_ERROR)
 
         e, kb = KnowledgebaseService.get_by_id(req["kb_id"])
         if not e:
             return get_data_error_result(
-                retmsg="Can't find this knowledgebase!")
+                message="Can't find this knowledgebase!")
+
+        if req.get("parser_id", "") == "tag" and os.environ.get('DOC_ENGINE', "elasticsearch") == "infinity":
+            return get_json_result(
+                data=False,
+                message='The chunk method Tag has not been supported by Infinity yet.',
+                code=settings.RetCode.OPERATING_ERROR
+            )
 
         if req["name"].lower() != kb.name.lower() \
-                and len(KnowledgebaseService.query(name=req["name"], tenant_id=current_user.id, status=StatusEnum.VALID.value)) > 1:
+                and len(
+            KnowledgebaseService.query(name=req["name"], tenant_id=current_user.id, status=StatusEnum.VALID.value)) > 1:
             return get_data_error_result(
-                retmsg="Duplicated knowledgebase name.")
+                message="Duplicated knowledgebase name.")
 
         del req["kb_id"]
         if not KnowledgebaseService.update_by_id(kb.id, req):
             return get_data_error_result()
 
+        if kb.pagerank != req.get("pagerank", 0):
+            if req.get("pagerank", 0) > 0:
+                settings.docStoreConn.update({"kb_id": kb.id}, {PAGERANK_FLD: req["pagerank"]},
+                                         search.index_name(kb.tenant_id), kb.id)
+            else:
+                # Elasticsearch requires PAGERANK_FLD be non-zero!
+                settings.docStoreConn.update({"exists": PAGERANK_FLD}, {"remove": PAGERANK_FLD},
+                                         search.index_name(kb.tenant_id), kb.id)
+
         e, kb = KnowledgebaseService.get_by_id(kb.id)
         if not e:
             return get_data_error_result(
-                retmsg="Database error (Knowledgebase rename)!")
+                message="Database error (Knowledgebase rename)!")
+        kb = kb.to_dict()
+        kb.update(req)
 
-        return get_json_result(data=kb.to_json())
+        return get_json_result(data=kb)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/detail', methods=['GET'])
+@manager.route('/detail', methods=['GET'])  # noqa: F821
 @login_required
 def detail():
     kb_id = request.args["kb_id"]
@@ -110,34 +147,37 @@ def detail():
                 break
         else:
             return get_json_result(
-                data=False, retmsg=f'Only owner of knowledgebase authorized for this operation.',
-                retcode=RetCode.OPERATING_ERROR)
+                data=False, message='Only owner of knowledgebase authorized for this operation.',
+                code=settings.RetCode.OPERATING_ERROR)
         kb = KnowledgebaseService.get_detail(kb_id)
         if not kb:
             return get_data_error_result(
-                retmsg="Can't find this knowledgebase!")
+                message="Can't find this knowledgebase!")
         return get_json_result(data=kb)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/list', methods=['GET'])
+@manager.route('/list', methods=['GET'])  # noqa: F821
 @login_required
 def list_kbs():
-    page_number = request.args.get("page", 1)
-    items_per_page = request.args.get("page_size", 150)
+    keywords = request.args.get("keywords", "")
+    page_number = int(request.args.get("page", 1))
+    items_per_page = int(request.args.get("page_size", 150))
+    parser_id = request.args.get("parser_id")
     orderby = request.args.get("orderby", "create_time")
     desc = request.args.get("desc", True)
     try:
         tenants = TenantService.get_joined_tenants_by_user_id(current_user.id)
-        kbs = KnowledgebaseService.get_by_tenant_ids(
-            [m["tenant_id"] for m in tenants], current_user.id, page_number, items_per_page, orderby, desc)
-        return get_json_result(data=kbs)
+        kbs, total = KnowledgebaseService.get_by_tenant_ids(
+            [m["tenant_id"] for m in tenants], current_user.id, page_number,
+            items_per_page, orderby, desc, keywords, parser_id)
+        return get_json_result(data={"kbs": kbs, "total": total})
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/rm', methods=['post'])
+@manager.route('/rm', methods=['post'])  # noqa: F821
 @login_required
 @validate_request("kb_id")
 def rm():
@@ -145,27 +185,139 @@ def rm():
     if not KnowledgebaseService.accessible4deletion(req["kb_id"], current_user.id):
         return get_json_result(
             data=False,
-            retmsg='No authorization.',
-            retcode=RetCode.AUTHENTICATION_ERROR
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
         )
     try:
         kbs = KnowledgebaseService.query(
             created_by=current_user.id, id=req["kb_id"])
         if not kbs:
             return get_json_result(
-                data=False, retmsg=f'Only owner of knowledgebase authorized for this operation.', retcode=RetCode.OPERATING_ERROR)
+                data=False, message='Only owner of knowledgebase authorized for this operation.',
+                code=settings.RetCode.OPERATING_ERROR)
 
         for doc in DocumentService.query(kb_id=req["kb_id"]):
             if not DocumentService.remove_document(doc, kbs[0].tenant_id):
                 return get_data_error_result(
-                    retmsg="Database error (Document removal)!")
+                    message="Database error (Document removal)!")
             f2d = File2DocumentService.get_by_document_id(doc.id)
+            if f2d:
                 FileService.filter_delete([File.source_type == FileSource.KNOWLEDGEBASE, File.id == f2d[0].file_id])
             File2DocumentService.delete_by_document_id(doc.id)
-
+        FileService.filter_delete(
+            [File.source_type == FileSource.KNOWLEDGEBASE, File.type == "folder", File.name == kbs[0].name])
         if not KnowledgebaseService.delete_by_id(req["kb_id"]):
             return get_data_error_result(
-                retmsg="Database error (Knowledgebase removal)!")
+                message="Database error (Knowledgebase removal)!")
+        for kb in kbs:
+            settings.docStoreConn.delete({"kb_id": kb.id}, search.index_name(kb.tenant_id), kb.id)
+            settings.docStoreConn.deleteIdx(search.index_name(kb.tenant_id), kb.id)
         return get_json_result(data=True)
     except Exception as e:
         return server_error_response(e)
+
+
+@manager.route('/<kb_id>/tags', methods=['GET'])  # noqa: F821
+@login_required
+def list_tags(kb_id):
+    if not KnowledgebaseService.accessible(kb_id, current_user.id):
+        return get_json_result(
+            data=False,
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
+        )
+
+    tags = settings.retrievaler.all_tags(current_user.id, [kb_id])
+    return get_json_result(data=tags)
+
+
+@manager.route('/tags', methods=['GET'])  # noqa: F821
+@login_required
+def list_tags_from_kbs():
+    kb_ids = request.args.get("kb_ids", "").split(",")
+    for kb_id in kb_ids:
+        if not KnowledgebaseService.accessible(kb_id, current_user.id):
+            return get_json_result(
+                data=False,
+                message='No authorization.',
+                code=settings.RetCode.AUTHENTICATION_ERROR
+            )
+
+    tags = settings.retrievaler.all_tags(current_user.id, kb_ids)
+    return get_json_result(data=tags)
+
+
+@manager.route('/<kb_id>/rm_tags', methods=['POST'])  # noqa: F821
+@login_required
+def rm_tags(kb_id):
+    req = request.json
+    if not KnowledgebaseService.accessible(kb_id, current_user.id):
+        return get_json_result(
+            data=False,
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
+        )
+    e, kb = KnowledgebaseService.get_by_id(kb_id)
+
+    for t in req["tags"]:
+        settings.docStoreConn.update({"tag_kwd": t, "kb_id": [kb_id]},
+                                     {"remove": {"tag_kwd": t}},
+                                     search.index_name(kb.tenant_id),
+                                     kb_id)
+    return get_json_result(data=True)
+
+
+@manager.route('/<kb_id>/rename_tag', methods=['POST'])  # noqa: F821
+@login_required
+def rename_tags(kb_id):
+    req = request.json
+    if not KnowledgebaseService.accessible(kb_id, current_user.id):
+        return get_json_result(
+            data=False,
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
+        )
+    e, kb = KnowledgebaseService.get_by_id(kb_id)
+
+    settings.docStoreConn.update({"tag_kwd": req["from_tag"], "kb_id": [kb_id]},
+                                     {"remove": {"tag_kwd": req["from_tag"].strip()}, "add": {"tag_kwd": req["to_tag"]}},
+                                     search.index_name(kb.tenant_id),
+                                     kb_id)
+    return get_json_result(data=True)
+
+
+@manager.route('/<kb_id>/knowledge_graph', methods=['GET'])  # noqa: F821
+@login_required
+def knowledge_graph(kb_id):
+    if not KnowledgebaseService.accessible(kb_id, current_user.id):
+        return get_json_result(
+            data=False,
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR
+        )
+    _, kb = KnowledgebaseService.get_by_id(kb_id)
+    req = {
+        "kb_id": [kb_id],
+        "knowledge_graph_kwd": ["graph"]
+    }
+    obj = {"graph": {}, "mind_map": {}}
+    try:
+        sres = settings.retrievaler.search(req, search.index_name(kb.tenant_id), [kb_id])
+    except Exception as e:
+        logging.exception(e)
+        return get_json_result(data=obj)
+
+    for id in sres.ids[:1]:
+        ty = sres.field[id]["knowledge_graph_kwd"]
+        try:
+            content_json = json.loads(sres.field[id]["content_with_weight"])
+        except Exception:
+            continue
+
+        obj[ty] = content_json
+
+    if "nodes" in obj["graph"]:
+        obj["graph"]["nodes"] = sorted(obj["graph"]["nodes"], key=lambda x: x.get("pagerank", 0), reverse=True)[:256]
+    if "edges" in obj["graph"]:
+        obj["graph"]["edges"] = sorted(obj["graph"]["edges"], key=lambda x: x.get("weight", 0), reverse=True)[:128]
+    return get_json_result(data=obj)

api/apps/llm_app.py

@@ -13,21 +13,22 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 import json
-
+import os
 from flask import request
 from flask_login import login_required, current_user
 from api.db.services.llm_service import LLMFactoriesService, TenantLLMService, LLMService
-from api.settings import LIGHTEN
+from api import settings
 from api.utils.api_utils import server_error_response, get_data_error_result, validate_request
 from api.db import StatusEnum, LLMType
 from api.db.db_models import TenantLLM
 from api.utils.api_utils import get_json_result
+from api.utils.file_utils import get_project_base_directory
 from rag.llm import EmbeddingModel, ChatModel, RerankModel, CvModel, TTSModel
-import requests
 
 
-@manager.route('/factories', methods=['GET'])
+@manager.route('/factories', methods=['GET'])  # noqa: F821
 @login_required
 def factories():
     try:
@@ -49,7 +50,7 @@ def factories():
         return server_error_response(e)
 
 
-@manager.route('/set_api_key', methods=['POST'])
+@manager.route('/set_api_key', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("llm_factory", "api_key")
 def set_api_key():
@@ -89,7 +90,7 @@ def set_api_key():
                 if len(arr) == 0 or tc == 0:
                     raise Exception("Fail")
                 rerank_passed = True
-                print(f'passed model rerank{llm.llm_name}',flush=True)
+                logging.debug(f'passed model rerank {llm.llm_name}')
             except Exception as e:
                 msg += f"\nFail to access model({llm.llm_name}) using this api key." + str(
                     e)
@@ -98,7 +99,7 @@ def set_api_key():
             break
 
     if msg:
-        return get_data_error_result(retmsg=msg)
+        return get_data_error_result(message=msg)
 
     llm_config = {
         "api_key": req["api_key"],
@@ -109,6 +110,7 @@ def set_api_key():
             llm_config[n] = req[n]
 
     for llm in LLMService.query(fid=factory):
+        llm_config["max_tokens"]=llm.max_tokens
         if not TenantLLMService.filter_update(
                 [TenantLLM.tenant_id == current_user.id,
                  TenantLLM.llm_factory == factory,
@@ -120,13 +122,14 @@ def set_api_key():
                 llm_name=llm.llm_name,
                 model_type=llm.model_type,
                 api_key=llm_config["api_key"],
-                api_base=llm_config["api_base"]
+                api_base=llm_config["api_base"],
+                max_tokens=llm_config["max_tokens"]
             )
 
     return get_json_result(data=True)
 
 
-@manager.route('/add_llm', methods=['POST'])
+@manager.route('/add_llm', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("llm_factory")
 def add_llm():
@@ -201,7 +204,8 @@ def add_llm():
         "model_type": req["model_type"],
         "llm_name": llm_name,
         "api_base": req.get("api_base", ""),
-        "api_key": api_key
+        "api_key": api_key,
+        "max_tokens": req.get("max_tokens")
     }
 
     msg = ""
@@ -212,7 +216,7 @@ def add_llm():
             base_url=llm["api_base"])
         try:
             arr, tc = mdl.encode(["Test if the api key is available"])
-            if len(arr[0]) == 0 or tc == 0:
+            if len(arr[0]) == 0:
                 raise Exception("Fail")
         except Exception as e:
             msg += f"\nFail to access embedding model({llm['llm_name']})." + str(e)
@@ -237,8 +241,8 @@ def add_llm():
             base_url=llm["api_base"]
         )
         try:
-            arr, tc = mdl.similarity("Hello~ Ragflower!", ["Hi, there!"])
-            if len(arr) == 0 or tc == 0:
+            arr, tc = mdl.similarity("Hello~ Ragflower!", ["Hi, there!", "Ohh, my friend!"])
+            if len(arr) == 0:
                 raise Exception("Not known.")
         except Exception as e:
             msg += f"\nFail to access model({llm['llm_name']})." + str(
@@ -250,18 +254,10 @@ def add_llm():
             base_url=llm["api_base"]
         )
         try:
-            img_url = (
-                "https://upload.wikimedia.org/wikipedia/comm"
-                "ons/thumb/d/dd/Gfp-wisconsin-madison-the-nature-boardwalk.jpg/256"
-                "0px-Gfp-wisconsin-madison-the-nature-boardwalk.jpg"
-            )
-            res = requests.get(img_url)
-            if res.status_code == 200:
-                m, tc = mdl.describe(res.content)
+            with open(os.path.join(get_project_base_directory(), "web/src/assets/yay.jpg"), "rb") as f:
+                m, tc = mdl.describe(f.read())
                 if not tc:
                     raise Exception(m)
-            else:
-                pass
         except Exception as e:
             msg += f"\nFail to access model({llm['llm_name']})." + str(e)
     elif llm["model_type"] == LLMType.TTS:
@@ -278,26 +274,28 @@ def add_llm():
         pass
 
     if msg:
-        return get_data_error_result(retmsg=msg)
+        return get_data_error_result(message=msg)
 
     if not TenantLLMService.filter_update(
-            [TenantLLM.tenant_id == current_user.id, TenantLLM.llm_factory == factory, TenantLLM.llm_name == llm["llm_name"]], llm):
+            [TenantLLM.tenant_id == current_user.id, TenantLLM.llm_factory == factory,
+             TenantLLM.llm_name == llm["llm_name"]], llm):
         TenantLLMService.save(**llm)
 
     return get_json_result(data=True)
 
 
-@manager.route('/delete_llm', methods=['POST'])
+@manager.route('/delete_llm', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("llm_factory", "llm_name")
 def delete_llm():
     req = request.json
     TenantLLMService.filter_delete(
-            [TenantLLM.tenant_id == current_user.id, TenantLLM.llm_factory == req["llm_factory"], TenantLLM.llm_name == req["llm_name"]])
+        [TenantLLM.tenant_id == current_user.id, TenantLLM.llm_factory == req["llm_factory"],
+         TenantLLM.llm_name == req["llm_name"]])
     return get_json_result(data=True)
 
 
-@manager.route('/delete_factory', methods=['POST'])
+@manager.route('/delete_factory', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("llm_factory")
 def delete_factory():
@@ -307,7 +305,7 @@ def delete_factory():
     return get_json_result(data=True)
 
 
-@manager.route('/my_llms', methods=['GET'])
+@manager.route('/my_llms', methods=['GET'])  # noqa: F821
 @login_required
 def my_llms():
     try:
@@ -328,11 +326,11 @@ def my_llms():
         return server_error_response(e)
 
 
-@manager.route('/list', methods=['GET'])
+@manager.route('/list', methods=['GET'])  # noqa: F821
 @login_required
 def list_app():
-    self_deploied = ["Youdao","FastEmbed", "BAAI", "Ollama", "Xinference", "LocalAI", "LM-Studio"]
-    weighted = ["Youdao","FastEmbed", "BAAI"] if LIGHTEN != 0 else []
+    self_deployed = ["Youdao", "FastEmbed", "BAAI", "Ollama", "Xinference", "LocalAI", "LM-Studio", "GPUStack"]
+    weighted = ["Youdao", "FastEmbed", "BAAI"] if settings.LIGHTEN != 0 else []
     model_type = request.args.get("model_type")
     try:
         objs = TenantLLMService.query(tenant_id=current_user.id)
@@ -341,12 +339,14 @@ def list_app():
         llms = [m.to_dict()
                 for m in llms if m.status == StatusEnum.VALID.value and m.fid not in weighted]
         for m in llms:
-            m["available"] = m["fid"] in facts or m["llm_name"].lower() == "flag-embedding" or m["fid"] in self_deploied
+            m["available"] = m["fid"] in facts or m["llm_name"].lower() == "flag-embedding" or m["fid"] in self_deployed
 
         llm_set = set([m["llm_name"] + "@" + m["fid"] for m in llms])
         for o in objs:
-            if not o.api_key:continue
-            if o.llm_name+"@"+o.llm_factory in llm_set:continue
+            if not o.api_key:
+                continue
+            if o.llm_name + "@" + o.llm_factory in llm_set:
+                continue
             llms.append({"llm_name": o.llm_name, "model_type": o.model_type, "fid": o.llm_factory, "available": True})
 
         res = {}

api/apps/sdk/agent.py

@@ -0,0 +1,39 @@
+#
+#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+
+from api.db.services.canvas_service import UserCanvasService
+from api.utils.api_utils import get_error_data_result, token_required
+from api.utils.api_utils import get_result
+from flask import request
+
+@manager.route('/agents', methods=['GET'])  # noqa: F821
+@token_required
+def list_agents(tenant_id):
+    id = request.args.get("id")
+    title = request.args.get("title")
+    if id or title:
+        canvas = UserCanvasService.query(id=id, title=title, user_id=tenant_id)
+        if not canvas:
+            return get_error_data_result("The agent doesn't exist.")
+    page_number = int(request.args.get("page", 1))
+    items_per_page = int(request.args.get("page_size", 30))
+    orderby = request.args.get("orderby", "update_time")
+    if request.args.get("desc") == "False" or request.args.get("desc") == "false":
+        desc = False
+    else:
+        desc = True
+    canvas = UserCanvasService.get_list(tenant_id,page_number,items_per_page,orderby,desc,id,title)
+    return get_result(data=canvas)

api/apps/sdk/chat.py

@@ -14,7 +14,7 @@
 #  limitations under the License.
 #
 from flask import request
-from api.settings import RetCode
+from api import settings
 from api.db import StatusEnum
 from api.db.services.dialog_service import DialogService
 from api.db.services.knowledgebase_service import KnowledgebaseService
@@ -25,25 +25,26 @@ from api.utils.api_utils import get_error_data_result, token_required
 from api.utils.api_utils import get_result
 
 
-
-@manager.route('/chats', methods=['POST'])
+@manager.route('/chats', methods=['POST'])  # noqa: F821
 @token_required
 def create(tenant_id):
     req = request.json
     ids = req.get("dataset_ids")
     if not ids:
-        return get_error_data_result(retmsg="`dataset_ids` is required")
+        return get_error_data_result(message="`dataset_ids` is required")
     for kb_id in ids:
-        kbs = KnowledgebaseService.query(id=kb_id,tenant_id=tenant_id)
+        kbs = KnowledgebaseService.accessible(kb_id=kb_id, user_id=tenant_id)
         if not kbs:
             return get_error_data_result(f"You don't own the dataset {kb_id}")
+        kbs = KnowledgebaseService.query(id=kb_id)
         kb = kbs[0]
         if kb.chunk_num == 0:
             return get_error_data_result(f"The dataset {kb_id} doesn't own parsed file")
     kbs = KnowledgebaseService.get_by_ids(ids)
     embd_count = list(set([kb.embd_id for kb in kbs]))
     if len(embd_count) != 1:
-        return get_result(retmsg='Datasets use different embedding models."',retcode=RetCode.AUTHENTICATION_ERROR)
+        return get_result(message='Datasets use different embedding models."',
+                          code=settings.RetCode.AUTHENTICATION_ERROR)
     req["kb_ids"] = ids
     # llm
     llm = req.get("llm")
@@ -55,7 +56,7 @@ def create(tenant_id):
         req["llm_setting"] = req.pop("llm")
     e, tenant = TenantService.get_by_id(tenant_id)
     if not e:
-        return get_error_data_result(retmsg="Tenant not found!")
+        return get_error_data_result(message="Tenant not found!")
     # prompt
     prompt = req.get("prompt")
     key_mapping = {"parameters": "variables",
@@ -64,7 +65,7 @@ def create(tenant_id):
                    "system": "prompt",
                    "rerank_id": "rerank_model",
                    "vector_similarity_weight": "keywords_similarity_weight"}
-    key_list = ["similarity_threshold", "vector_similarity_weight", "top_n", "rerank_id"]
+    key_list = ["similarity_threshold", "vector_similarity_weight", "top_n", "rerank_id","top_k"]
     if prompt:
         for new_key, old_key in key_mapping.items():
             if old_key in prompt:
@@ -81,17 +82,20 @@ def create(tenant_id):
     req["top_k"] = req.get("top_k", 1024)
     req["rerank_id"] = req.get("rerank_id", "")
     if req.get("rerank_id"):
-        if not TenantLLMService.query(tenant_id=tenant_id,llm_name=req.get("rerank_id"),model_type="rerank"):
+        value_rerank_model = ["BAAI/bge-reranker-v2-m3", "maidalun1020/bce-reranker-base_v1"]
+        if req["rerank_id"] not in value_rerank_model and not TenantLLMService.query(tenant_id=tenant_id,
+                                                                                     llm_name=req.get("rerank_id"),
+                                                                                     model_type="rerank"):
             return get_error_data_result(f"`rerank_model` {req.get('rerank_id')} doesn't exist")
     if not req.get("llm_id"):
         req["llm_id"] = tenant.llm_id
     if not req.get("name"):
-        return get_error_data_result(retmsg="`name` is required.")
+        return get_error_data_result(message="`name` is required.")
     if DialogService.query(name=req["name"], tenant_id=tenant_id, status=StatusEnum.VALID.value):
-        return get_error_data_result(retmsg="Duplicated chat name in creating chat.")
+        return get_error_data_result(message="Duplicated chat name in creating chat.")
     # tenant_id
     if req.get("tenant_id"):
-        return get_error_data_result(retmsg="`tenant_id` must not be provided.")
+        return get_error_data_result(message="`tenant_id` must not be provided.")
     req["tenant_id"] = tenant_id
     # prompt more parameter
     default_prompt = {
@@ -103,28 +107,31 @@ def create(tenant_id):
         "parameters": [
             {"key": "knowledge", "optional": False}
         ],
-        "empty_response": "Sorry! No relevant content was found in the knowledge base!"
+        "empty_response": "Sorry! No relevant content was found in the knowledge base!",
+        "quote": True,
+        "tts": False,
+        "refine_multiturn": True
     }
-    key_list_2 = ["system", "prologue", "parameters", "empty_response"]
+    key_list_2 = ["system", "prologue", "parameters", "empty_response", "quote", "tts", "refine_multiturn"]
     if "prompt_config" not in req:
         req['prompt_config'] = {}
     for key in key_list_2:
         temp = req['prompt_config'].get(key)
-        if not temp:
+        if (not temp and key == 'system') or (key not in req["prompt_config"]):
             req['prompt_config'][key] = default_prompt[key]
     for p in req['prompt_config']["parameters"]:
         if p["optional"]:
             continue
         if req['prompt_config']["system"].find("{%s}" % p["key"]) < 0:
             return get_error_data_result(
-                retmsg="Parameter '{}' is not used".format(p["key"]))
+                message="Parameter '{}' is not used".format(p["key"]))
     # save
     if not DialogService.save(**req):
-        return get_error_data_result(retmsg="Fail to new a chat!")
+        return get_error_data_result(message="Fail to new a chat!")
     # response
     e, res = DialogService.get_by_id(req["id"])
     if not e:
-        return get_error_data_result(retmsg="Fail to new a chat!")
+        return get_error_data_result(message="Fail to new a chat!")
     res = res.to_json()
     renamed_dict = {}
     for key, value in res["prompt_config"].items():
@@ -133,7 +140,7 @@ def create(tenant_id):
     res["prompt"] = renamed_dict
     del res["prompt_config"]
     new_dict = {"similarity_threshold": res["similarity_threshold"],
-                "keywords_similarity_weight": res["vector_similarity_weight"],
+                "keywords_similarity_weight": 1-res["vector_similarity_weight"],
                 "top_n": res["top_n"],
                 "rerank_model": res['rerank_id']}
     res["prompt"].update(new_dict)
@@ -146,23 +153,25 @@ def create(tenant_id):
     res["avatar"] = res.pop("icon")
     return get_result(data=res)
 
-@manager.route('/chats/<chat_id>', methods=['PUT'])
+
+@manager.route('/chats/<chat_id>', methods=['PUT'])  # noqa: F821
 @token_required
 def update(tenant_id, chat_id):
     if not DialogService.query(tenant_id=tenant_id, id=chat_id, status=StatusEnum.VALID.value):
-        return get_error_data_result(retmsg='You do not own the chat')
+        return get_error_data_result(message='You do not own the chat')
     req = request.json
     ids = req.get("dataset_ids")
     if "show_quotation" in req:
         req["do_refer"] = req.pop("show_quotation")
     if "dataset_ids" in req:
         if not ids:
-            return get_error_data_result("`datasets` can't be empty")
+            return get_error_data_result("`dataset_ids` can't be empty")
         if ids:
             for kb_id in ids:
-                kbs = KnowledgebaseService.query(id=kb_id, tenant_id=tenant_id)
+                kbs = KnowledgebaseService.accessible(kb_id=kb_id, user_id=tenant_id)
                 if not kbs:
                     return get_error_data_result(f"You don't own the dataset {kb_id}")
+                kbs = KnowledgebaseService.query(id=kb_id)
                 kb = kbs[0]
                 if kb.chunk_num == 0:
                     return get_error_data_result(f"The dataset {kb_id} doesn't own parsed file")
@@ -170,8 +179,8 @@ def update(tenant_id,chat_id):
             embd_count = list(set([kb.embd_id for kb in kbs]))
             if len(embd_count) != 1:
                 return get_result(
-                    retmsg='Datasets use different embedding models."',
-                    retcode=RetCode.AUTHENTICATION_ERROR)
+                    message='Datasets use different embedding models."',
+                    code=settings.RetCode.AUTHENTICATION_ERROR)
             req["kb_ids"] = ids
     llm = req.get("llm")
     if llm:
@@ -182,10 +191,7 @@ def update(tenant_id,chat_id):
         req["llm_setting"] = req.pop("llm")
     e, tenant = TenantService.get_by_id(tenant_id)
     if not e:
-        return get_error_data_result(retmsg="Tenant not found!")
-    if req.get("rerank_model"):
-        if not TenantLLMService.query(tenant_id=tenant_id,llm_name=req.get("rerank_model"),model_type="rerank"):
-            return get_error_data_result(f"`rerank_model` {req.get('rerank_model')} doesn't exist")
+        return get_error_data_result(message="Tenant not found!")
     # prompt
     prompt = req.get("prompt")
     key_mapping = {"parameters": "variables",
@@ -194,7 +200,7 @@ def update(tenant_id,chat_id):
                    "system": "prompt",
                    "rerank_id": "rerank_model",
                    "vector_similarity_weight": "keywords_similarity_weight"}
-    key_list = ["similarity_threshold", "vector_similarity_weight", "top_n", "rerank_id"]
+    key_list = ["similarity_threshold", "vector_similarity_weight", "top_n", "rerank_id","top_k"]
     if prompt:
         for new_key, old_key in key_mapping.items():
             if old_key in prompt:
@@ -205,20 +211,26 @@ def update(tenant_id,chat_id):
         req["prompt_config"] = req.pop("prompt")
     e, res = DialogService.get_by_id(chat_id)
     res = res.to_json()
+    if req.get("rerank_id"):
+        value_rerank_model = ["BAAI/bge-reranker-v2-m3", "maidalun1020/bce-reranker-base_v1"]
+        if req["rerank_id"] not in value_rerank_model and not TenantLLMService.query(tenant_id=tenant_id,
+                                                                                     llm_name=req.get("rerank_id"),
+                                                                                     model_type="rerank"):
+            return get_error_data_result(f"`rerank_model` {req.get('rerank_id')} doesn't exist")
     if "name" in req:
         if not req.get("name"):
-            return get_error_data_result(retmsg="`name` is not empty.")
+            return get_error_data_result(message="`name` is not empty.")
         if req["name"].lower() != res["name"].lower() \
                 and len(
             DialogService.query(name=req["name"], tenant_id=tenant_id, status=StatusEnum.VALID.value)) > 0:
-            return get_error_data_result(retmsg="Duplicated chat name in updating dataset.")
+            return get_error_data_result(message="Duplicated chat name in updating dataset.")
     if "prompt_config" in req:
         res["prompt_config"].update(req["prompt_config"])
         for p in res["prompt_config"]["parameters"]:
             if p["optional"]:
                 continue
             if res["prompt_config"]["system"].find("{%s}" % p["key"]) < 0:
-                return get_error_data_result(retmsg="Parameter '{}' is not used".format(p["key"]))
+                return get_error_data_result(message="Parameter '{}' is not used".format(p["key"]))
     if "llm_setting" in req:
         res["llm_setting"].update(req["llm_setting"])
     req["prompt_config"] = res["prompt_config"]
@@ -229,11 +241,11 @@ def update(tenant_id,chat_id):
     if "dataset_ids" in req:
         req.pop("dataset_ids")
     if not DialogService.update_by_id(chat_id, req):
-        return get_error_data_result(retmsg="Chat not found!")
+        return get_error_data_result(message="Chat not found!")
     return get_result()
 
 
-@manager.route('/chats', methods=['DELETE'])
+@manager.route('/chats', methods=['DELETE'])  # noqa: F821
 @token_required
 def delete(tenant_id):
     req = request.json
@@ -250,21 +262,23 @@ def delete(tenant_id):
         id_list = ids
     for id in id_list:
         if not DialogService.query(tenant_id=tenant_id, id=id, status=StatusEnum.VALID.value):
-            return get_error_data_result(retmsg=f"You don't own the chat {id}")
+            return get_error_data_result(message=f"You don't own the chat {id}")
         temp_dict = {"status": StatusEnum.INVALID.value}
         DialogService.update_by_id(id, temp_dict)
     return get_result()
 
-@manager.route('/chats', methods=['GET'])
+
+@manager.route('/chats', methods=['GET'])  # noqa: F821
 @token_required
 def list_chat(tenant_id):
     id = request.args.get("id")
     name = request.args.get("name")
-    chat = DialogService.query(id=id,name=name,status=StatusEnum.VALID.value)
+    if id or name:
+        chat = DialogService.query(id=id, name=name, status=StatusEnum.VALID.value, tenant_id=tenant_id)
         if not chat:
-        return get_error_data_result(retmsg="The chat doesn't exist")
+            return get_error_data_result(message="The chat doesn't exist")
     page_number = int(request.args.get("page", 1))
-    items_per_page = int(request.args.get("page_size", 1024))
+    items_per_page = int(request.args.get("page_size", 30))
     orderby = request.args.get("orderby", "create_time")
     if request.args.get("desc") == "False" or request.args.get("desc") == "false":
         desc = False
@@ -274,7 +288,6 @@ def list_chat(tenant_id):
     if not chats:
         return get_result(data=[])
     list_assts = []
-    renamed_dict = {}
     key_mapping = {"parameters": "variables",
                    "prologue": "opener",
                    "quote": "show_quote",
@@ -284,13 +297,14 @@ def list_chat(tenant_id):
                    "do_refer": "show_quotation"}
     key_list = ["similarity_threshold", "vector_similarity_weight", "top_n", "rerank_id"]
     for res in chats:
+        renamed_dict = {}
         for key, value in res["prompt_config"].items():
             new_key = key_mapping.get(key, key)
             renamed_dict[new_key] = value
         res["prompt"] = renamed_dict
         del res["prompt_config"]
         new_dict = {"similarity_threshold": res["similarity_threshold"],
-                    "keywords_similarity_weight": res["vector_similarity_weight"],
+                    "keywords_similarity_weight": 1-res["vector_similarity_weight"],
                     "top_n": res["top_n"],
                     "rerank_model": res['rerank_id']}
         res["prompt"].update(new_dict)
@@ -302,7 +316,7 @@ def list_chat(tenant_id):
         for kb_id in res["kb_ids"]:
             kb = KnowledgebaseService.query(id=kb_id)
             if not kb:
-                return get_error_data_result(retmsg=f"Don't exist the kb {kb_id}")
+                return get_error_data_result(message=f"Don't exist the kb {kb_id}")
             kb_list.append(kb[0].to_json())
         del res["kb_ids"]
         res["datasets"] = kb_list

api/apps/sdk/dataset.py

@@ -23,14 +23,71 @@ from api.db.services.file_service import FileService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.llm_service import TenantLLMService, LLMService
 from api.db.services.user_service import TenantService
-from api.settings import RetCode
+from api import settings
 from api.utils import get_uuid
-from api.utils.api_utils import get_result, token_required, get_error_data_result, valid,get_parser_config
+from api.utils.api_utils import (
+    get_result,
+    token_required,
+    get_error_data_result,
+    valid,
+    get_parser_config,
+)
 
 
-@manager.route('/datasets', methods=['POST'])
+@manager.route("/datasets", methods=["POST"])  # noqa: F821
 @token_required
 def create(tenant_id):
+    """
+    Create a new dataset.
+    ---
+    tags:
+      - Datasets
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: header
+        name: Authorization
+        type: string
+        required: true
+        description: Bearer token for authentication.
+      - in: body
+        name: body
+        description: Dataset creation parameters.
+        required: true
+        schema:
+          type: object
+          required:
+            - name
+          properties:
+            name:
+              type: string
+              description: Name of the dataset.
+            permission:
+              type: string
+              enum: ['me', 'team']
+              description: Dataset permission.
+            language:
+              type: string
+              enum: ['Chinese', 'English']
+              description: Language of the dataset.
+            chunk_method:
+              type: string
+              enum: ["naive", "manual", "qa", "table", "paper", "book", "laws",
+                     "presentation", "picture", "one", "knowledge_graph", "email", "tag"
+                     ]
+              description: Chunking method.
+            parser_config:
+              type: object
+              description: Parser configuration.
+    responses:
+      200:
+        description: Successful operation.
+        schema:
+          type: object
+          properties:
+            data:
+              type: object
+    """
     req = request.json
     e, t = TenantService.get_by_id(tenant_id)
     permission = req.get("permission")
@@ -39,51 +96,94 @@ def create(tenant_id):
     parser_config = req.get("parser_config")
     valid_permission = ["me", "team"]
     valid_language = ["Chinese", "English"]
-    valid_chunk_method = ["naive","manual","qa","table","paper","book","laws","presentation","picture","one","knowledge_graph","email"]
-    check_validation=valid(permission,valid_permission,language,valid_language,chunk_method,valid_chunk_method)
+    valid_chunk_method = [
+        "naive",
+        "manual",
+        "qa",
+        "table",
+        "paper",
+        "book",
+        "laws",
+        "presentation",
+        "picture",
+        "one",
+        "knowledge_graph",
+        "email",
+        "tag"
+    ]
+    check_validation = valid(
+        permission,
+        valid_permission,
+        language,
+        valid_language,
+        chunk_method,
+        valid_chunk_method,
+    )
     if check_validation:
         return check_validation
     req["parser_config"] = get_parser_config(chunk_method, parser_config)
     if "tenant_id" in req:
-        return get_error_data_result(
-            retmsg="`tenant_id` must not be provided")
+        return get_error_data_result(message="`tenant_id` must not be provided")
     if "chunk_count" in req or "document_count" in req:
-        return get_error_data_result(retmsg="`chunk_count` or `document_count` must not be provided")
-    if "name" not in req:
         return get_error_data_result(
-            retmsg="`name` is not empty!")
-    req['id'] = get_uuid()
+            message="`chunk_count` or `document_count` must not be provided"
+        )
+    if "name" not in req:
+        return get_error_data_result(message="`name` is not empty!")
+    req["id"] = get_uuid()
     req["name"] = req["name"].strip()
     if req["name"] == "":
+        return get_error_data_result(message="`name` is not empty string!")
+    if KnowledgebaseService.query(
+        name=req["name"], tenant_id=tenant_id, status=StatusEnum.VALID.value
+    ):
         return get_error_data_result(
-            retmsg="`name` is not empty string!")
-    if KnowledgebaseService.query(name=req["name"], tenant_id=tenant_id, status=StatusEnum.VALID.value):
-        return get_error_data_result(
-            retmsg="Duplicated dataset name in creating dataset.")
-    req["tenant_id"] = req['created_by'] = tenant_id
+            message="Duplicated dataset name in creating dataset."
+        )
+    req["tenant_id"] = req["created_by"] = tenant_id
     if not req.get("embedding_model"):
-        req['embedding_model'] = t.embd_id
+        req["embedding_model"] = t.embd_id
     else:
-        valid_embedding_models=["BAAI/bge-large-zh-v1.5","BAAI/bge-base-en-v1.5","BAAI/bge-large-en-v1.5","BAAI/bge-small-en-v1.5",
-                                "BAAI/bge-small-zh-v1.5","jinaai/jina-embeddings-v2-base-en","jinaai/jina-embeddings-v2-small-en",
-                                "nomic-ai/nomic-embed-text-v1.5","sentence-transformers/all-MiniLM-L6-v2","text-embedding-v2",
-                                "text-embedding-v3","maidalun1020/bce-embedding-base_v1"]
-        embd_model=LLMService.query(llm_name=req["embedding_model"],model_type="embedding")
-        if not embd_model:
-            return get_error_data_result(f"`embedding_model` {req.get('embedding_model')} doesn't exist")
+        valid_embedding_models = [
+            "BAAI/bge-large-zh-v1.5",
+            "BAAI/bge-base-en-v1.5",
+            "BAAI/bge-large-en-v1.5",
+            "BAAI/bge-small-en-v1.5",
+            "BAAI/bge-small-zh-v1.5",
+            "jinaai/jina-embeddings-v2-base-en",
+            "jinaai/jina-embeddings-v2-small-en",
+            "nomic-ai/nomic-embed-text-v1.5",
+            "sentence-transformers/all-MiniLM-L6-v2",
+            "text-embedding-v2",
+            "text-embedding-v3",
+            "maidalun1020/bce-embedding-base_v1",
+        ]
+        embd_model = LLMService.query(
+            llm_name=req["embedding_model"], model_type="embedding"
+        )
         if embd_model:
-            if req["embedding_model"] not in valid_embedding_models and not TenantLLMService.query(tenant_id=tenant_id,model_type="embedding", llm_name=req.get("embedding_model")):
+            if req["embedding_model"] not in valid_embedding_models and not TenantLLMService.query(tenant_id=tenant_id,model_type="embedding",llm_name=req.get("embedding_model"),):
                 return get_error_data_result(f"`embedding_model` {req.get('embedding_model')} doesn't exist")
+        if not embd_model:
+            embd_model=TenantLLMService.query(tenant_id=tenant_id,model_type="embedding", llm_name=req.get("embedding_model"))
+        if not embd_model:
+            return get_error_data_result(
+                f"`embedding_model` {req.get('embedding_model')} doesn't exist"
+            )
     key_mapping = {
         "chunk_num": "chunk_count",
         "doc_num": "document_count",
         "parser_id": "chunk_method",
-        "embd_id": "embedding_model"
+        "embd_id": "embedding_model",
+    }
+    mapped_keys = {
+        new_key: req[old_key]
+        for new_key, old_key in key_mapping.items()
+        if old_key in req
     }
-    mapped_keys = {new_key: req[old_key] for new_key, old_key in key_mapping.items() if old_key in req}
     req.update(mapped_keys)
     if not KnowledgebaseService.save(**req):
-        return get_error_data_result(retmsg="Create dataset error.(Database error)")
+        return get_error_data_result(message="Create dataset error.(Database error)")
     renamed_data = {}
     e, k = KnowledgebaseService.get_by_id(req["id"])
     for key, value in k.to_dict().items():
@@ -91,9 +191,41 @@ def create(tenant_id):
         renamed_data[new_key] = value
     return get_result(data=renamed_data)
 
-@manager.route('/datasets', methods=['DELETE'])
+
+@manager.route("/datasets", methods=["DELETE"])  # noqa: F821
 @token_required
 def delete(tenant_id):
+    """
+    Delete datasets.
+    ---
+    tags:
+      - Datasets
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: header
+        name: Authorization
+        type: string
+        required: true
+        description: Bearer token for authentication.
+      - in: body
+        name: body
+        description: Dataset deletion parameters.
+        required: true
+        schema:
+          type: object
+          properties:
+            ids:
+              type: array
+              items:
+                type: string
+              description: List of dataset IDs to delete.
+    responses:
+      200:
+        description: Successful operation.
+        schema:
+          type: object
+    """
     req = request.json
     if not req:
         ids = None
@@ -109,44 +241,122 @@ def delete(tenant_id):
     for id in id_list:
         kbs = KnowledgebaseService.query(id=id, tenant_id=tenant_id)
         if not kbs:
-            return get_error_data_result(retmsg=f"You don't own the dataset {id}")
+            return get_error_data_result(message=f"You don't own the dataset {id}")
         for doc in DocumentService.query(kb_id=id):
             if not DocumentService.remove_document(doc, tenant_id):
                 return get_error_data_result(
-                    retmsg="Remove document error.(Database error)")
+                    message="Remove document error.(Database error)"
+                )
             f2d = File2DocumentService.get_by_document_id(doc.id)
-            FileService.filter_delete([File.source_type == FileSource.KNOWLEDGEBASE, File.id == f2d[0].file_id])
+            FileService.filter_delete(
+                [
+                    File.source_type == FileSource.KNOWLEDGEBASE,
+                    File.id == f2d[0].file_id,
+                ]
+            )
             File2DocumentService.delete_by_document_id(doc.id)
+        FileService.filter_delete(
+            [File.source_type == FileSource.KNOWLEDGEBASE, File.type == "folder", File.name == kbs[0].name])
         if not KnowledgebaseService.delete_by_id(id):
-            return get_error_data_result(
-                retmsg="Delete dataset error.(Database error)")
-    return get_result(retcode=RetCode.SUCCESS)
+            return get_error_data_result(message="Delete dataset error.(Database error)")
+    return get_result(code=settings.RetCode.SUCCESS)
 
-@manager.route('/datasets/<dataset_id>', methods=['PUT'])
+
+@manager.route("/datasets/<dataset_id>", methods=["PUT"])  # noqa: F821
 @token_required
 def update(tenant_id, dataset_id):
+    """
+    Update a dataset.
+    ---
+    tags:
+      - Datasets
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: path
+        name: dataset_id
+        type: string
+        required: true
+        description: ID of the dataset to update.
+      - in: header
+        name: Authorization
+        type: string
+        required: true
+        description: Bearer token for authentication.
+      - in: body
+        name: body
+        description: Dataset update parameters.
+        required: true
+        schema:
+          type: object
+          properties:
+            name:
+              type: string
+              description: New name of the dataset.
+            permission:
+              type: string
+              enum: ['me', 'team']
+              description: Updated permission.
+            language:
+              type: string
+              enum: ['Chinese', 'English']
+              description: Updated language.
+            chunk_method:
+              type: string
+              enum: ["naive", "manual", "qa", "table", "paper", "book", "laws",
+                     "presentation", "picture", "one", "knowledge_graph", "email", "tag"
+                     ]
+              description: Updated chunking method.
+            parser_config:
+              type: object
+              description: Updated parser configuration.
+    responses:
+      200:
+        description: Successful operation.
+        schema:
+          type: object
+    """
     if not KnowledgebaseService.query(id=dataset_id, tenant_id=tenant_id):
-        return get_error_data_result(retmsg="You don't own the dataset")
+        return get_error_data_result(message="You don't own the dataset")
     req = request.json
     e, t = TenantService.get_by_id(tenant_id)
     invalid_keys = {"id", "embd_id", "chunk_num", "doc_num", "parser_id"}
     if any(key in req for key in invalid_keys):
-        return get_error_data_result(retmsg="The input parameters are invalid.")
+        return get_error_data_result(message="The input parameters are invalid.")
     permission = req.get("permission")
     language = req.get("language")
     chunk_method = req.get("chunk_method")
     parser_config = req.get("parser_config")
     valid_permission = ["me", "team"]
     valid_language = ["Chinese", "English"]
-    valid_chunk_method = ["naive", "manual", "qa", "table", "paper", "book", "laws", "presentation", "picture", "one",
-                          "knowledge_graph", "email"]
-    check_validation = valid(permission, valid_permission, language, valid_language, chunk_method, valid_chunk_method)
+    valid_chunk_method = [
+        "naive",
+        "manual",
+        "qa",
+        "table",
+        "paper",
+        "book",
+        "laws",
+        "presentation",
+        "picture",
+        "one",
+        "knowledge_graph",
+        "email",
+        "tag"
+    ]
+    check_validation = valid(
+        permission,
+        valid_permission,
+        language,
+        valid_language,
+        chunk_method,
+        valid_chunk_method,
+    )
     if check_validation:
         return check_validation
     if "tenant_id" in req:
         if req["tenant_id"] != tenant_id:
-            return get_error_data_result(
-                retmsg="Can't change `tenant_id`.")
+            return get_error_data_result(message="Can't change `tenant_id`.")
     e, kb = KnowledgebaseService.get_by_id(dataset_id)
     if "parser_config" in req:
         temp_dict = kb.parser_config
@@ -154,60 +364,145 @@ def update(tenant_id,dataset_id):
         req["parser_config"] = temp_dict
     if "chunk_count" in req:
         if req["chunk_count"] != kb.chunk_num:
-            return get_error_data_result(
-                retmsg="Can't change `chunk_count`.")
+            return get_error_data_result(message="Can't change `chunk_count`.")
         req.pop("chunk_count")
     if "document_count" in req:
-        if req['document_count'] != kb.doc_num:
-            return get_error_data_result(
-                retmsg="Can't change `document_count`.")
+        if req["document_count"] != kb.doc_num:
+            return get_error_data_result(message="Can't change `document_count`.")
         req.pop("document_count")
     if "chunk_method" in req:
-        if kb.chunk_num != 0 and req['chunk_method'] != kb.parser_id:
+        if kb.chunk_num != 0 and req["chunk_method"] != kb.parser_id:
             return get_error_data_result(
-                retmsg="If `chunk_count` is not 0, `chunk_method` is not changeable.")
-        req['parser_id'] = req.pop('chunk_method')
-        if req['parser_id'] != kb.parser_id:
+                message="If `chunk_count` is not 0, `chunk_method` is not changeable."
+            )
+        req["parser_id"] = req.pop("chunk_method")
+        if req["parser_id"] != kb.parser_id:
             if not req.get("parser_config"):
                 req["parser_config"] = get_parser_config(chunk_method, parser_config)
     if "embedding_model" in req:
-        if kb.chunk_num != 0 and req['embedding_model'] != kb.embd_id:
+        if kb.chunk_num != 0 and req["embedding_model"] != kb.embd_id:
             return get_error_data_result(
-                retmsg="If `chunk_count` is not 0, `embedding_model` is not changeable.")
+                message="If `chunk_count` is not 0, `embedding_model` is not changeable."
+            )
         if not req.get("embedding_model"):
             return get_error_data_result("`embedding_model` can't be empty")
-        valid_embedding_models=["BAAI/bge-large-zh-v1.5","BAAI/bge-base-en-v1.5","BAAI/bge-large-en-v1.5","BAAI/bge-small-en-v1.5",
-                                "BAAI/bge-small-zh-v1.5","jinaai/jina-embeddings-v2-base-en","jinaai/jina-embeddings-v2-small-en",
-                                "nomic-ai/nomic-embed-text-v1.5","sentence-transformers/all-MiniLM-L6-v2","text-embedding-v2",
-                                "text-embedding-v3","maidalun1020/bce-embedding-base_v1"]
-        embd_model=LLMService.query(llm_name=req["embedding_model"],model_type="embedding")
-        if not embd_model:
-            return get_error_data_result(f"`embedding_model` {req.get('embedding_model')} doesn't exist")
+        valid_embedding_models = [
+            "BAAI/bge-large-zh-v1.5",
+            "BAAI/bge-base-en-v1.5",
+            "BAAI/bge-large-en-v1.5",
+            "BAAI/bge-small-en-v1.5",
+            "BAAI/bge-small-zh-v1.5",
+            "jinaai/jina-embeddings-v2-base-en",
+            "jinaai/jina-embeddings-v2-small-en",
+            "nomic-ai/nomic-embed-text-v1.5",
+            "sentence-transformers/all-MiniLM-L6-v2",
+            "text-embedding-v2",
+            "text-embedding-v3",
+            "maidalun1020/bce-embedding-base_v1",
+        ]
+        embd_model = LLMService.query(
+            llm_name=req["embedding_model"], model_type="embedding"
+        )
         if embd_model:
-            if req["embedding_model"] not in valid_embedding_models and not TenantLLMService.query(tenant_id=tenant_id,model_type="embedding", llm_name=req.get("embedding_model")):
+            if req["embedding_model"] not in valid_embedding_models and not TenantLLMService.query(tenant_id=tenant_id,model_type="embedding",llm_name=req.get("embedding_model"),):
                 return get_error_data_result(f"`embedding_model` {req.get('embedding_model')} doesn't exist")
-        req['embd_id'] = req.pop('embedding_model')
+        if not embd_model:
+            embd_model=TenantLLMService.query(tenant_id=tenant_id,model_type="embedding", llm_name=req.get("embedding_model"))
+
+        if not embd_model:
+            return get_error_data_result(
+                f"`embedding_model` {req.get('embedding_model')} doesn't exist"
+            )
+        req["embd_id"] = req.pop("embedding_model")
     if "name" in req:
         req["name"] = req["name"].strip()
-        if req["name"].lower() != kb.name.lower() \
-                and len(KnowledgebaseService.query(name=req["name"], tenant_id=tenant_id,
-                                                   status=StatusEnum.VALID.value)) > 0:
+        if (
+            req["name"].lower() != kb.name.lower()
+            and len(
+                KnowledgebaseService.query(
+                    name=req["name"], tenant_id=tenant_id, status=StatusEnum.VALID.value
+                )
+            )
+            > 0
+        ):
             return get_error_data_result(
-                retmsg="Duplicated dataset name in updating dataset.")
+                message="Duplicated dataset name in updating dataset."
+            )
     if not KnowledgebaseService.update_by_id(kb.id, req):
-        return get_error_data_result(retmsg="Update dataset error.(Database error)")
-    return get_result(retcode=RetCode.SUCCESS)
+        return get_error_data_result(message="Update dataset error.(Database error)")
+    return get_result(code=settings.RetCode.SUCCESS)
 
-@manager.route('/datasets', methods=['GET'])
+
+@manager.route("/datasets", methods=["GET"])  # noqa: F821
 @token_required
 def list(tenant_id):
+    """
+    List datasets.
+    ---
+    tags:
+      - Datasets
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: query
+        name: id
+        type: string
+        required: false
+        description: Dataset ID to filter.
+      - in: query
+        name: name
+        type: string
+        required: false
+        description: Dataset name to filter.
+      - in: query
+        name: page
+        type: integer
+        required: false
+        default: 1
+        description: Page number.
+      - in: query
+        name: page_size
+        type: integer
+        required: false
+        default: 1024
+        description: Number of items per page.
+      - in: query
+        name: orderby
+        type: string
+        required: false
+        default: "create_time"
+        description: Field to order by.
+      - in: query
+        name: desc
+        type: boolean
+        required: false
+        default: true
+        description: Order in descending.
+      - in: header
+        name: Authorization
+        type: string
+        required: true
+        description: Bearer token for authentication.
+    responses:
+      200:
+        description: Successful operation.
+        schema:
+          type: array
+          items:
+            type: object
+    """
     id = request.args.get("id")
     name = request.args.get("name")
-    kbs = KnowledgebaseService.query(id=id,name=name,status=1)
+    if id:
+        kbs = KnowledgebaseService.get_kb_by_id(id,tenant_id)
         if not kbs:
-        return get_error_data_result(retmsg="The dataset doesn't exist")
+            return get_error_data_result(f"You don't own the dataset {id}")
+    if name:
+        kbs = KnowledgebaseService.get_kb_by_name(name,tenant_id)
+        if not kbs:
+            return get_error_data_result(f"You don't own the dataset {name}")
     page_number = int(request.args.get("page", 1))
-    items_per_page = int(request.args.get("page_size", 1024))
+    items_per_page = int(request.args.get("page_size", 30))
     orderby = request.args.get("orderby", "create_time")
     if request.args.get("desc") == "False" or request.args.get("desc") == "false":
         desc = False
@@ -215,14 +510,22 @@ def list(tenant_id):
         desc = True
     tenants = TenantService.get_joined_tenants_by_user_id(tenant_id)
     kbs = KnowledgebaseService.get_list(
-        [m["tenant_id"] for m in tenants], tenant_id, page_number, items_per_page, orderby, desc, id, name)
+        [m["tenant_id"] for m in tenants],
+        tenant_id,
+        page_number,
+        items_per_page,
+        orderby,
+        desc,
+        id,
+        name,
+    )
     renamed_list = []
     for kb in kbs:
         key_mapping = {
             "chunk_num": "chunk_count",
             "doc_num": "document_count",
             "parser_id": "chunk_method",
-            "embd_id": "embedding_model"
+            "embd_id": "embedding_model",
         }
         renamed_data = {}
         for key, value in kb.items():

api/apps/sdk/dify_retrieval.py

@@ -15,20 +15,22 @@
 #
 from flask import request, jsonify
 
-from api.db import LLMType, ParserType
+from api.db import LLMType
+from api.db.services.dialog_service import label_question
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.llm_service import LLMBundle
-from api.settings import retrievaler, kg_retrievaler, RetCode
+from api import settings
 from api.utils.api_utils import validate_request, build_error_result, apikey_required
 
 
-@manager.route('/dify/retrieval', methods=['POST'])
+@manager.route('/dify/retrieval', methods=['POST'])  # noqa: F821
 @apikey_required
 @validate_request("knowledge_id", "query")
 def retrieval(tenant_id):
     req = request.json
     question = req["query"]
     kb_id = req["knowledge_id"]
+    use_kg = req.get("use_kg", False)
     retrieval_setting = req.get("retrieval_setting", {})
     similarity_threshold = float(retrieval_setting.get("score_threshold", 0.0))
     top = int(retrieval_setting.get("top_k", 1024))
@@ -37,15 +39,14 @@ def retrieval(tenant_id):
 
         e, kb = KnowledgebaseService.get_by_id(kb_id)
         if not e:
-            return build_error_result(error_msg="Knowledgebase not found!", retcode=RetCode.NOT_FOUND)
+            return build_error_result(message="Knowledgebase not found!", code=settings.RetCode.NOT_FOUND)
 
         if kb.tenant_id != tenant_id:
-            return build_error_result(error_msg="Knowledgebase not found!", retcode=RetCode.NOT_FOUND)
+            return build_error_result(message="Knowledgebase not found!", code=settings.RetCode.NOT_FOUND)
 
         embd_mdl = LLMBundle(kb.tenant_id, LLMType.EMBEDDING.value, llm_name=kb.embd_id)
 
-        retr = retrievaler if kb.parser_id != ParserType.KG else kg_retrievaler
-        ranks = retr.retrieval(
+        ranks = settings.retrievaler.retrieval(
             question,
             embd_mdl,
             kb.tenant_id,
@@ -54,14 +55,24 @@ def retrieval(tenant_id):
             page_size=top,
             similarity_threshold=similarity_threshold,
             vector_similarity_weight=0.3,
-            top=top
+            top=top,
+            rank_feature=label_question(question, [kb])
         )
+
+        if use_kg:
+            ck = settings.kg_retrievaler.retrieval(question,
+                                                   [tenant_id],
+                                                   [kb_id],
+                                                   embd_mdl,
+                                                   LLMBundle(kb.tenant_id, LLMType.CHAT))
+            if ck["content_with_weight"]:
+                ranks["chunks"].insert(0, ck)
+
         records = []
         for c in ranks["chunks"]:
-            if "vector" in c:
-                del c["vector"]
+            c.pop("vector", None)
             records.append({
-                "content": c["content_ltks"],
+                "content": c["content_with_weight"],
                 "score": c["similarity"],
                 "title": c["docnm_kwd"],
                 "metadata": {}
@@ -71,7 +82,7 @@ def retrieval(tenant_id):
     except Exception as e:
         if str(e).find("not_found") > 0:
             return build_error_result(
-                error_msg=f'No chunk found! Check the chunk status please!',
-                retcode=RetCode.NOT_FOUND
+                message='No chunk found! Check the chunk status please!',
+                code=settings.RetCode.NOT_FOUND
             )
-        return build_error_result(error_msg=str(e), retcode=RetCode.SERVER_ERROR)
+        return build_error_result(message=str(e), code=settings.RetCode.SERVER_ERROR)

api/apps/sdk/session.py

@@ -13,44 +13,104 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import re
 import json
-from uuid import uuid4
-
+from api.db import LLMType
 from flask import request, Response
 
+from api.db.services.conversation_service import ConversationService, iframe_completion
+from api.db.services.conversation_service import completion as rag_completion
+from api.db.services.canvas_service import completion as agent_completion
+from api.db.services.dialog_service import ask
+from agent.canvas import Canvas
 from api.db import StatusEnum
-from api.db.services.dialog_service import DialogService, ConversationService, chat
+from api.db.db_models import APIToken
+from api.db.services.api_service import API4ConversationService
+from api.db.services.canvas_service import UserCanvasService
+from api.db.services.dialog_service import DialogService
+from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.utils import get_uuid
 from api.utils.api_utils import get_error_data_result
 from api.utils.api_utils import get_result, token_required
+from api.db.services.llm_service import LLMBundle
 
-@manager.route('/chats/<chat_id>/sessions', methods=['POST'])
+
+
+@manager.route('/chats/<chat_id>/sessions', methods=['POST'])  # noqa: F821
 @token_required
 def create(tenant_id, chat_id):
     req = request.json
     req["dialog_id"] = chat_id
     dia = DialogService.query(tenant_id=tenant_id, id=req["dialog_id"], status=StatusEnum.VALID.value)
     if not dia:
-        return get_error_data_result(retmsg="You do not own the assistant")
+        return get_error_data_result(message="You do not own the assistant.")
     conv = {
         "id": get_uuid(),
         "dialog_id": req["dialog_id"],
         "name": req.get("name", "New session"),
-        "message": [{"role": "assistant", "content": "Hi! I am your assistant，can I help you?"}]
+        "message": [{"role": "assistant", "content": dia[0].prompt_config.get("prologue")}],
+        "user_id": req.get("user_id", "")
     }
     if not conv.get("name"):
-        return get_error_data_result(retmsg="`name` can not be empty.")
+        return get_error_data_result(message="`name` can not be empty.")
     ConversationService.save(**conv)
     e, conv = ConversationService.get_by_id(conv["id"])
     if not e:
-        return get_error_data_result(retmsg="Fail to create a session!")
+        return get_error_data_result(message="Fail to create a session!")
     conv = conv.to_dict()
     conv['messages'] = conv.pop("message")
     conv["chat_id"] = conv.pop("dialog_id")
     del conv["reference"]
     return get_result(data=conv)
 
-@manager.route('/chats/<chat_id>/sessions/<session_id>', methods=['PUT'])
+
+@manager.route('/agents/<agent_id>/sessions', methods=['POST'])  # noqa: F821
+@token_required
+def create_agent_session(tenant_id, agent_id):
+    req = request.json
+    e, cvs = UserCanvasService.get_by_id(agent_id)
+    if not e:
+        return get_error_data_result("Agent not found.")
+
+    if not UserCanvasService.query(user_id=tenant_id, id=agent_id):
+        return get_error_data_result("You cannot access the agent.")
+
+    if not isinstance(cvs.dsl, str):
+        cvs.dsl = json.dumps(cvs.dsl, ensure_ascii=False)
+
+    canvas = Canvas(cvs.dsl, tenant_id)
+    canvas.reset()
+    query = canvas.get_preset_param()
+    if query:
+        for ele in query:
+            if not ele["optional"]:
+                if not req.get(ele["key"]):
+                    return get_error_data_result(f"`{ele['key']}` is required")
+                ele["value"] = req[ele["key"]]
+            if ele["optional"]:
+                if req.get(ele["key"]):
+                    ele["value"] = req[ele['key']]
+                else:
+                    if "value" in ele:
+                        ele.pop("value")
+    else:
+        for ans in canvas.run(stream=False):
+            pass
+    cvs.dsl = json.loads(str(canvas))
+    conv = {
+        "id": get_uuid(),
+        "dialog_id": cvs.id,
+        "user_id": req.get("user_id", "") if isinstance(req, dict) else "",
+        "message": [{"role": "assistant", "content": canvas.get_prologue()}],
+        "source": "agent",
+        "dsl": cvs.dsl
+    }
+    API4ConversationService.save(**conv)
+    conv["agent_id"] = conv.pop("dialog_id")
+    return get_result(data=conv)
+
+
+@manager.route('/chats/<chat_id>/sessions/<session_id>', methods=['PUT'])  # noqa: F821
 @token_required
 def update(tenant_id, chat_id, session_id):
     req = request.json
@@ -58,121 +118,96 @@ def update(tenant_id,chat_id,session_id):
     conv_id = session_id
     conv = ConversationService.query(id=conv_id, dialog_id=chat_id)
     if not conv:
-        return get_error_data_result(retmsg="Session does not exist")
+        return get_error_data_result(message="Session does not exist")
     if not DialogService.query(id=chat_id, tenant_id=tenant_id, status=StatusEnum.VALID.value):
-        return get_error_data_result(retmsg="You do not own the session")
+        return get_error_data_result(message="You do not own the session")
     if "message" in req or "messages" in req:
-        return get_error_data_result(retmsg="`message` can not be change")
+        return get_error_data_result(message="`message` can not be change")
     if "reference" in req:
-        return get_error_data_result(retmsg="`reference` can not be change")
+        return get_error_data_result(message="`reference` can not be change")
     if "name" in req and not req.get("name"):
-        return get_error_data_result(retmsg="`name` can not be empty.")
+        return get_error_data_result(message="`name` can not be empty.")
     if not ConversationService.update_by_id(conv_id, req):
-        return get_error_data_result(retmsg="Session updates error")
+        return get_error_data_result(message="Session updates error")
     return get_result()
 
 
-@manager.route('/chats/<chat_id>/completions', methods=['POST'])
+@manager.route('/chats/<chat_id>/completions', methods=['POST'])  # noqa: F821
 @token_required
-def completion(tenant_id,chat_id):
+def chat_completion(tenant_id, chat_id):
     req = request.json
-    if not req.get("session_id"):
-        conv = {
-            "id": get_uuid(),
-            "dialog_id": chat_id,
-            "name": req.get("name", "New session"),
-            "message": [{"role": "assistant", "content": "Hi! I am your assistant，can I help you?"}]
-        }
-        if not conv.get("name"):
-            return get_error_data_result(retmsg="`name` can not be empty.")
-        ConversationService.save(**conv)
-        e, conv = ConversationService.get_by_id(conv["id"])
-        session_id=conv.id
-    else:
-        session_id = req.get("session_id")
-    if not req.get("question"):
-        return get_error_data_result(retmsg="Please input your question.")
-    conv = ConversationService.query(id=session_id,dialog_id=chat_id)
-    if not conv:
-        return get_error_data_result(retmsg="Session does not exist")
-    conv = conv[0]
-    if not DialogService.query(id=chat_id, tenant_id=tenant_id, status=StatusEnum.VALID.value):
-        return get_error_data_result(retmsg="You do not own the chat")
-    msg = []
-    question = {
-        "content": req.get("question"),
-        "role": "user",
-        "id": str(uuid4())
-    }
-    conv.message.append(question)
-    for m in conv.message:
-        if m["role"] == "system": continue
-        if m["role"] == "assistant" and not msg: continue
-        msg.append(m)
-    message_id = msg[-1].get("id")
-    e, dia = DialogService.get_by_id(conv.dialog_id)
-
-    if not conv.reference:
-        conv.reference = []
-    conv.message.append({"role": "assistant", "content": "", "id": message_id})
-    conv.reference.append({"chunks": [], "doc_aggs": []})
-
-    def fillin_conv(ans):
-        nonlocal conv, message_id
-        if not conv.reference:
-            conv.reference.append(ans["reference"])
-        else:
-            conv.reference[-1] = ans["reference"]
-        conv.message[-1] = {"role": "assistant", "content": ans["answer"],
-                            "id": message_id, "prompt": ans.get("prompt", "")}
-        ans["id"] = message_id
-        ans["session_id"]=session_id
-
-    def stream():
-        nonlocal dia, msg, req, conv
-        try:
-            for ans in chat(dia, msg, **req):
-                fillin_conv(ans)
-                yield "data:" + json.dumps({"code": 0,  "data": ans}, ensure_ascii=False) + "\n\n"
-            ConversationService.update_by_id(conv.id, conv.to_dict())
-        except Exception as e:
-            yield "data:" + json.dumps({"code": 500, "message": str(e),
-                                        "data": {"answer": "**ERROR**: " + str(e),"reference": []}},
-                                       ensure_ascii=False) + "\n\n"
-        yield "data:" + json.dumps({"code": 0, "data": True}, ensure_ascii=False) + "\n\n"
-
+    if not req or not req.get("session_id"):
+        req = {"question": ""}
+    if not DialogService.query(tenant_id=tenant_id, id=chat_id, status=StatusEnum.VALID.value):
+        return get_error_data_result(f"You don't own the chat {chat_id}")
+    if req.get("session_id"):
+        if not ConversationService.query(id=req["session_id"], dialog_id=chat_id):
+            return get_error_data_result(f"You don't own the session {req['session_id']}")
     if req.get("stream", True):
-        resp = Response(stream(), mimetype="text/event-stream")
+        resp = Response(rag_completion(tenant_id, chat_id, **req), mimetype="text/event-stream")
+        resp.headers.add_header("Cache-control", "no-cache")
+        resp.headers.add_header("Connection", "keep-alive")
+        resp.headers.add_header("X-Accel-Buffering", "no")
+        resp.headers.add_header("Content-Type", "text/event-stream; charset=utf-8")
+
+        return resp
+    else:
+        answer = None
+        for ans in rag_completion(tenant_id, chat_id, **req):
+            answer = ans
+            break
+        return get_result(data=answer)
+
+
+@manager.route('/agents/<agent_id>/completions', methods=['POST'])  # noqa: F821
+@token_required
+def agent_completions(tenant_id, agent_id):
+    req = request.json
+    cvs = UserCanvasService.query(user_id=tenant_id, id=agent_id)
+    if not cvs:
+        return get_error_data_result(f"You don't own the agent {agent_id}")
+    if req.get("session_id"):
+        dsl = cvs[0].dsl
+        if not isinstance(dsl, str):
+            dsl = json.dumps(dsl)
+        #canvas = Canvas(dsl, tenant_id)
+        #if canvas.get_preset_param():
+        #    req["question"] = ""
+        conv = API4ConversationService.query(id=req["session_id"], dialog_id=agent_id)
+        if not conv:
+            return get_error_data_result(f"You don't own the session {req['session_id']}")
+    else:
+        req["question"] = ""
+    if req.get("stream", True):
+        resp = Response(agent_completion(tenant_id, agent_id, **req), mimetype="text/event-stream")
         resp.headers.add_header("Cache-control", "no-cache")
         resp.headers.add_header("Connection", "keep-alive")
         resp.headers.add_header("X-Accel-Buffering", "no")
         resp.headers.add_header("Content-Type", "text/event-stream; charset=utf-8")
         return resp
-
-    else:
-        answer = None
-        for ans in chat(dia, msg, **req):
-            answer = ans
-            fillin_conv(ans)
-            ConversationService.update_by_id(conv.id, conv.to_dict())
-            break
+    try:
+        for answer in agent_completion(tenant_id, agent_id, **req):
             return get_result(data=answer)
+    except Exception as e:
+        return get_error_data_result(str(e))
 
-@manager.route('/chats/<chat_id>/sessions', methods=['GET'])
+
+@manager.route('/chats/<chat_id>/sessions', methods=['GET'])  # noqa: F821
 @token_required
-def list(chat_id,tenant_id):
+def list_session(tenant_id, chat_id):
     if not DialogService.query(tenant_id=tenant_id, id=chat_id, status=StatusEnum.VALID.value):
-        return get_error_data_result(retmsg=f"You don't own the assistant {chat_id}.")
+        return get_error_data_result(message=f"You don't own the assistant {chat_id}.")
     id = request.args.get("id")
     name = request.args.get("name")
     page_number = int(request.args.get("page", 1))
-    items_per_page = int(request.args.get("page_size", 1024))
+    items_per_page = int(request.args.get("page_size", 30))
     orderby = request.args.get("orderby", "create_time")
+    user_id = request.args.get("user_id")
     if request.args.get("desc") == "False" or request.args.get("desc") == "false":
         desc = False
     else:
         desc = True
-    convs = ConversationService.get_list(chat_id,page_number,items_per_page,orderby,desc,id,name)
+    convs = ConversationService.get_list(chat_id, page_number, items_per_page, orderby, desc, id, name, user_id)
     if not convs:
         return get_result(data=[])
     for conv in convs:
@@ -181,7 +216,7 @@ def list(chat_id,tenant_id):
         for info in infos:
             if "prompt" in info:
                 info.pop("prompt")
-        conv["chat"] = conv.pop("dialog_id")
+        conv["chat_id"] = conv.pop("dialog_id")
         if conv["reference"]:
             messages = conv["messages"]
             message_num = 0
@@ -193,16 +228,65 @@ def list(chat_id,tenant_id):
                         chunks = conv["reference"][chunk_num]["chunks"]
                         for chunk in chunks:
                             new_chunk = {
-                                "id": chunk["chunk_id"],
-                                "content": chunk["content_with_weight"],
-                                "document_id": chunk["doc_id"],
-                                "document_name": chunk["docnm_kwd"],
-                                "dataset_id": chunk["kb_id"],
-                                "image_id": chunk["img_id"],
-                                "similarity": chunk["similarity"],
-                                "vector_similarity": chunk["vector_similarity"],
-                                "term_similarity": chunk["term_similarity"],
-                                "positions": chunk["positions"],
+                                "id": chunk.get("chunk_id", chunk.get("id")),
+                                "content": chunk.get("content_with_weight", chunk.get("content")),
+                                "document_id": chunk.get("doc_id", chunk.get("document_id")),
+                                "document_name": chunk.get("docnm_kwd", chunk.get("document_name")),
+                                "dataset_id": chunk.get("kb_id", chunk.get("dataset_id")),
+                                "image_id": chunk.get("image_id", chunk.get("img_id")),
+                                "positions": chunk.get("positions", chunk.get("position_int")),
+                            }
+
+                            chunk_list.append(new_chunk)
+                    chunk_num += 1
+                    messages[message_num]["reference"] = chunk_list
+                message_num += 1
+        del conv["reference"]
+    return get_result(data=convs)
+
+
+@manager.route('/agents/<agent_id>/sessions', methods=['GET'])  # noqa: F821
+@token_required
+def list_agent_session(tenant_id, agent_id):
+    if not UserCanvasService.query(user_id=tenant_id, id=agent_id):
+        return get_error_data_result(message=f"You don't own the agent {agent_id}.")
+    id = request.args.get("id")
+    user_id = request.args.get("user_id")
+    page_number = int(request.args.get("page", 1))
+    items_per_page = int(request.args.get("page_size", 30))
+    orderby = request.args.get("orderby", "update_time")
+    if request.args.get("desc") == "False" or request.args.get("desc") == "false":
+        desc = False
+    else:
+        desc = True
+    convs = API4ConversationService.get_list(agent_id, tenant_id, page_number, items_per_page, orderby, desc, id, user_id)
+    if not convs:
+        return get_result(data=[])
+    for conv in convs:
+        conv['messages'] = conv.pop("message")
+        infos = conv["messages"]
+        for info in infos:
+            if "prompt" in info:
+                info.pop("prompt")
+        conv["agent_id"] = conv.pop("dialog_id")
+        if conv["reference"]:
+            messages = conv["messages"]
+            message_num = 0
+            chunk_num = 0
+            while message_num < len(messages):
+                if message_num != 0 and messages[message_num]["role"] != "user":
+                    chunk_list = []
+                    if "chunks" in conv["reference"][chunk_num]:
+                        chunks = conv["reference"][chunk_num]["chunks"]
+                        for chunk in chunks:
+                            new_chunk = {
+                                "id": chunk.get("chunk_id", chunk.get("id")),
+                                "content": chunk.get("content_with_weight", chunk.get("content")),
+                                "document_id": chunk.get("doc_id", chunk.get("document_id")),
+                                "document_name": chunk.get("docnm_kwd", chunk.get("document_name")),
+                                "dataset_id": chunk.get("kb_id", chunk.get("dataset_id")),
+                                "image_id": chunk.get("image_id", chunk.get("img_id")),
+                                "positions": chunk.get("positions", chunk.get("position_int")),
                             }
                             chunk_list.append(new_chunk)
                     chunk_num += 1
@@ -211,11 +295,12 @@ def list(chat_id,tenant_id):
         del conv["reference"]
     return get_result(data=convs)
 
-@manager.route('/chats/<chat_id>/sessions', methods=["DELETE"])
+
+@manager.route('/chats/<chat_id>/sessions', methods=["DELETE"])  # noqa: F821
 @token_required
 def delete(tenant_id, chat_id):
     if not DialogService.query(id=chat_id, tenant_id=tenant_id, status=StatusEnum.VALID.value):
-        return get_error_data_result(retmsg="You don't own the chat")
+        return get_error_data_result(message="You don't own the chat")
     req = request.json
     convs = ConversationService.query(dialog_id=chat_id)
     if not req:
@@ -232,6 +317,138 @@ def delete(tenant_id,chat_id):
     for id in conv_list:
         conv = ConversationService.query(id=id, dialog_id=chat_id)
         if not conv:
-            return get_error_data_result(retmsg="The chat doesn't own the session")
+            return get_error_data_result(message="The chat doesn't own the session")
         ConversationService.delete_by_id(id)
     return get_result()
+
+
+@manager.route('/sessions/ask', methods=['POST'])  # noqa: F821
+@token_required
+def ask_about(tenant_id):
+    req = request.json
+    if not req.get("question"):
+        return get_error_data_result("`question` is required.")
+    if not req.get("dataset_ids"):
+        return get_error_data_result("`dataset_ids` is required.")
+    if not isinstance(req.get("dataset_ids"), list):
+        return get_error_data_result("`dataset_ids` should be a list.")
+    req["kb_ids"] = req.pop("dataset_ids")
+    for kb_id in req["kb_ids"]:
+        if not KnowledgebaseService.accessible(kb_id, tenant_id):
+            return get_error_data_result(f"You don't own the dataset {kb_id}.")
+        kbs = KnowledgebaseService.query(id=kb_id)
+        kb = kbs[0]
+        if kb.chunk_num == 0:
+            return get_error_data_result(f"The dataset {kb_id} doesn't own parsed file")
+    uid = tenant_id
+
+    def stream():
+        nonlocal req, uid
+        try:
+            for ans in ask(req["question"], req["kb_ids"], uid):
+                yield "data:" + json.dumps({"code": 0, "message": "", "data": ans}, ensure_ascii=False) + "\n\n"
+        except Exception as e:
+            yield "data:" + json.dumps({"code": 500, "message": str(e),
+                                        "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
+                                       ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+
+    resp = Response(stream(), mimetype="text/event-stream")
+    resp.headers.add_header("Cache-control", "no-cache")
+    resp.headers.add_header("Connection", "keep-alive")
+    resp.headers.add_header("X-Accel-Buffering", "no")
+    resp.headers.add_header("Content-Type", "text/event-stream; charset=utf-8")
+    return resp
+
+
+@manager.route('/sessions/related_questions', methods=['POST'])  # noqa: F821
+@token_required
+def related_questions(tenant_id):
+    req = request.json
+    if not req.get("question"):
+        return get_error_data_result("`question` is required.")
+    question = req["question"]
+    chat_mdl = LLMBundle(tenant_id, LLMType.CHAT)
+    prompt = """
+Objective: To generate search terms related to the user's search keywords, helping users find more valuable information.
+Instructions:
+ - Based on the keywords provided by the user, generate 5-10 related search terms.
+ - Each search term should be directly or indirectly related to the keyword, guiding the user to find more valuable information.
+ - Use common, general terms as much as possible, avoiding obscure words or technical jargon.
+ - Keep the term length between 2-4 words, concise and clear.
+ - DO NOT translate, use the language of the original keywords.
+
+### Example:
+Keywords: Chinese football
+Related search terms:
+1. Current status of Chinese football
+2. Reform of Chinese football
+3. Youth training of Chinese football
+4. Chinese football in the Asian Cup
+5. Chinese football in the World Cup
+
+Reason:
+ - When searching, users often only use one or two keywords, making it difficult to fully express their information needs.
+ - Generating related search terms can help users dig deeper into relevant information and improve search efficiency. 
+ - At the same time, related terms can also help search engines better understand user needs and return more accurate search results.
+
+"""
+    ans = chat_mdl.chat(prompt, [{"role": "user", "content": f"""
+Keywords: {question}
+Related search terms:
+    """}], {"temperature": 0.9})
+    return get_result(data=[re.sub(r"^[0-9]\. ", "", a) for a in ans.split("\n") if re.match(r"^[0-9]\. ", a)])
+
+
+@manager.route('/chatbots/<dialog_id>/completions', methods=['POST'])  # noqa: F821
+def chatbot_completions(dialog_id):
+    req = request.json
+
+    token = request.headers.get('Authorization').split()
+    if len(token) != 2:
+        return get_error_data_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_error_data_result(message='Authentication error: API key is invalid!"')
+
+    if "quote" not in req:
+        req["quote"] = False
+
+    if req.get("stream", True):
+        resp = Response(iframe_completion(dialog_id, **req), mimetype="text/event-stream")
+        resp.headers.add_header("Cache-control", "no-cache")
+        resp.headers.add_header("Connection", "keep-alive")
+        resp.headers.add_header("X-Accel-Buffering", "no")
+        resp.headers.add_header("Content-Type", "text/event-stream; charset=utf-8")
+        return resp
+
+    for answer in iframe_completion(dialog_id, **req):
+        return get_result(data=answer)
+
+
+@manager.route('/agentbots/<agent_id>/completions', methods=['POST'])  # noqa: F821
+def agent_bot_completions(agent_id):
+    req = request.json
+
+    token = request.headers.get('Authorization').split()
+    if len(token) != 2:
+        return get_error_data_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_error_data_result(message='Authentication error: API key is invalid!"')
+
+    if "quote" not in req:
+        req["quote"] = False
+
+    if req.get("stream", True):
+        resp = Response(agent_completion(objs[0].tenant_id, agent_id, **req), mimetype="text/event-stream")
+        resp.headers.add_header("Cache-control", "no-cache")
+        resp.headers.add_header("Connection", "keep-alive")
+        resp.headers.add_header("X-Accel-Buffering", "no")
+        resp.headers.add_header("Content-Type", "text/event-stream; charset=utf-8")
+        return resp
+
+    for answer in agent_completion(objs[0].tenant_id, agent_id, **req):
+        return get_result(data=answer)

api/apps/system_app.py

@@ -13,8 +13,9 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License
 #
-import json
+import logging
 from datetime import datetime
+import json
 
 from flask_login import login_required, current_user
 
@@ -22,120 +23,278 @@ from api.db.db_models import APIToken
 from api.db.services.api_service import APITokenService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.user_service import UserTenantService
-from api.settings import DATABASE_TYPE
+from api import settings
 from api.utils import current_timestamp, datetime_format
-from api.utils.api_utils import get_json_result, get_data_error_result, server_error_response, \
-    generate_confirmation_token, request, validate_request
-from api.versions import get_rag_version
-from rag.utils.es_conn import ELASTICSEARCH
+from api.utils.api_utils import (
+    get_json_result,
+    get_data_error_result,
+    server_error_response,
+    generate_confirmation_token,
+)
+from api.versions import get_ragflow_version
 from rag.utils.storage_factory import STORAGE_IMPL, STORAGE_IMPL_TYPE
 from timeit import default_timer as timer
 
 from rag.utils.redis_conn import REDIS_CONN
 
 
-@manager.route('/version', methods=['GET'])
+@manager.route("/version", methods=["GET"])  # noqa: F821
 @login_required
 def version():
-    return get_json_result(data=get_rag_version())
+    """
+    Get the current version of the application.
+    ---
+    tags:
+      - System
+    security:
+      - ApiKeyAuth: []
+    responses:
+      200:
+        description: Version retrieved successfully.
+        schema:
+          type: object
+          properties:
+            version:
+              type: string
+              description: Version number.
+    """
+    return get_json_result(data=get_ragflow_version())
 
 
-@manager.route('/status', methods=['GET'])
+@manager.route("/status", methods=["GET"])  # noqa: F821
 @login_required
 def status():
+    """
+    Get the system status.
+    ---
+    tags:
+      - System
+    security:
+      - ApiKeyAuth: []
+    responses:
+      200:
+        description: System is operational.
+        schema:
+          type: object
+          properties:
+            es:
+              type: object
+              description: Elasticsearch status.
+            storage:
+              type: object
+              description: Storage status.
+            database:
+              type: object
+              description: Database status.
+      503:
+        description: Service unavailable.
+        schema:
+          type: object
+          properties:
+            error:
+              type: string
+              description: Error message.
+    """
     res = {}
     st = timer()
     try:
-        res["es"] = ELASTICSEARCH.health()
-        res["es"]["elapsed"] = "{:.1f}".format((timer() - st)*1000.)
+        res["doc_engine"] = settings.docStoreConn.health()
+        res["doc_engine"]["elapsed"] = "{:.1f}".format((timer() - st) * 1000.0)
     except Exception as e:
-        res["es"] = {"status": "red", "elapsed": "{:.1f}".format((timer() - st)*1000.), "error": str(e)}
+        res["doc_engine"] = {
+            "type": "unknown",
+            "status": "red",
+            "elapsed": "{:.1f}".format((timer() - st) * 1000.0),
+            "error": str(e),
+        }
 
     st = timer()
     try:
         STORAGE_IMPL.health()
-        res["storage"] = {"storage": STORAGE_IMPL_TYPE.lower(), "status": "green", "elapsed": "{:.1f}".format((timer() - st)*1000.)}
+        res["storage"] = {
+            "storage": STORAGE_IMPL_TYPE.lower(),
+            "status": "green",
+            "elapsed": "{:.1f}".format((timer() - st) * 1000.0),
+        }
     except Exception as e:
-        res["storage"] = {"storage": STORAGE_IMPL_TYPE.lower(), "status": "red", "elapsed": "{:.1f}".format((timer() - st)*1000.), "error": str(e)}
+        res["storage"] = {
+            "storage": STORAGE_IMPL_TYPE.lower(),
+            "status": "red",
+            "elapsed": "{:.1f}".format((timer() - st) * 1000.0),
+            "error": str(e),
+        }
 
     st = timer()
     try:
         KnowledgebaseService.get_by_id("x")
-        res["database"] = {"database": DATABASE_TYPE.lower(), "status": "green", "elapsed": "{:.1f}".format((timer() - st)*1000.)}
+        res["database"] = {
+            "database": settings.DATABASE_TYPE.lower(),
+            "status": "green",
+            "elapsed": "{:.1f}".format((timer() - st) * 1000.0),
+        }
     except Exception as e:
-        res["database"] = {"database": DATABASE_TYPE.lower(), "status": "red", "elapsed": "{:.1f}".format((timer() - st)*1000.), "error": str(e)}
+        res["database"] = {
+            "database": settings.DATABASE_TYPE.lower(),
+            "status": "red",
+            "elapsed": "{:.1f}".format((timer() - st) * 1000.0),
+            "error": str(e),
+        }
 
     st = timer()
     try:
         if not REDIS_CONN.health():
             raise Exception("Lost connection!")
-        res["redis"] = {"status": "green", "elapsed": "{:.1f}".format((timer() - st)*1000.)}
+        res["redis"] = {
+            "status": "green",
+            "elapsed": "{:.1f}".format((timer() - st) * 1000.0),
+        }
     except Exception as e:
-        res["redis"] = {"status": "red", "elapsed": "{:.1f}".format((timer() - st)*1000.), "error": str(e)}
+        res["redis"] = {
+            "status": "red",
+            "elapsed": "{:.1f}".format((timer() - st) * 1000.0),
+            "error": str(e),
+        }
 
+    task_executor_heartbeats = {}
     try:
-        v = REDIS_CONN.get("TASKEXE")
-        if not v:
-            raise Exception("No task executor running!")
-        obj = json.loads(v)
-        color = "green"
-        for id in obj.keys():
-            arr = obj[id]
-            if len(arr) == 1:
-                obj[id] = [0]
-            else:
-                obj[id] = [arr[i+1]-arr[i] for i in range(len(arr)-1)]
-            elapsed = max(obj[id])
-            if elapsed > 50: color = "yellow"
-            if elapsed > 120: color = "red"
-        res["task_executor"] = {"status": color, "elapsed": obj}
-    except Exception as e:
-        res["task_executor"] = {"status": "red", "error": str(e)}
+        task_executors = REDIS_CONN.smembers("TASKEXE")
+        now = datetime.now().timestamp()
+        for task_executor_id in task_executors:
+            heartbeats = REDIS_CONN.zrangebyscore(task_executor_id, now - 60*30, now)
+            heartbeats = [json.loads(heartbeat) for heartbeat in heartbeats]
+            task_executor_heartbeats[task_executor_id] = heartbeats
+    except Exception:
+        logging.exception("get task executor heartbeats failed!")
+    res["task_executor_heartbeats"] = task_executor_heartbeats
 
     return get_json_result(data=res)
 
 
-@manager.route('/new_token', methods=['POST'])
+@manager.route("/new_token", methods=["POST"])  # noqa: F821
 @login_required
 def new_token():
+    """
+    Generate a new API token.
+    ---
+    tags:
+      - API Tokens
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: query
+        name: name
+        type: string
+        required: false
+        description: Name of the token.
+    responses:
+      200:
+        description: Token generated successfully.
+        schema:
+          type: object
+          properties:
+            token:
+              type: string
+              description: The generated API token.
+    """
     try:
         tenants = UserTenantService.query(user_id=current_user.id)
         if not tenants:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
 
         tenant_id = tenants[0].tenant_id
-        obj = {"tenant_id": tenant_id, "token": generate_confirmation_token(tenant_id),
+        obj = {
+            "tenant_id": tenant_id,
+            "token": generate_confirmation_token(tenant_id),
+            "beta": generate_confirmation_token(generate_confirmation_token(tenant_id)).replace("ragflow-", "")[:32],
             "create_time": current_timestamp(),
             "create_date": datetime_format(datetime.now()),
             "update_time": None,
-               "update_date": None
+            "update_date": None,
         }
 
         if not APITokenService.save(**obj):
-            return get_data_error_result(retmsg="Fail to new a dialog!")
+            return get_data_error_result(message="Fail to new a dialog!")
 
         return get_json_result(data=obj)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/token_list', methods=['GET'])
+@manager.route("/token_list", methods=["GET"])  # noqa: F821
 @login_required
 def token_list():
+    """
+    List all API tokens for the current user.
+    ---
+    tags:
+      - API Tokens
+    security:
+      - ApiKeyAuth: []
+    responses:
+      200:
+        description: List of API tokens.
+        schema:
+          type: object
+          properties:
+            tokens:
+              type: array
+              items:
+                type: object
+                properties:
+                  token:
+                    type: string
+                    description: The API token.
+                  name:
+                    type: string
+                    description: Name of the token.
+                  create_time:
+                    type: string
+                    description: Token creation time.
+    """
     try:
         tenants = UserTenantService.query(user_id=current_user.id)
         if not tenants:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
 
-        objs = APITokenService.query(tenant_id=tenants[0].tenant_id)
-        return get_json_result(data=[o.to_dict() for o in objs])
+        tenant_id = tenants[0].tenant_id
+        objs = APITokenService.query(tenant_id=tenant_id)
+        objs = [o.to_dict() for o in objs]
+        for o in objs:
+            if not o["beta"]:
+                o["beta"] = generate_confirmation_token(generate_confirmation_token(tenants[0].tenant_id)).replace("ragflow-", "")[:32]
+                APITokenService.filter_update([APIToken.tenant_id == tenant_id, APIToken.token == o["token"]], o)
+        return get_json_result(data=objs)
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route('/token/<token>', methods=['DELETE'])
+@manager.route("/token/<token>", methods=["DELETE"])  # noqa: F821
 @login_required
 def rm(token):
+    """
+    Remove an API token.
+    ---
+    tags:
+      - API Tokens
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: path
+        name: token
+        type: string
+        required: true
+        description: The API token to remove.
+    responses:
+      200:
+        description: Token removed successfully.
+        schema:
+          type: object
+          properties:
+            success:
+              type: boolean
+              description: Deletion status.
+    """
     APITokenService.filter_delete(
-                [APIToken.tenant_id == current_user.id, APIToken.token == token])
+        [APIToken.tenant_id == current_user.id, APIToken.token == token]
+    )
     return get_json_result(data=True)

api/apps/tenant_app.py

@@ -17,6 +17,7 @@
 from flask import request
 from flask_login import login_required, current_user
 
+from api import settings
 from api.db import UserTenantRole, StatusEnum
 from api.db.db_models import UserTenant
 from api.db.services.user_service import UserTenantService, UserService
@@ -25,9 +26,15 @@ from api.utils import get_uuid, delta_seconds
 from api.utils.api_utils import get_json_result, validate_request, server_error_response, get_data_error_result
 
 
-@manager.route("/<tenant_id>/user/list", methods=["GET"])
+@manager.route("/<tenant_id>/user/list", methods=["GET"])  # noqa: F821
 @login_required
 def user_list(tenant_id):
+    if current_user.id != tenant_id:
+        return get_json_result(
+            data=False,
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR)
+
     try:
         users = UserTenantService.get_by_tenant_id(tenant_id)
         for u in users:
@@ -37,39 +44,55 @@ def user_list(tenant_id):
         return server_error_response(e)
 
 
-@manager.route('/<tenant_id>/user', methods=['POST'])
+@manager.route('/<tenant_id>/user', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("email")
 def create(tenant_id):
-    req = request.json
-    usrs = UserService.query(email=req["email"])
-    if not usrs:
-        return get_data_error_result(retmsg="User not found.")
+    if current_user.id != tenant_id:
+        return get_json_result(
+            data=False,
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR)
 
-    user_id = usrs[0].id
-    user_tenants = UserTenantService.query(user_id=user_id, tenant_id=tenant_id)
+    req = request.json
+    invite_user_email = req["email"]
+    invite_users = UserService.query(email=invite_user_email)
+    if not invite_users:
+        return get_data_error_result(message="User not found.")
+
+    user_id_to_invite = invite_users[0].id
+    user_tenants = UserTenantService.query(user_id=user_id_to_invite, tenant_id=tenant_id)
     if user_tenants:
-        if user_tenants[0].status == UserTenantRole.NORMAL.value:
-            return get_data_error_result(retmsg="This user is in the team already.")
-        return get_data_error_result(retmsg="Invitation notification is sent.")
+        user_tenant_role = user_tenants[0].role
+        if user_tenant_role == UserTenantRole.NORMAL:
+            return get_data_error_result(message=f"{invite_user_email} is already in the team.")
+        if user_tenant_role == UserTenantRole.OWNER:
+            return get_data_error_result(message=f"{invite_user_email} is the owner of the team.")
+        return get_data_error_result(message=f"{invite_user_email} is in the team, but the role: {user_tenant_role} is invalid.")
 
     UserTenantService.save(
         id=get_uuid(),
-        user_id=user_id,
+        user_id=user_id_to_invite,
         tenant_id=tenant_id,
         invited_by=current_user.id,
         role=UserTenantRole.INVITE,
         status=StatusEnum.VALID.value)
 
-    usr = usrs[0].to_dict()
+    usr = invite_users[0].to_dict()
     usr = {k: v for k, v in usr.items() if k in ["id", "avatar", "email", "nickname"]}
 
     return get_json_result(data=usr)
 
 
-@manager.route('/<tenant_id>/user/<user_id>', methods=['DELETE'])
+@manager.route('/<tenant_id>/user/<user_id>', methods=['DELETE'])  # noqa: F821
 @login_required
 def rm(tenant_id, user_id):
+    if current_user.id != tenant_id and current_user.id != user_id:
+        return get_json_result(
+            data=False,
+            message='No authorization.',
+            code=settings.RetCode.AUTHENTICATION_ERROR)
+
     try:
         UserTenantService.filter_delete([UserTenant.tenant_id == tenant_id, UserTenant.user_id == user_id])
         return get_json_result(data=True)
@@ -77,7 +100,7 @@ def rm(tenant_id, user_id):
         return server_error_response(e)
 
 
-@manager.route("/list", methods=["GET"])
+@manager.route("/list", methods=["GET"])  # noqa: F821
 @login_required
 def tenant_list():
     try:
@@ -89,7 +112,7 @@ def tenant_list():
         return server_error_response(e)
 
 
-@manager.route("/agree/<tenant_id>", methods=["PUT"])
+@manager.route("/agree/<tenant_id>", methods=["PUT"])  # noqa: F821
 @login_required
 def agree(tenant_id):
     try:

api/apps/user_app.py

@@ -13,6 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 import json
 import re
 from datetime import datetime
@@ -23,65 +24,127 @@ from flask_login import login_required, current_user, login_user, logout_user
 
 from api.db.db_models import TenantLLM
 from api.db.services.llm_service import TenantLLMService, LLMService
-from api.utils.api_utils import server_error_response, validate_request, get_data_error_result
-from api.utils import get_uuid, get_format_time, decrypt, download_img, current_timestamp, datetime_format
-from api.db import UserTenantRole, LLMType, FileType
-from api.settings import RetCode, GITHUB_OAUTH, FEISHU_OAUTH, CHAT_MDL, EMBEDDING_MDL, ASR_MDL, IMAGE2TEXT_MDL, PARSERS, \
-    API_KEY, \
-    LLM_FACTORY, LLM_BASE_URL, RERANK_MDL
+from api.utils.api_utils import (
+    server_error_response,
+    validate_request,
+    get_data_error_result,
+)
+from api.utils import (
+    get_uuid,
+    get_format_time,
+    decrypt,
+    download_img,
+    current_timestamp,
+    datetime_format,
+)
+from api.db import UserTenantRole, FileType
+from api import settings
 from api.db.services.user_service import UserService, TenantService, UserTenantService
 from api.db.services.file_service import FileService
-from api.settings import stat_logger
 from api.utils.api_utils import get_json_result, construct_response
 
 
-@manager.route('/login', methods=['POST', 'GET'])
+@manager.route("/login", methods=["POST", "GET"])  # noqa: F821
 def login():
+    """
+    User login endpoint.
+    ---
+    tags:
+      - User
+    parameters:
+      - in: body
+        name: body
+        description: Login credentials.
+        required: true
+        schema:
+          type: object
+          properties:
+            email:
+              type: string
+              description: User email.
+            password:
+              type: string
+              description: User password.
+    responses:
+      200:
+        description: Login successful.
+        schema:
+          type: object
+      401:
+        description: Authentication failed.
+        schema:
+          type: object
+    """
     if not request.json:
-        return get_json_result(data=False,
-                               retcode=RetCode.AUTHENTICATION_ERROR,
-                               retmsg='Unauthorized!')
+        return get_json_result(
+            data=False, code=settings.RetCode.AUTHENTICATION_ERROR, message="Unauthorized!"
+        )
 
-    email = request.json.get('email', "")
+    email = request.json.get("email", "")
     users = UserService.query(email=email)
     if not users:
-        return get_json_result(data=False,
-                               retcode=RetCode.AUTHENTICATION_ERROR,
-                               retmsg=f'Email: {email} is not registered!')
+        return get_json_result(
+            data=False,
+            code=settings.RetCode.AUTHENTICATION_ERROR,
+            message=f"Email: {email} is not registered!",
+        )
 
-    password = request.json.get('password')
+    password = request.json.get("password")
     try:
         password = decrypt(password)
     except BaseException:
-        return get_json_result(data=False,
-                               retcode=RetCode.SERVER_ERROR,
-                               retmsg='Fail to crypt password')
+        return get_json_result(
+            data=False, code=settings.RetCode.SERVER_ERROR, message="Fail to crypt password"
+        )
 
     user = UserService.query_user(email, password)
     if user:
         response_data = user.to_json()
         user.access_token = get_uuid()
         login_user(user)
-        user.update_time = current_timestamp(),
-        user.update_date = datetime_format(datetime.now()),
+        user.update_time = (current_timestamp(),)
+        user.update_date = (datetime_format(datetime.now()),)
         user.save()
         msg = "Welcome back!"
-        return construct_response(data=response_data, auth=user.get_id(), retmsg=msg)
+        return construct_response(data=response_data, auth=user.get_id(), message=msg)
     else:
-        return get_json_result(data=False,
-                               retcode=RetCode.AUTHENTICATION_ERROR,
-                               retmsg='Email and password do not match!')
+        return get_json_result(
+            data=False,
+            code=settings.RetCode.AUTHENTICATION_ERROR,
+            message="Email and password do not match!",
+        )
 
 
-@manager.route('/github_callback', methods=['GET'])
+@manager.route("/github_callback", methods=["GET"])  # noqa: F821
 def github_callback():
+    """
+    GitHub OAuth callback endpoint.
+    ---
+    tags:
+      - OAuth
+    parameters:
+      - in: query
+        name: code
+        type: string
+        required: true
+        description: Authorization code from GitHub.
+    responses:
+      200:
+        description: Authentication successful.
+        schema:
+          type: object
+    """
     import requests
-    res = requests.post(GITHUB_OAUTH.get("url"),
+
+    res = requests.post(
+        settings.GITHUB_OAUTH.get("url"),
         data={
-                            "client_id": GITHUB_OAUTH.get("client_id"),
-                            "client_secret": GITHUB_OAUTH.get("secret_key"),
-                            "code": request.args.get('code')},
-                        headers={"Accept": "application/json"})
+            "client_id": settings.GITHUB_OAUTH.get("client_id"),
+            "client_secret": settings.GITHUB_OAUTH.get("secret_key"),
+            "code": request.args.get("code"),
+        },
+        headers={"Accept": "application/json"},
+    )
     res = res.json()
     if "error" in res:
         return redirect("/?error=%s" % res["error_description"])
@@ -101,9 +164,11 @@ def github_callback():
             try:
                 avatar = download_img(user_info["avatar_url"])
             except Exception as e:
-                stat_logger.exception(e)
+                logging.exception(e)
                 avatar = ""
-            users = user_register(user_id, {
+            users = user_register(
+                user_id,
+                {
                     "access_token": session["access_token"],
                     "email": email_address,
                     "avatar": avatar,
@@ -111,11 +176,12 @@ def github_callback():
                     "login_channel": "github",
                     "last_login_time": get_format_time(),
                     "is_superuser": False,
-            })
+                },
+            )
             if not users:
-                raise Exception(f'Fail to register {email_address}.')
+                raise Exception(f"Fail to register {email_address}.")
             if len(users) > 1:
-                raise Exception(f'Same email: {email_address} exists!')
+                raise Exception(f"Same email: {email_address} exists!")
 
             # Try to log in
             user = users[0]
@@ -123,7 +189,7 @@ def github_callback():
             return redirect("/?auth=%s" % user.get_id())
         except Exception as e:
             rollback_user_registration(user_id)
-            stat_logger.exception(e)
+            logging.exception(e)
             return redirect("/?error=%s" % str(e))
 
     # User has already registered, try to log in
@@ -134,33 +200,59 @@ def github_callback():
     return redirect("/?auth=%s" % user.get_id())
 
 
-@manager.route('/feishu_callback', methods=['GET'])
+@manager.route("/feishu_callback", methods=["GET"])  # noqa: F821
 def feishu_callback():
+    """
+    Feishu OAuth callback endpoint.
+    ---
+    tags:
+      - OAuth
+    parameters:
+      - in: query
+        name: code
+        type: string
+        required: true
+        description: Authorization code from Feishu.
+    responses:
+      200:
+        description: Authentication successful.
+        schema:
+          type: object
+    """
     import requests
-    app_access_token_res = requests.post(FEISHU_OAUTH.get("app_access_token_url"),
-                                         data=json.dumps({
-                                             "app_id": FEISHU_OAUTH.get("app_id"),
-                                             "app_secret": FEISHU_OAUTH.get("app_secret")
-                                         }),
-                                         headers={"Content-Type": "application/json; charset=utf-8"})
+
+    app_access_token_res = requests.post(
+        settings.FEISHU_OAUTH.get("app_access_token_url"),
+        data=json.dumps(
+            {
+                "app_id": settings.FEISHU_OAUTH.get("app_id"),
+                "app_secret": settings.FEISHU_OAUTH.get("app_secret"),
+            }
+        ),
+        headers={"Content-Type": "application/json; charset=utf-8"},
+    )
     app_access_token_res = app_access_token_res.json()
-    if app_access_token_res['code'] != 0:
+    if app_access_token_res["code"] != 0:
         return redirect("/?error=%s" % app_access_token_res)
 
-    res = requests.post(FEISHU_OAUTH.get("user_access_token_url"),
-                        data=json.dumps({
-                            "grant_type": FEISHU_OAUTH.get("grant_type"),
-                            "code": request.args.get('code')
-                        }),
+    res = requests.post(
+        settings.FEISHU_OAUTH.get("user_access_token_url"),
+        data=json.dumps(
+            {
+                "grant_type": settings.FEISHU_OAUTH.get("grant_type"),
+                "code": request.args.get("code"),
+            }
+        ),
         headers={
             "Content-Type": "application/json; charset=utf-8",
-                            'Authorization': f"Bearer {app_access_token_res['app_access_token']}"
-                        })
+            "Authorization": f"Bearer {app_access_token_res['app_access_token']}",
+        },
+    )
     res = res.json()
-    if res['code'] != 0:
+    if res["code"] != 0:
         return redirect("/?error=%s" % res["message"])
 
-    if "contact:user.email:readonly" not in res["data"]["scope"].split(" "):
+    if "contact:user.email:readonly" not in res["data"]["scope"].split():
         return redirect("/?error=contact:user.email:readonly not in scope")
     session["access_token"] = res["data"]["access_token"]
     session["access_token_from"] = "feishu"
@@ -174,9 +266,11 @@ def feishu_callback():
             try:
                 avatar = download_img(user_info["avatar_url"])
             except Exception as e:
-                stat_logger.exception(e)
+                logging.exception(e)
                 avatar = ""
-            users = user_register(user_id, {
+            users = user_register(
+                user_id,
+                {
                     "access_token": session["access_token"],
                     "email": email_address,
                     "avatar": avatar,
@@ -184,11 +278,12 @@ def feishu_callback():
                     "login_channel": "feishu",
                     "last_login_time": get_format_time(),
                     "is_superuser": False,
-            })
+                },
+            )
             if not users:
-                raise Exception(f'Fail to register {email_address}.')
+                raise Exception(f"Fail to register {email_address}.")
             if len(users) > 1:
-                raise Exception(f'Same email: {email_address} exists!')
+                raise Exception(f"Same email: {email_address} exists!")
 
             # Try to log in
             user = users[0]
@@ -196,7 +291,7 @@ def feishu_callback():
             return redirect("/?auth=%s" % user.get_id())
         except Exception as e:
             rollback_user_registration(user_id)
-            stat_logger.exception(e)
+            logging.exception(e)
             return redirect("/?error=%s" % str(e))
 
     # User has already registered, try to log in
@@ -209,11 +304,14 @@ def feishu_callback():
 
 def user_info_from_feishu(access_token):
     import requests
-    headers = {"Content-Type": "application/json; charset=utf-8",
-               'Authorization': f"Bearer {access_token}"}
+
+    headers = {
+        "Content-Type": "application/json; charset=utf-8",
+        "Authorization": f"Bearer {access_token}",
+    }
     res = requests.get(
-        f"https://open.feishu.cn/open-apis/authen/v1/user_info",
-        headers=headers)
+        "https://open.feishu.cn/open-apis/authen/v1/user_info", headers=headers
+    )
     user_info = res.json()["data"]
     user_info["email"] = None if user_info.get("email") == "" else user_info["email"]
     return user_info
@@ -221,47 +319,103 @@ def user_info_from_feishu(access_token):
 
 def user_info_from_github(access_token):
     import requests
-    headers = {"Accept": "application/json",
-               'Authorization': f"token {access_token}"}
+
+    headers = {"Accept": "application/json", "Authorization": f"token {access_token}"}
     res = requests.get(
-        f"https://api.github.com/user?access_token={access_token}",
-        headers=headers)
+        f"https://api.github.com/user?access_token={access_token}", headers=headers
+    )
     user_info = res.json()
     email_info = requests.get(
         f"https://api.github.com/user/emails?access_token={access_token}",
-        headers=headers).json()
+        headers=headers,
+    ).json()
     user_info["email"] = next(
-        (email for email in email_info if email['primary'] == True),
-        None)["email"]
+        (email for email in email_info if email["primary"]), None
+    )["email"]
     return user_info
 
 
-@manager.route("/logout", methods=['GET'])
+@manager.route("/logout", methods=["GET"])  # noqa: F821
 @login_required
 def log_out():
+    """
+    User logout endpoint.
+    ---
+    tags:
+      - User
+    security:
+      - ApiKeyAuth: []
+    responses:
+      200:
+        description: Logout successful.
+        schema:
+          type: object
+    """
     current_user.access_token = ""
     current_user.save()
     logout_user()
     return get_json_result(data=True)
 
 
-@manager.route("/setting", methods=["POST"])
+@manager.route("/setting", methods=["POST"])  # noqa: F821
 @login_required
 def setting_user():
+    """
+    Update user settings.
+    ---
+    tags:
+      - User
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: body
+        name: body
+        description: User settings to update.
+        required: true
+        schema:
+          type: object
+          properties:
+            nickname:
+              type: string
+              description: New nickname.
+            email:
+              type: string
+              description: New email.
+    responses:
+      200:
+        description: Settings updated successfully.
+        schema:
+          type: object
+    """
     update_dict = {}
     request_data = request.json
     if request_data.get("password"):
         new_password = request_data.get("new_password")
         if not check_password_hash(
-                current_user.password, decrypt(request_data["password"])):
-            return get_json_result(data=False, retcode=RetCode.AUTHENTICATION_ERROR, retmsg='Password error!')
+                current_user.password, decrypt(request_data["password"])
+        ):
+            return get_json_result(
+                data=False,
+                code=settings.RetCode.AUTHENTICATION_ERROR,
+                message="Password error!",
+            )
 
         if new_password:
             update_dict["password"] = generate_password_hash(decrypt(new_password))
 
     for k in request_data.keys():
-        if k in ["password", "new_password", "email", "status", "is_superuser", "login_channel", "is_anonymous",
-                 "is_active", "is_authenticated", "last_login_time"]:
+        if k in [
+            "password",
+            "new_password",
+            "email",
+            "status",
+            "is_superuser",
+            "login_channel",
+            "is_anonymous",
+            "is_active",
+            "is_authenticated",
+            "last_login_time",
+        ]:
             continue
         update_dict[k] = request_data[k]
 
@@ -269,34 +423,59 @@ def setting_user():
         UserService.update_by_id(current_user.id, update_dict)
         return get_json_result(data=True)
     except Exception as e:
-        stat_logger.exception(e)
-        return get_json_result(data=False, retmsg='Update failure!', retcode=RetCode.EXCEPTION_ERROR)
+        logging.exception(e)
+        return get_json_result(
+            data=False, message="Update failure!", code=settings.RetCode.EXCEPTION_ERROR
+        )
 
 
-@manager.route("/info", methods=["GET"])
+@manager.route("/info", methods=["GET"])  # noqa: F821
 @login_required
 def user_profile():
+    """
+    Get user profile information.
+    ---
+    tags:
+      - User
+    security:
+      - ApiKeyAuth: []
+    responses:
+      200:
+        description: User profile retrieved successfully.
+        schema:
+          type: object
+          properties:
+            id:
+              type: string
+              description: User ID.
+            nickname:
+              type: string
+              description: User nickname.
+            email:
+              type: string
+              description: User email.
+    """
     return get_json_result(data=current_user.to_dict())
 
 
 def rollback_user_registration(user_id):
     try:
         UserService.delete_by_id(user_id)
-    except Exception as e:
+    except Exception:
         pass
     try:
         TenantService.delete_by_id(user_id)
-    except Exception as e:
+    except Exception:
         pass
     try:
         u = UserTenantService.query(tenant_id=user_id)
         if u:
             UserTenantService.delete_by_id(u[0].id)
-    except Exception as e:
+    except Exception:
         pass
     try:
         TenantLLM.delete().where(TenantLLM.tenant_id == user_id).execute()
-    except Exception as e:
+    except Exception:
         pass
 
 
@@ -305,18 +484,18 @@ def user_register(user_id, user):
     tenant = {
         "id": user_id,
         "name": user["nickname"] + "‘s Kingdom",
-        "llm_id": CHAT_MDL,
-        "embd_id": EMBEDDING_MDL,
-        "asr_id": ASR_MDL,
-        "parser_ids": PARSERS,
-        "img2txt_id": IMAGE2TEXT_MDL,
-        "rerank_id": RERANK_MDL
+        "llm_id": settings.CHAT_MDL,
+        "embd_id": settings.EMBEDDING_MDL,
+        "asr_id": settings.ASR_MDL,
+        "parser_ids": settings.PARSERS,
+        "img2txt_id": settings.IMAGE2TEXT_MDL,
+        "rerank_id": settings.RERANK_MDL,
     }
     usr_tenant = {
         "tenant_id": user_id,
         "user_id": user_id,
         "invited_by": user_id,
-        "role": UserTenantRole.OWNER
+        "role": UserTenantRole.OWNER,
     }
     file_id = get_uuid()
     file = {
@@ -330,14 +509,18 @@ def user_register(user_id, user):
         "location": "",
     }
     tenant_llm = []
-    for llm in LLMService.query(fid=LLM_FACTORY):
-        tenant_llm.append({"tenant_id": user_id,
-                           "llm_factory": LLM_FACTORY,
+    for llm in LLMService.query(fid=settings.LLM_FACTORY):
+        tenant_llm.append(
+            {
+                "tenant_id": user_id,
+                "llm_factory": settings.LLM_FACTORY,
                 "llm_name": llm.llm_name,
                 "model_type": llm.model_type,
-                           "api_key": API_KEY,
-                           "api_base": LLM_BASE_URL
-                           })
+                "api_key": settings.API_KEY,
+                "api_base": settings.LLM_BASE_URL,
+                "max_tokens": llm.max_tokens if llm.max_tokens else 8192
+            }
+        )
 
     if not UserService.save(**user):
         return
@@ -348,24 +531,55 @@ def user_register(user_id, user):
     return UserService.query(email=user["email"])
 
 
-@manager.route("/register", methods=["POST"])
+@manager.route("/register", methods=["POST"])  # noqa: F821
 @validate_request("nickname", "email", "password")
 def user_add():
+    """
+    Register a new user.
+    ---
+    tags:
+      - User
+    parameters:
+      - in: body
+        name: body
+        description: Registration details.
+        required: true
+        schema:
+          type: object
+          properties:
+            nickname:
+              type: string
+              description: User nickname.
+            email:
+              type: string
+              description: User email.
+            password:
+              type: string
+              description: User password.
+    responses:
+      200:
+        description: Registration successful.
+        schema:
+          type: object
+    """
     req = request.json
     email_address = req["email"]
 
     # Validate the email address
     if not re.match(r"^[\w\._-]+@([\w_-]+\.)+[\w-]{2,5}$", email_address):
-        return get_json_result(data=False,
-                               retmsg=f'Invalid email address: {email_address}!',
-                               retcode=RetCode.OPERATING_ERROR)
+        return get_json_result(
+            data=False,
+            message=f"Invalid email address: {email_address}!",
+            code=settings.RetCode.OPERATING_ERROR,
+        )
 
     # Check if the email address is already used
     if UserService.query(email=email_address):
         return get_json_result(
             data=False,
-            retmsg=f'Email: {email_address} has already registered!',
-            retcode=RetCode.OPERATING_ERROR)
+            message=f"Email: {email_address} has already registered!",
+            code=settings.RetCode.OPERATING_ERROR,
+        )
 
     # Construct user info data
     nickname = req["nickname"]
@@ -383,42 +597,107 @@ def user_add():
     try:
         users = user_register(user_id, user_dict)
         if not users:
-            raise Exception(f'Fail to register {email_address}.')
+            raise Exception(f"Fail to register {email_address}.")
         if len(users) > 1:
-            raise Exception(f'Same email: {email_address} exists!')
+            raise Exception(f"Same email: {email_address} exists!")
         user = users[0]
         login_user(user)
-        return construct_response(data=user.to_json(),
+        return construct_response(
+            data=user.to_json(),
             auth=user.get_id(),
-                                  retmsg=f"{nickname}, welcome aboard!")
+            message=f"{nickname}, welcome aboard!",
+        )
     except Exception as e:
         rollback_user_registration(user_id)
-        stat_logger.exception(e)
-        return get_json_result(data=False,
-                               retmsg=f'User registration failure, error: {str(e)}',
-                               retcode=RetCode.EXCEPTION_ERROR)
+        logging.exception(e)
+        return get_json_result(
+            data=False,
+            message=f"User registration failure, error: {str(e)}",
+            code=settings.RetCode.EXCEPTION_ERROR,
+        )
 
 
-@manager.route("/tenant_info", methods=["GET"])
+@manager.route("/tenant_info", methods=["GET"])  # noqa: F821
 @login_required
 def tenant_info():
+    """
+    Get tenant information.
+    ---
+    tags:
+      - Tenant
+    security:
+      - ApiKeyAuth: []
+    responses:
+      200:
+        description: Tenant information retrieved successfully.
+        schema:
+          type: object
+          properties:
+            tenant_id:
+              type: string
+              description: Tenant ID.
+            name:
+              type: string
+              description: Tenant name.
+            llm_id:
+              type: string
+              description: LLM ID.
+            embd_id:
+              type: string
+              description: Embedding model ID.
+    """
     try:
         tenants = TenantService.get_info_by(current_user.id)
         if not tenants:
-            return get_data_error_result(retmsg="Tenant not found!")
+            return get_data_error_result(message="Tenant not found!")
         return get_json_result(data=tenants[0])
     except Exception as e:
         return server_error_response(e)
 
 
-@manager.route("/set_tenant_info", methods=["POST"])
+@manager.route("/set_tenant_info", methods=["POST"])  # noqa: F821
 @login_required
 @validate_request("tenant_id", "asr_id", "embd_id", "img2txt_id", "llm_id")
 def set_tenant_info():
+    """
+    Update tenant information.
+    ---
+    tags:
+      - Tenant
+    security:
+      - ApiKeyAuth: []
+    parameters:
+      - in: body
+        name: body
+        description: Tenant information to update.
+        required: true
+        schema:
+          type: object
+          properties:
+            tenant_id:
+              type: string
+              description: Tenant ID.
+            llm_id:
+              type: string
+              description: LLM ID.
+            embd_id:
+              type: string
+              description: Embedding model ID.
+            asr_id:
+              type: string
+              description: ASR model ID.
+            img2txt_id:
+              type: string
+              description: Image to Text model ID.
+    responses:
+      200:
+        description: Tenant information updated successfully.
+        schema:
+          type: object
+    """
     req = request.json
     try:
-        tid = req["tenant_id"]
-        del req["tenant_id"]
+        tid = req.pop("tenant_id")
         TenantService.update_by_id(tid, req)
         return get_json_result(data=True)
     except Exception as e:

api/constants.py

@@ -13,9 +13,15 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 
-class DataSet:
-    def __init__(self, user_key, dataset_url, uuid, name):
-        self.user_key = user_key
-        self.dataset_url = dataset_url
-        self.uuid = uuid
-        self.name = name
+NAME_LENGTH_LIMIT = 2 ** 10
+
+IMG_BASE64_PREFIX = 'data:image/png;base64,'
+
+SERVICE_CONF = "service_conf.yaml"
+
+API_VERSION = "v1"
+RAG_FLOW_SERVICE_NAME = "ragflow"
+REQUEST_WAIT_SEC = 2
+REQUEST_MAX_WAIT_SEC = 300
+
+DATASET_NAME_LIMIT = 128

api/db/__init__.py

@@ -89,6 +89,7 @@ class ParserType(StrEnum):
     AUDIO = "audio"
     EMAIL = "email"
     KG = "knowledge_graph"
+    TAG = "tag"
 
 
 class FileSource(StrEnum):

api/db/db_models.py

@@ -13,6 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 import inspect
 import os
 import sys
@@ -29,12 +30,10 @@ from peewee import (
     Field, Model, Metadata
 )
 from playhouse.pool import PooledMySQLDatabase, PooledPostgresqlDatabase
-from api.db import SerializedType, ParserType
-from api.settings import DATABASE, stat_logger, SECRET_KEY, DATABASE_TYPE
-from api.utils.log_utils import getLogger
-from api import utils
 
-LOGGER = getLogger()
+from api.db import SerializedType, ParserType
+from api import settings
+from api import utils
 
 
 def singleton(cls, *args, **kw):
@@ -65,7 +64,7 @@ class TextFieldType(Enum):
 
 
 class LongTextField(TextField):
-    field_type = TextFieldType[DATABASE_TYPE.upper()].value
+    field_type = TextFieldType[settings.DATABASE_TYPE.upper()].value
 
 
 class JSONField(LongTextField):
@@ -130,7 +129,7 @@ def is_continuous_field(cls: typing.Type) -> bool:
     for p in cls.__bases__:
         if p in CONTINUOUS_FIELD_TYPE:
             return True
-        elif p != Field and p != object:
+        elif p is not Field and p is not object:
             if is_continuous_field(p):
                 return True
     else:
@@ -272,6 +271,7 @@ class JsonSerializedField(SerializedField):
         super(JsonSerializedField, self).__init__(serialized_type=SerializedType.JSON, object_hook=object_hook,
                                                   object_pairs_hook=object_pairs_hook, **kwargs)
 
+
 class PooledDatabase(Enum):
     MYSQL = PooledMySQLDatabase
     POSTGRES = PooledPostgresqlDatabase
@@ -285,10 +285,11 @@ class DatabaseMigrator(Enum):
 @singleton
 class BaseDataBase:
     def __init__(self):
-        database_config = DATABASE.copy()
+        database_config = settings.DATABASE.copy()
         db_name = database_config.pop("name")
-        self.database_connection = PooledDatabase[DATABASE_TYPE.upper()].value(db_name, **database_config)
-        stat_logger.info('init database on cluster mode successfully')
+        self.database_connection = PooledDatabase[settings.DATABASE_TYPE.upper()].value(db_name, **database_config)
+        logging.info('init database on cluster mode successfully')
+
 
 class PostgresDatabaseLock:
     def __init__(self, lock_name, timeout=10, db=None):
@@ -334,6 +335,7 @@ class PostgresDatabaseLock:
 
         return magic
 
+
 class MysqlDatabaseLock:
     def __init__(self, lock_name, timeout=10, db=None):
         self.lock_name = lock_name
@@ -388,7 +390,7 @@ class DatabaseLock(Enum):
 
 
 DB = BaseDataBase().database_connection
-DB.lock = DatabaseLock[DATABASE_TYPE.upper()].value
+DB.lock = DatabaseLock[settings.DATABASE_TYPE.upper()].value
 
 
 def close_connection():
@@ -396,7 +398,7 @@ def close_connection():
         if DB:
             DB.close_stale(age=30)
     except Exception as e:
-        LOGGER.exception(e)
+        logging.exception(e)
 
 
 class DataBaseModel(BaseModel):
@@ -412,15 +414,15 @@ def init_database_tables(alter_fields=[]):
     for name, obj in members:
         if obj != DataBaseModel and issubclass(obj, DataBaseModel):
             table_objs.append(obj)
-            LOGGER.info(f"start create table {obj.__name__}")
+            logging.debug(f"start create table {obj.__name__}")
             try:
                 obj.create_table()
-                LOGGER.info(f"create table success: {obj.__name__}")
+                logging.debug(f"create table success: {obj.__name__}")
             except Exception as e:
-                LOGGER.exception(e)
+                logging.exception(e)
                 create_failed_list.append(obj.__name__)
     if create_failed_list:
-        LOGGER.info(f"create tables failed: {create_failed_list}")
+        logging.error(f"create tables failed: {create_failed_list}")
         raise Exception(f"create tables failed: {create_failed_list}")
     migrate_db()
 
@@ -470,7 +472,7 @@ class User(DataBaseModel, UserMixin):
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
     is_superuser = BooleanField(null=True, help_text="is root", default=False, index=True)
@@ -479,7 +481,7 @@ class User(DataBaseModel, UserMixin):
         return self.email
 
     def get_id(self):
-        jwt = Serializer(secret_key=SECRET_KEY)
+        jwt = Serializer(secret_key=settings.SECRET_KEY)
         return jwt.dumps(str(self.access_token))
 
     class Meta:
@@ -525,7 +527,7 @@ class Tenant(DataBaseModel):
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
 
@@ -542,7 +544,7 @@ class UserTenant(DataBaseModel):
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
 
@@ -559,7 +561,7 @@ class InvitationCode(DataBaseModel):
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
 
@@ -582,7 +584,7 @@ class LLMFactories(DataBaseModel):
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
 
@@ -616,7 +618,7 @@ class LLM(DataBaseModel):
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
 
@@ -648,7 +650,7 @@ class TenantLLM(DataBaseModel):
         index=True)
     api_key = CharField(max_length=1024, null=True, help_text="API KEY", index=True)
     api_base = CharField(max_length=255, null=True, help_text="API Base")
-
+    max_tokens = IntegerField(default=8192, index=True)
     used_tokens = IntegerField(default=0, index=True)
 
     def __str__(self):
@@ -700,10 +702,11 @@ class Knowledgebase(DataBaseModel):
         default=ParserType.NAIVE.value,
         index=True)
     parser_config = JSONField(null=False, default={"pages": [[1, 1000000]]})
+    pagerank = IntegerField(default=0, index=False)
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
 
@@ -757,6 +760,7 @@ class Document(DataBaseModel):
         default="")
     process_begin_at = DateTimeField(null=True, index=True)
     process_duation = FloatField(default=0)
+    meta_fields = JSONField(null=True, default={})
 
     run = CharField(
         max_length=1,
@@ -767,7 +771,7 @@ class Document(DataBaseModel):
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
 
@@ -840,7 +844,7 @@ class Task(DataBaseModel):
     doc_id = CharField(max_length=32, null=False, index=True)
     from_page = IntegerField(default=0)
 
-    to_page = IntegerField(default=-1)
+    to_page = IntegerField(default=100000000)
 
     begin_at = DateTimeField(null=True, index=True)
     process_duation = FloatField(default=0)
@@ -851,6 +855,8 @@ class Task(DataBaseModel):
         help_text="process message",
         default="")
     retry_count = IntegerField(default=0)
+    digest = TextField(null=True, help_text="task digest", default="")
+    chunk_ids = LongTextField(null=True, help_text="chunk ids", default="")
 
 
 class Dialog(DataBaseModel):
@@ -879,8 +885,10 @@ class Dialog(DataBaseModel):
         default="simple",
         help_text="simple|advanced",
         index=True)
-    prompt_config = JSONField(null=False, default={"system": "", "prologue": "Hi! I'm your assistant, what can I do for you?",
-                                                   "parameters": [], "empty_response": "Sorry! No relevant content was found in the knowledge base!"})
+    prompt_config = JSONField(null=False,
+                              default={"system": "", "prologue": "Hi! I'm your assistant, what can I do for you?",
+                                       "parameters": [],
+                                       "empty_response": "Sorry! No relevant content was found in the knowledge base!"})
 
     similarity_threshold = FloatField(default=0.2)
     vector_similarity_weight = FloatField(default=0.3)
@@ -904,7 +912,7 @@ class Dialog(DataBaseModel):
     status = CharField(
         max_length=1,
         null=True,
-        help_text="is it validate(0: wasted，1: validate)",
+        help_text="is it validate(0: wasted, 1: validate)",
         default="1",
         index=True)
 
@@ -918,6 +926,7 @@ class Conversation(DataBaseModel):
     name = CharField(max_length=255, null=True, help_text="converastion name", index=True)
     message = JSONField(null=True)
     reference = JSONField(null=True, default=[])
+    user_id = CharField(max_length=255, null=True, help_text="user_id", index=True)
 
     class Meta:
         db_table = "conversation"
@@ -928,6 +937,7 @@ class APIToken(DataBaseModel):
     token = CharField(max_length=255, null=False, index=True)
     dialog_id = CharField(max_length=32, null=False, index=True)
     source = CharField(max_length=16, null=True, help_text="none|agent|dialog", index=True)
+    beta = CharField(max_length=255, null=True, index=True)
 
     class Meta:
         db_table = "api_token"
@@ -942,7 +952,7 @@ class API4Conversation(DataBaseModel):
     reference = JSONField(null=True, default=[])
     tokens = IntegerField(default=0)
     source = CharField(max_length=16, null=True, help_text="none|agent|dialog", index=True)
-
+    dsl = JSONField(null=True, default={})
     duration = FloatField(default=0, index=True)
     round = IntegerField(default=0, index=True)
     thumb_up = IntegerField(default=0, index=True)
@@ -980,14 +990,14 @@ class CanvasTemplate(DataBaseModel):
 
 def migrate_db():
     with DB.transaction():
-        migrator = DatabaseMigrator[DATABASE_TYPE.upper()].value(DB)
+        migrator = DatabaseMigrator[settings.DATABASE_TYPE.upper()].value(DB)
         try:
             migrate(
                 migrator.add_column('file', 'source_type', CharField(max_length=128, null=False, default="",
                                                                      help_text="where dose this document come from",
                                                                      index=True))
             )
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
@@ -996,7 +1006,7 @@ def migrate_db():
                                               help_text="default rerank model ID"))
 
             )
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
@@ -1004,59 +1014,104 @@ def migrate_db():
                                                                      help_text="default rerank model ID"))
 
             )
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
                 migrator.add_column('dialog', 'top_k', IntegerField(default=1024))
 
             )
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
                 migrator.alter_column_type('tenant_llm', 'api_key',
                                            CharField(max_length=1024, null=True, help_text="API KEY", index=True))
             )
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
                 migrator.add_column('api_token', 'source',
                                     CharField(max_length=16, null=True, help_text="none|agent|dialog", index=True))
             )
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
                 migrator.add_column("tenant", "tts_id",
                                     CharField(max_length=256, null=True, help_text="default tts model ID", index=True))
             )
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
                 migrator.add_column('api_4_conversation', 'source',
                                     CharField(max_length=16, null=True, help_text="none|agent|dialog", index=True))
             )
-        except Exception as e:
-            pass
-        try:
-            DB.execute_sql('ALTER TABLE llm DROP PRIMARY KEY;')
-            DB.execute_sql('ALTER TABLE llm ADD PRIMARY KEY (llm_name,fid);')
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
                 migrator.add_column('task', 'retry_count', IntegerField(default=0))
             )
-        except Exception as e:
+        except Exception:
             pass
         try:
             migrate(
                 migrator.alter_column_type('api_token', 'dialog_id',
                                            CharField(max_length=32, null=True, index=True))
             )
-        except Exception as e:
+        except Exception:
+            pass
+        try:
+            migrate(
+                migrator.add_column("tenant_llm", "max_tokens", IntegerField(default=8192, index=True))
+            )
+        except Exception:
+            pass
+        try:
+            migrate(
+                migrator.add_column("api_4_conversation", "dsl", JSONField(null=True, default={}))
+            )
+        except Exception:
+            pass
+        try:
+            migrate(
+                migrator.add_column("knowledgebase", "pagerank", IntegerField(default=0, index=False))
+            )
+        except Exception:
+            pass
+        try:
+            migrate(
+                migrator.add_column("api_token", "beta", CharField(max_length=255, null=True, index=True))
+            )
+        except Exception:
+            pass
+        try:
+            migrate(
+                migrator.add_column("task", "digest", TextField(null=True, help_text="task digest", default=""))
+            )
+        except Exception:
             pass
 
+        try:
+            migrate(
+                migrator.add_column("task", "chunk_ids", LongTextField(null=True, help_text="chunk ids", default=""))
+            )
+        except Exception:
+            pass
+        try:
+            migrate(
+                migrator.add_column("conversation", "user_id",
+                                    CharField(max_length=255, null=True, help_text="user_id", index=True))
+            )
+        except Exception:
+            pass
+        try:
+            migrate(
+                migrator.add_column("document", "meta_fields",
+                                    JSONField(null=True, default={}))
+            )
+        except Exception:
+            pass

api/db/db_utils.py

@@ -15,19 +15,12 @@
 #
 import operator
 from functools import reduce
-from typing import Dict, Type, Union
 
 from playhouse.pool import PooledMySQLDatabase
 
 from api.utils import current_timestamp, timestamp_to_date
 
 from api.db.db_models import DB, DataBaseModel
-from api.db.runtime_config import RuntimeConfig
-from api.utils.log_utils import getLogger
-from enum import Enum
-
-
-LOGGER = getLogger()
 
 
 @DB.connection_context()
@@ -93,7 +86,7 @@ supported_operators = {
 
 
 def query_dict2expression(
-        model: Type[DataBaseModel], query: Dict[str, Union[bool, int, str, list, tuple]]):
+        model: type[DataBaseModel], query: dict[str, bool | int | str | list | tuple]):
     expression = []
 
     for field, value in query.items():
@@ -111,8 +104,8 @@ def query_dict2expression(
     return reduce(operator.iand, expression)
 
 
-def query_db(model: Type[DataBaseModel], limit: int = 0, offset: int = 0,
-             query: dict = None, order_by: Union[str, list, tuple] = None):
+def query_db(model: type[DataBaseModel], limit: int = 0, offset: int = 0,
+             query: dict = None, order_by: str | list | tuple | None = None):
     data = model.select()
     if query:
         data = data.where(query_dict2expression(model, query))

api/db/init_data.py

@@ -13,6 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 import base64
 import json
 import os
@@ -28,7 +29,7 @@ from api.db.services.document_service import DocumentService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.llm_service import LLMFactoriesService, LLMService, TenantLLMService, LLMBundle
 from api.db.services.user_service import TenantService, UserTenantService
-from api.settings import CHAT_MDL, EMBEDDING_MDL, ASR_MDL, IMAGE2TEXT_MDL, PARSERS, LLM_FACTORY, API_KEY, LLM_BASE_URL
+from api import settings
 from api.utils.file_utils import get_project_base_directory
 
 
@@ -50,11 +51,11 @@ def init_superuser():
     tenant = {
         "id": user_info["id"],
         "name": user_info["nickname"] + "‘s Kingdom",
-        "llm_id": CHAT_MDL,
-        "embd_id": EMBEDDING_MDL,
-        "asr_id": ASR_MDL,
-        "parser_ids": PARSERS,
-        "img2txt_id": IMAGE2TEXT_MDL
+        "llm_id": settings.CHAT_MDL,
+        "embd_id": settings.EMBEDDING_MDL,
+        "asr_id": settings.ASR_MDL,
+        "parser_ids": settings.PARSERS,
+        "img2txt_id": settings.IMAGE2TEXT_MDL
     }
     usr_tenant = {
         "tenant_id": user_info["id"],
@@ -63,34 +64,33 @@ def init_superuser():
         "role": UserTenantRole.OWNER
     }
     tenant_llm = []
-    for llm in LLMService.query(fid=LLM_FACTORY):
+    for llm in LLMService.query(fid=settings.LLM_FACTORY):
         tenant_llm.append(
-            {"tenant_id": user_info["id"], "llm_factory": LLM_FACTORY, "llm_name": llm.llm_name, "model_type": llm.model_type,
-             "api_key": API_KEY, "api_base": LLM_BASE_URL})
+            {"tenant_id": user_info["id"], "llm_factory": settings.LLM_FACTORY, "llm_name": llm.llm_name,
+             "model_type": llm.model_type,
+             "api_key": settings.API_KEY, "api_base": settings.LLM_BASE_URL})
 
     if not UserService.save(**user_info):
-        print("\033[93m【ERROR】\033[0mcan't init admin.")
+        logging.error("can't init admin.")
         return
     TenantService.insert(**tenant)
     UserTenantService.insert(**usr_tenant)
     TenantLLMService.insert_many(tenant_llm)
-    print(
-        "【INFO】Super user initialized. \033[93memail: admin@ragflow.io, password: admin\033[0m. Changing the password after logining is strongly recomanded.")
+    logging.info(
+        "Super user initialized. email: admin@ragflow.io, password: admin. Changing the password after login is strongly recommended.")
 
     chat_mdl = LLMBundle(tenant["id"], LLMType.CHAT, tenant["llm_id"])
     msg = chat_mdl.chat(system="", history=[
         {"role": "user", "content": "Hello!"}], gen_conf={})
     if msg.find("ERROR: ") == 0:
-        print(
-            "\33[91m【ERROR】\33[0m: ",
+        logging.error(
             "'{}' dosen't work. {}".format(
                 tenant["llm_id"],
                 msg))
     embd_mdl = LLMBundle(tenant["id"], LLMType.EMBEDDING, tenant["embd_id"])
     v, c = embd_mdl.encode(["Hello!"])
     if c == 0:
-        print(
-            "\33[91m【ERROR】\33[0m:",
+        logging.error(
             "'{}' dosen't work!".format(
                 tenant["embd_id"]))
 
@@ -98,7 +98,9 @@ def init_superuser():
 def init_llm_factory():
     try:
         LLMService.filter_delete([(LLM.fid == "MiniMax" or LLM.fid == "Minimax")])
-    except Exception as e:
+        LLMService.filter_delete([(LLM.fid == "cohere")])
+        LLMFactoriesService.filter_delete([LLMFactories.name == "cohere"])
+    except Exception:
         pass
 
     factory_llm_infos = json.load(
@@ -111,14 +113,14 @@ def init_llm_factory():
         llm_infos = factory_llm_info.pop("llm")
         try:
             LLMFactoriesService.save(**factory_llm_info)
-        except Exception as e:
+        except Exception:
             pass
         LLMService.filter_delete([LLM.fid == factory_llm_info["name"]])
         for llm_info in llm_infos:
             llm_info["fid"] = factory_llm_info["name"]
             try:
                 LLMService.save(**llm_info)
-            except Exception as e:
+            except Exception:
                 pass
 
     LLMFactoriesService.filter_delete([LLMFactories.name == "Local"])
@@ -129,8 +131,9 @@ def init_llm_factory():
     LLMFactoriesService.filter_delete([LLMFactoriesService.model.name == "QAnything"])
     LLMService.filter_delete([LLMService.model.fid == "QAnything"])
     TenantLLMService.filter_update([TenantLLMService.model.llm_factory == "QAnything"], {"llm_factory": "Youdao"})
+    TenantLLMService.filter_update([TenantLLMService.model.llm_factory == "cohere"], {"llm_factory": "Cohere"})
     TenantService.filter_update([1 == 1], {
-        "parser_ids": "naive:General,qa:Q&A,resume:Resume,manual:Manual,table:Table,paper:Paper,book:Book,laws:Laws,presentation:Presentation,picture:Picture,one:One,audio:Audio,knowledge_graph:Knowledge Graph,email:Email"})
+        "parser_ids": "naive:General,qa:Q&A,resume:Resume,manual:Manual,table:Table,paper:Paper,book:Book,laws:Laws,presentation:Presentation,picture:Picture,one:One,audio:Audio,email:Email,tag:Tag"})
     ## insert openai two embedding models to the current openai user.
     # print("Start to insert 2 OpenAI embedding models...")
     tenant_ids = set([row["tenant_id"] for row in TenantLLMService.get_openai_models()])
@@ -145,19 +148,12 @@ def init_llm_factory():
                 row = deepcopy(row)
                 row["llm_name"] = "text-embedding-3-large"
                 TenantLLMService.save(**row)
-            except Exception as e:
+            except Exception:
                 pass
             break
     for kb_id in KnowledgebaseService.get_all_ids():
         KnowledgebaseService.update_by_id(kb_id, {"doc_num": DocumentService.get_kb_doc_count(kb_id)})
-    """
-    drop table llm;
-    drop table llm_factories;
-    update tenant set parser_ids='naive:General,qa:Q&A,resume:Resume,manual:Manual,table:Table,paper:Paper,book:Book,laws:Laws,presentation:Presentation,picture:Picture,one:One,audio:Audio,knowledge_graph:Knowledge Graph';
-    alter table knowledgebase modify avatar longtext;
-    alter table user modify avatar longtext;
-    alter table dialog modify icon longtext;
-    """
+
 
 
 def add_graph_templates():
@@ -167,11 +163,10 @@ def add_graph_templates():
             cnvs = json.load(open(os.path.join(dir, fnm), "r"))
             try:
                 CanvasTemplateService.save(**cnvs)
-            except:
+            except Exception:
                 CanvasTemplateService.update_by_id(cnvs["id"], cnvs)
-        except Exception as e:
-            print("Add graph templates error: ", e)
-            print("------------", flush=True)
+        except Exception:
+            logging.exception("Add graph templates error: ")
 
 
 def init_web_data():
@@ -182,7 +177,7 @@ def init_web_data():
     #    init_superuser()
 
     add_graph_templates()
-    print("init web data success:{}".format(time.time() - start_time))
+    logging.info("init web data success:{}".format(time.time() - start_time))
 
 
 if __name__ == '__main__':

api/db/runtime_config.py

@@ -13,7 +13,7 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-from api.versions import get_versions
+from api.versions import get_ragflow_version
 from .reload_config_base import ReloadConfigBase
 
 
@@ -35,7 +35,7 @@ class RuntimeConfig(ReloadConfigBase):
 
     @classmethod
     def init_env(cls):
-        cls.ENV.update(get_versions())
+        cls.ENV.update({"version": get_ragflow_version()})
 
     @classmethod
     def load_config_manager(cls):

api/db/services/__init__.py

@@ -15,13 +15,14 @@
 #
 import pathlib
 import re
-from .user_service import UserService
+from .user_service import UserService as UserService
 
 
 def duplicate_name(query_func, **kwargs):
     fnm = kwargs["name"]
     objs = query_func(**kwargs)
-    if not objs: return fnm
+    if not objs:
+        return fnm
     ext = pathlib.Path(fnm).suffix #.jpg
     nm = re.sub(r"%s$"%ext, "", fnm)
     r = re.search(r"\(([0-9]+)\)$", nm)
@@ -31,8 +32,8 @@ def duplicate_name(query_func, **kwargs):
         nm = re.sub(r"\([0-9]+\)$", "", nm)
     c += 1
     nm = f"{nm}({c})"
-    if ext: nm += f"{ext}"
+    if ext:
+        nm += f"{ext}"
 
     kwargs["name"] = nm
     return duplicate_name(query_func, **kwargs)
-

api/db/services/api_service.py

@@ -39,6 +39,24 @@ class APITokenService(CommonService):
 class API4ConversationService(CommonService):
     model = API4Conversation
 
+    @classmethod
+    @DB.connection_context()
+    def get_list(cls, dialog_id, tenant_id,
+                 page_number, items_per_page,
+                 orderby, desc, id, user_id=None):
+        sessions = cls.model.select().where(cls.model.dialog_id == dialog_id)
+        if id:
+            sessions = sessions.where(cls.model.id == id)
+        if user_id:
+            sessions = sessions.where(cls.model.user_id == user_id)
+        if desc:
+            sessions = sessions.order_by(cls.model.getter_by(orderby).desc())
+        else:
+            sessions = sessions.order_by(cls.model.getter_by(orderby).asc())
+        sessions = sessions.paginate(page_number, items_per_page)
+
+        return list(sessions.dicts())
+
     @classmethod
     @DB.connection_context()
     def append_message(cls, id, conversation):
@@ -48,7 +66,8 @@ class API4ConversationService(CommonService):
     @classmethod
     @DB.connection_context()
     def stats(cls, tenant_id, from_date, to_date, source=None):
-        if len(to_date) == 10: to_date += " 23:59:59"
+        if len(to_date) == 10:
+            to_date += " 23:59:59"
         return cls.model.select(
             cls.model.create_date.truncate("day").alias("dt"),
             peewee.fn.COUNT(

api/db/services/canvas_service.py

@@ -13,14 +13,155 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-from datetime import datetime
-import peewee
-from api.db.db_models import DB, API4Conversation, APIToken, Dialog, CanvasTemplate, UserCanvas
+import json
+import time
+import traceback
+from uuid import uuid4
+from agent.canvas import Canvas
+from api.db.db_models import DB, CanvasTemplate, UserCanvas, API4Conversation
+from api.db.services.api_service import API4ConversationService
 from api.db.services.common_service import CommonService
+from api.db.services.conversation_service import structure_answer
+from api.utils import get_uuid
 
 
 class CanvasTemplateService(CommonService):
     model = CanvasTemplate
 
+
 class UserCanvasService(CommonService):
     model = UserCanvas
+
+    @classmethod
+    @DB.connection_context()
+    def get_list(cls, tenant_id,
+                 page_number, items_per_page, orderby, desc, id, title):
+        agents = cls.model.select()
+        if id:
+            agents = agents.where(cls.model.id == id)
+        if title:
+            agents = agents.where(cls.model.title == title)
+        agents = agents.where(cls.model.user_id == tenant_id)
+        if desc:
+            agents = agents.order_by(cls.model.getter_by(orderby).desc())
+        else:
+            agents = agents.order_by(cls.model.getter_by(orderby).asc())
+
+        agents = agents.paginate(page_number, items_per_page)
+
+        return list(agents.dicts())
+
+
+def completion(tenant_id, agent_id, question, session_id=None, stream=True, **kwargs):
+    e, cvs = UserCanvasService.get_by_id(agent_id)
+    assert e, "Agent not found."
+    assert cvs.user_id == tenant_id, "You do not own the agent."
+    if not isinstance(cvs.dsl,str):
+        cvs.dsl = json.dumps(cvs.dsl, ensure_ascii=False)
+    canvas = Canvas(cvs.dsl, tenant_id)
+    canvas.reset()
+    message_id = str(uuid4())
+    if not session_id:
+        query = canvas.get_preset_param()
+        if query:
+            for ele in query:
+                if not ele["optional"]:
+                    if not kwargs.get(ele["key"]):
+                        assert False, f"`{ele['key']}` is required"
+                    ele["value"] = kwargs[ele["key"]]
+                if ele["optional"]:
+                    if kwargs.get(ele["key"]):
+                        ele["value"] = kwargs[ele['key']]
+                    else:
+                        if "value" in ele:
+                            ele.pop("value")
+        cvs.dsl = json.loads(str(canvas))
+        session_id=get_uuid()
+        conv = {
+            "id": session_id,
+            "dialog_id": cvs.id,
+            "user_id": kwargs.get("user_id", "") if isinstance(kwargs, dict) else "",
+            "message": [{"role": "assistant", "content": canvas.get_prologue(), "created_at": time.time()}],
+            "source": "agent",
+            "dsl": cvs.dsl
+        }
+        API4ConversationService.save(**conv)
+        if query:
+            yield "data:" + json.dumps({"code": 0,
+                                        "message": "",
+                                        "data": {
+                                            "session_id": session_id,
+                                            "answer": canvas.get_prologue(),
+                                            "reference": [],
+                                            "param": canvas.get_preset_param()
+                                        }
+                                        },
+                                       ensure_ascii=False) + "\n\n"
+            yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+            return
+        else:
+            conv = API4Conversation(**conv)
+    else:
+        e, conv = API4ConversationService.get_by_id(session_id)
+        assert e, "Session not found!"
+        canvas = Canvas(json.dumps(conv.dsl), tenant_id)
+        canvas.messages.append({"role": "user", "content": question, "id": message_id})
+        canvas.add_user_input(question)
+        if not conv.message:
+            conv.message = []
+        conv.message.append({
+            "role": "user",
+            "content": question,
+            "id": message_id
+        })
+        if not conv.reference:
+            conv.reference = []
+        conv.reference.append({"chunks": [], "doc_aggs": []})
+
+    final_ans = {"reference": [], "content": ""}
+    if stream:
+        try:
+            for ans in canvas.run(stream=stream):
+                if ans.get("running_status"):
+                    yield "data:" + json.dumps({"code": 0, "message": "",
+                                                "data": {"answer": ans["content"],
+                                                         "running_status": True}},
+                                               ensure_ascii=False) + "\n\n"
+                    continue
+                for k in ans.keys():
+                    final_ans[k] = ans[k]
+                ans = {"answer": ans["content"], "reference": ans.get("reference", [])}
+                ans = structure_answer(conv, ans, message_id, session_id)
+                yield "data:" + json.dumps({"code": 0, "message": "", "data": ans},
+                                           ensure_ascii=False) + "\n\n"
+
+            canvas.messages.append({"role": "assistant", "content": final_ans["content"], "created_at": time.time(), "id": message_id})
+            canvas.history.append(("assistant", final_ans["content"]))
+            if final_ans.get("reference"):
+                canvas.reference.append(final_ans["reference"])
+            conv.dsl = json.loads(str(canvas))
+            API4ConversationService.append_message(conv.id, conv.to_dict())
+        except Exception as e:
+            traceback.print_exc()
+            conv.dsl = json.loads(str(canvas))
+            API4ConversationService.append_message(conv.id, conv.to_dict())
+            yield "data:" + json.dumps({"code": 500, "message": str(e),
+                                        "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
+                                       ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+
+    else:
+        for answer in canvas.run(stream=False):
+            if answer.get("running_status"):
+                continue
+            final_ans["content"] = "\n".join(answer["content"]) if "content" in answer else ""
+            canvas.messages.append({"role": "assistant", "content": final_ans["content"], "id": message_id})
+            if final_ans.get("reference"):
+                canvas.reference.append(final_ans["reference"])
+            conv.dsl = json.loads(str(canvas))
+
+            result = {"answer": final_ans["content"], "reference": final_ans.get("reference", [])}
+            result = structure_answer(conv, result, message_id, session_id)
+            API4ConversationService.append_message(conv.id, conv.to_dict())
+            yield result
+            break

api/db/services/common_service.py

@@ -115,7 +115,7 @@ class CommonService:
         try:
             obj = cls.model.query(id=pid)[0]
             return True, obj
-        except Exception as e:
+        except Exception:
             return False, None
 
     @classmethod

api/db/services/conversation_service.py

@@ -0,0 +1,233 @@
+#
+#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+import time
+from uuid import uuid4
+from api.db import StatusEnum
+from api.db.db_models import Conversation, DB
+from api.db.services.api_service import API4ConversationService
+from api.db.services.common_service import CommonService
+from api.db.services.dialog_service import DialogService, chat
+from api.utils import get_uuid
+import json
+
+
+class ConversationService(CommonService):
+    model = Conversation
+
+    @classmethod
+    @DB.connection_context()
+    def get_list(cls, dialog_id, page_number, items_per_page, orderby, desc, id, name, user_id=None):
+        sessions = cls.model.select().where(cls.model.dialog_id == dialog_id)
+        if id:
+            sessions = sessions.where(cls.model.id == id)
+        if name:
+            sessions = sessions.where(cls.model.name == name)
+        if user_id:
+            sessions = sessions.where(cls.model.user_id == user_id)
+        if desc:
+            sessions = sessions.order_by(cls.model.getter_by(orderby).desc())
+        else:
+            sessions = sessions.order_by(cls.model.getter_by(orderby).asc())
+
+        sessions = sessions.paginate(page_number, items_per_page)
+
+        return list(sessions.dicts())
+
+
+def structure_answer(conv, ans, message_id, session_id):
+    reference = ans["reference"]
+    if not isinstance(reference, dict):
+        reference = {}
+        ans["reference"] = {}
+
+    def get_value(d, k1, k2):
+        return d.get(k1, d.get(k2))
+
+    chunk_list = [{
+        "id": get_value(chunk, "chunk_id", "id"),
+        "content": get_value(chunk, "content", "content_with_weight"),
+        "document_id": get_value(chunk, "doc_id", "document_id"),
+        "document_name": get_value(chunk, "docnm_kwd", "document_name"),
+        "dataset_id": get_value(chunk, "kb_id", "dataset_id"),
+        "image_id": get_value(chunk, "image_id", "img_id"),
+        "positions": get_value(chunk, "positions", "position_int"),
+    } for chunk in reference.get("chunks", [])]
+
+    reference["chunks"] = chunk_list
+    ans["id"] = message_id
+    ans["session_id"] = session_id
+
+    if not conv:
+        return ans
+
+    if not conv.message:
+        conv.message = []
+    if not conv.message or conv.message[-1].get("role", "") != "assistant":
+        conv.message.append({"role": "assistant", "content": ans["answer"], "created_at": time.time(), "id": message_id})
+    else:
+        conv.message[-1] = {"role": "assistant", "content": ans["answer"], "created_at": time.time(), "id": message_id}
+    if conv.reference:
+        conv.reference[-1] = reference
+    return ans
+
+
+def completion(tenant_id, chat_id, question, name="New session", session_id=None, stream=True, **kwargs):
+    assert name, "`name` can not be empty."
+    dia = DialogService.query(id=chat_id, tenant_id=tenant_id, status=StatusEnum.VALID.value)
+    assert dia, "You do not own the chat."
+
+    if not session_id:
+        session_id = get_uuid()
+        conv = {
+            "id": session_id,
+            "dialog_id": chat_id,
+            "name": name,
+            "message": [{"role": "assistant", "content": dia[0].prompt_config.get("prologue"), "created_at": time.time()}],
+            "user_id": kwargs.get("user_id", "")
+        }
+        ConversationService.save(**conv)
+        yield "data:" + json.dumps({"code": 0, "message": "",
+                                    "data": {
+                                        "answer": conv["message"][0]["content"],
+                                        "reference": {},
+                                        "audio_binary": None,
+                                        "id": None,
+                                        "session_id": session_id
+                                    }},
+                                   ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+        return
+
+    conv = ConversationService.query(id=session_id, dialog_id=chat_id)
+    if not conv:
+        raise LookupError("Session does not exist")
+
+    conv = conv[0]
+    msg = []
+    question = {
+        "content": question,
+        "role": "user",
+        "id": str(uuid4())
+    }
+    conv.message.append(question)
+    for m in conv.message:
+        if m["role"] == "system":
+            continue
+        if m["role"] == "assistant" and not msg:
+            continue
+        msg.append(m)
+    message_id = msg[-1].get("id")
+    e, dia = DialogService.get_by_id(conv.dialog_id)
+
+    if not conv.reference:
+        conv.reference = []
+    conv.message.append({"role": "assistant", "content": "", "id": message_id})
+    conv.reference.append({"chunks": [], "doc_aggs": []})
+
+    if stream:
+        try:
+            for ans in chat(dia, msg, True, **kwargs):
+                ans = structure_answer(conv, ans, message_id, session_id)
+                yield "data:" + json.dumps({"code": 0, "data": ans}, ensure_ascii=False) + "\n\n"
+            ConversationService.update_by_id(conv.id, conv.to_dict())
+        except Exception as e:
+            yield "data:" + json.dumps({"code": 500, "message": str(e),
+                                        "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
+                                       ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "data": True}, ensure_ascii=False) + "\n\n"
+
+    else:
+        answer = None
+        for ans in chat(dia, msg, False, **kwargs):
+            answer = structure_answer(conv, ans, message_id, session_id)
+            ConversationService.update_by_id(conv.id, conv.to_dict())
+            break
+        yield answer
+
+
+def iframe_completion(dialog_id, question, session_id=None, stream=True, **kwargs):
+    e, dia = DialogService.get_by_id(dialog_id)
+    assert e, "Dialog not found"
+    if not session_id:
+        session_id = get_uuid()
+        conv = {
+            "id": session_id,
+            "dialog_id": dialog_id,
+            "user_id": kwargs.get("user_id", ""),
+            "message": [{"role": "assistant", "content": dia.prompt_config["prologue"], "created_at": time.time()}]
+        }
+        API4ConversationService.save(**conv)
+        yield "data:" + json.dumps({"code": 0, "message": "",
+                                    "data": {
+                                        "answer": conv["message"][0]["content"],
+                                        "reference": {},
+                                        "audio_binary": None,
+                                        "id": None,
+                                        "session_id": session_id
+                                    }},
+                                   ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+        return
+    else:
+        session_id = session_id
+        e, conv = API4ConversationService.get_by_id(session_id)
+        assert e, "Session not found!"
+
+    if not conv.message:
+        conv.message = []
+    messages = conv.message
+    question = {
+        "role": "user",
+        "content": question,
+        "id": str(uuid4())
+    }
+    messages.append(question)
+
+    msg = []
+    for m in messages:
+        if m["role"] == "system":
+            continue
+        if m["role"] == "assistant" and not msg:
+            continue
+        msg.append(m)
+    if not msg[-1].get("id"):
+        msg[-1]["id"] = get_uuid()
+    message_id = msg[-1]["id"]
+
+    if not conv.reference:
+        conv.reference = []
+    conv.reference.append({"chunks": [], "doc_aggs": []})
+
+    if stream:
+        try:
+            for ans in chat(dia, msg, True, **kwargs):
+                ans = structure_answer(conv, ans, message_id, session_id)
+                yield "data:" + json.dumps({"code": 0, "message": "", "data": ans},
+                                           ensure_ascii=False) + "\n\n"
+            API4ConversationService.append_message(conv.id, conv.to_dict())
+        except Exception as e:
+            yield "data:" + json.dumps({"code": 500, "message": str(e),
+                                        "data": {"answer": "**ERROR**: " + str(e), "reference": []}},
+                                       ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+
+    else:
+        answer = None
+        for ans in chat(dia, msg, False, **kwargs):
+            answer = structure_answer(conv, ans, message_id, session_id)
+            API4ConversationService.append_message(conv.id, conv.to_dict())
+            break
+        yield answer

api/db/services/dialog_service.py

@@ -13,22 +13,28 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
 import binascii
 import os
 import json
+import json_repair
 import re
+from collections import defaultdict
 from copy import deepcopy
 from timeit import default_timer as timer
-
-
+import datetime
+from datetime import timedelta
 from api.db import LLMType, ParserType, StatusEnum
-from api.db.db_models import Dialog, Conversation,DB
+from api.db.db_models import Dialog, DB
 from api.db.services.common_service import CommonService
+from api.db.services.document_service import DocumentService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.llm_service import LLMService, TenantLLMService, LLMBundle
-from api.settings import chat_logger, retrievaler, kg_retrievaler
+from api import settings
+from graphrag.utils import get_tags_from_cache, set_tags_to_cache
 from rag.app.resume import forbidden_select_fields4resume
 from rag.nlp.search import index_name
+from rag.settings import TAG_FLD
 from rag.utils import rmSpace, num_tokens_from_string, encoder
 from api.utils.file_utils import get_project_base_directory
 
@@ -59,27 +65,6 @@ class DialogService(CommonService):
         return list(chats.dicts())
 
 
-class ConversationService(CommonService):
-    model = Conversation
-
-    @classmethod
-    @DB.connection_context()
-    def get_list(cls,dialog_id,page_number, items_per_page, orderby, desc, id , name):
-        sessions = cls.model.select().where(cls.model.dialog_id ==dialog_id)
-        if id:
-            sessions = sessions.where(cls.model.id == id)
-        if name:
-            sessions = sessions.where(cls.model.name == name)
-        if desc:
-            sessions = sessions.order_by(cls.model.getter_by(orderby).desc())
-        else:
-            sessions = sessions.order_by(cls.model.getter_by(orderby).asc())
-
-        sessions = sessions.paginate(page_number, items_per_page)
-
-        return list(sessions.dicts())
-
-
 def message_fit_in(msg, max_length=4000):
     def count():
         nonlocal msg
@@ -97,6 +82,7 @@ def message_fit_in(msg, max_length=4000):
         return c, msg
 
     msg_ = [m for m in msg[:-1] if m["role"] == "system"]
+    if len(msg) > 1:
         msg_.append(msg[-1])
     msg = msg_
     c = count()
@@ -104,21 +90,21 @@ def message_fit_in(msg, max_length=4000):
         return c, msg
 
     ll = num_tokens_from_string(msg_[0]["content"])
-    l = num_tokens_from_string(msg_[-1]["content"])
-    if ll / (ll + l) > 0.8:
+    ll2 = num_tokens_from_string(msg_[-1]["content"])
+    if ll / (ll + ll2) > 0.8:
         m = msg_[0]["content"]
-        m = encoder.decode(encoder.encode(m)[:max_length - l])
+        m = encoder.decode(encoder.encode(m)[:max_length - ll2])
         msg[0]["content"] = m
         return max_length, msg
 
     m = msg_[1]["content"]
-    m = encoder.decode(encoder.encode(m)[:max_length - l])
+    m = encoder.decode(encoder.encode(m)[:max_length - ll2])
     msg[1]["content"] = m
     return max_length, msg
 
 
 def llm_id2llm_type(llm_id):
-    llm_id = llm_id.split("@")[0]
+    llm_id, _ = TenantLLMService.split_model_name_and_factory(llm_id)
     fnm = os.path.join(get_project_base_directory(), "conf")
     llm_factories = json.load(open(os.path.join(fnm, "llm_factories.json"), "r"))
     for llm_factory in llm_factories["factory_llm_infos"]:
@@ -127,31 +113,92 @@ def llm_id2llm_type(llm_id):
                 return llm["model_type"].strip(",")[-1]
 
 
+def kb_prompt(kbinfos, max_tokens):
+    knowledges = [ck["content_with_weight"] for ck in kbinfos["chunks"]]
+    used_token_count = 0
+    chunks_num = 0
+    for i, c in enumerate(knowledges):
+        used_token_count += num_tokens_from_string(c)
+        chunks_num += 1
+        if max_tokens * 0.97 < used_token_count:
+            knowledges = knowledges[:i]
+            break
+
+    docs = DocumentService.get_by_ids([ck["doc_id"] for ck in kbinfos["chunks"][:chunks_num]])
+    docs = {d.id: d.meta_fields for d in docs}
+
+    doc2chunks = defaultdict(lambda: {"chunks": [], "meta": []})
+    for ck in kbinfos["chunks"][:chunks_num]:
+        doc2chunks[ck["docnm_kwd"]]["chunks"].append(ck["content_with_weight"])
+        doc2chunks[ck["docnm_kwd"]]["meta"] = docs.get(ck["doc_id"], {})
+
+    knowledges = []
+    for nm, cks_meta in doc2chunks.items():
+        txt = f"Document: {nm} \n"
+        for k,v in cks_meta["meta"].items():
+            txt += f"{k}: {v}\n"
+        txt += "Relevant fragments as following:\n"
+        for i, chunk in enumerate(cks_meta["chunks"], 1):
+            txt += f"{i}. {chunk}\n"
+        knowledges.append(txt)
+    return knowledges
+
+
+def label_question(question, kbs):
+    tags = None
+    tag_kb_ids = []
+    for kb in kbs:
+        if kb.parser_config.get("tag_kb_ids"):
+            tag_kb_ids.extend(kb.parser_config["tag_kb_ids"])
+    if tag_kb_ids:
+        all_tags = get_tags_from_cache(tag_kb_ids)
+        if not all_tags:
+            all_tags = settings.retrievaler.all_tags_in_portion(kb.tenant_id, tag_kb_ids)
+            set_tags_to_cache(all_tags, tag_kb_ids)
+        else:
+            all_tags = json.loads(all_tags)
+        tag_kbs = KnowledgebaseService.get_by_ids(tag_kb_ids)
+        tags = settings.retrievaler.tag_query(question,
+                                              list(set([kb.tenant_id for kb in tag_kbs])),
+                                              tag_kb_ids,
+                                              all_tags,
+                                              kb.parser_config.get("topn_tags", 3)
+                                              )
+    return tags
+
+
 def chat(dialog, messages, stream=True, **kwargs):
     assert messages[-1]["role"] == "user", "The last content of this conversation is not from user."
-    st = timer()
-    tmp = dialog.llm_id.split("@")
-    fid = None
-    llm_id = tmp[0]
-    if len(tmp)>1: fid = tmp[1]
 
-    llm = LLMService.query(llm_name=llm_id) if not fid else LLMService.query(llm_name=llm_id, fid=fid)
+    chat_start_ts = timer()
+
+    # Get llm model name and model provider name
+    llm_id, model_provider = TenantLLMService.split_model_name_and_factory(dialog.llm_id)
+
+    # Get llm model instance by model and provide name
+    llm = LLMService.query(llm_name=llm_id) if not model_provider else LLMService.query(llm_name=llm_id, fid=model_provider)
+
     if not llm:
-        llm = TenantLLMService.query(tenant_id=dialog.tenant_id, llm_name=llm_id) if not fid else \
-            TenantLLMService.query(tenant_id=dialog.tenant_id, llm_name=llm_id, llm_factory=fid)
+        # Model name is provided by tenant, but not system built-in
+        llm = TenantLLMService.query(tenant_id=dialog.tenant_id, llm_name=llm_id) if not model_provider else \
+            TenantLLMService.query(tenant_id=dialog.tenant_id, llm_name=llm_id, llm_factory=model_provider)
         if not llm:
             raise LookupError("LLM(%s) not found" % dialog.llm_id)
         max_tokens = 8192
     else:
         max_tokens = llm[0].max_tokens
+
+    check_llm_ts = timer()
+
     kbs = KnowledgebaseService.get_by_ids(dialog.kb_ids)
-    embd_nms = list(set([kb.embd_id for kb in kbs]))
-    if len(embd_nms) != 1:
+    embedding_list = list(set([kb.embd_id for kb in kbs]))
+    if len(embedding_list) != 1:
         yield {"answer": "**ERROR**: Knowledge bases use different embedding models.", "reference": []}
         return {"answer": "**ERROR**: Knowledge bases use different embedding models.", "reference": []}
 
-    is_kg = all([kb.parser_id == ParserType.KG for kb in kbs])
-    retr = retrievaler if not is_kg else kg_retrievaler
+    embedding_model_name = embedding_list[0]
+
+    retriever = settings.retrievaler
 
     questions = [m["content"] for m in messages if m["role"] == "user"][-3:]
     attachments = kwargs["doc_ids"].split(",") if "doc_ids" in kwargs else None
@@ -161,12 +208,21 @@ def chat(dialog, messages, stream=True, **kwargs):
             if "doc_ids" in m:
                 attachments.extend(m["doc_ids"])
 
-    embd_mdl = LLMBundle(dialog.tenant_id, LLMType.EMBEDDING, embd_nms[0])
+    create_retriever_ts = timer()
+
+    embd_mdl = LLMBundle(dialog.tenant_id, LLMType.EMBEDDING, embedding_model_name)
+    if not embd_mdl:
+        raise LookupError("Embedding model(%s) not found" % embedding_model_name)
+
+    bind_embedding_ts = timer()
+
     if llm_id2llm_type(dialog.llm_id) == "image2text":
         chat_mdl = LLMBundle(dialog.tenant_id, LLMType.IMAGE2TEXT, dialog.llm_id)
     else:
         chat_mdl = LLMBundle(dialog.tenant_id, LLMType.CHAT, dialog.llm_id)
 
+    bind_llm_ts = timer()
+
     prompt_config = dialog.prompt_config
     field_map = KnowledgebaseService.get_field_map(dialog.kb_ids)
     tts_mdl = None
@@ -174,7 +230,7 @@ def chat(dialog, messages, stream=True, **kwargs):
         tts_mdl = LLMBundle(dialog.tenant_id, LLMType.TTS)
     # try to use sql if field mapping is good to go
     if field_map:
-        chat_logger.info("Use SQL to retrieval:{}".format(questions[-1]))
+        logging.debug("Use SQL to retrieval:{}".format(questions[-1]))
         ans = use_sql(questions[-1], field_map, dialog.tenant_id, chat_mdl, prompt_config.get("quote", True))
         if ans:
             yield ans
@@ -194,35 +250,52 @@ def chat(dialog, messages, stream=True, **kwargs):
     else:
         questions = questions[-1:]
 
+    refine_question_ts = timer()
+
     rerank_mdl = None
     if dialog.rerank_id:
         rerank_mdl = LLMBundle(dialog.tenant_id, LLMType.RERANK, dialog.rerank_id)
 
-    for _ in range(len(questions) // 2):
-        questions.append(questions[-1])
+    bind_reranker_ts = timer()
+    generate_keyword_ts = bind_reranker_ts
+
     if "knowledge" not in [p["key"] for p in prompt_config["parameters"]]:
         kbinfos = {"total": 0, "chunks": [], "doc_aggs": []}
     else:
         if prompt_config.get("keyword", False):
             questions[-1] += keyword_extraction(chat_mdl, questions[-1])
+            generate_keyword_ts = timer()
 
         tenant_ids = list(set([kb.tenant_id for kb in kbs]))
-        kbinfos = retr.retrieval(" ".join(questions), embd_mdl, tenant_ids, dialog.kb_ids, 1, dialog.top_n,
+
+        kbinfos = retriever.retrieval(" ".join(questions), embd_mdl, tenant_ids, dialog.kb_ids, 1, dialog.top_n,
                                       dialog.similarity_threshold,
                                       dialog.vector_similarity_weight,
                                       doc_ids=attachments,
-                                        top=dialog.top_k, aggs=False, rerank_mdl=rerank_mdl)
-    knowledges = [ck["content_with_weight"] for ck in kbinfos["chunks"]]
-    chat_logger.info(
+                                      top=dialog.top_k, aggs=False, rerank_mdl=rerank_mdl,
+                                      rank_feature=label_question(" ".join(questions), kbs)
+                                      )
+        if prompt_config.get("use_kg"):
+            ck = settings.kg_retrievaler.retrieval(" ".join(questions),
+                                              tenant_ids,
+                                              dialog.kb_ids,
+                                              embd_mdl,
+                                              LLMBundle(dialog.tenant_id, LLMType.CHAT))
+            if ck["content_with_weight"]:
+                kbinfos["chunks"].insert(0, ck)
+
+    retrieval_ts = timer()
+
+    knowledges = kb_prompt(kbinfos, max_tokens)
+    logging.debug(
         "{}->{}".format(" ".join(questions), "\n->".join(knowledges)))
-    retrieval_tm = timer()
 
     if not knowledges and prompt_config.get("empty_response"):
         empty_res = prompt_config["empty_response"]
         yield {"answer": empty_res, "reference": kbinfos, "audio_binary": tts(tts_mdl, empty_res)}
         return {"answer": prompt_config["empty_response"], "reference": kbinfos}
 
-    kwargs["knowledge"] = "\n\n------\n\n".join(knowledges)
+    kwargs["knowledge"] = "\n------\n" + "\n\n------\n\n".join(knowledges)
     gen_conf = dialog.llm_setting
 
     msg = [{"role": "system", "content": prompt_config["system"].format(**kwargs)}]
@@ -239,10 +312,13 @@ def chat(dialog, messages, stream=True, **kwargs):
             max_tokens - used_token_count)
 
     def decorate_answer(answer):
-        nonlocal prompt_config, knowledges, kwargs, kbinfos, prompt, retrieval_tm
+        nonlocal prompt_config, knowledges, kwargs, kbinfos, prompt, retrieval_ts
+
+        finish_chat_ts = timer()
+
         refs = []
         if knowledges and (prompt_config.get("quote", True) and kwargs.get("quote", True)):
-            answer, idx = retr.insert_citations(answer,
+            answer, idx = retriever.insert_citations(answer,
                                                      [ck["content_ltks"]
                                                       for ck in kbinfos["chunks"]],
                                                      [ck["vector"]
@@ -253,7 +329,8 @@ def chat(dialog, messages, stream=True, **kwargs):
             idx = set([kbinfos["chunks"][int(i)]["doc_id"] for i in idx])
             recall_docs = [
                 d for d in kbinfos["doc_aggs"] if d["doc_id"] in idx]
-            if not recall_docs: recall_docs = kbinfos["doc_aggs"]
+            if not recall_docs:
+                recall_docs = kbinfos["doc_aggs"]
             kbinfos["doc_aggs"] = recall_docs
 
             refs = deepcopy(kbinfos)
@@ -262,10 +339,22 @@ def chat(dialog, messages, stream=True, **kwargs):
                     del c["vector"]
 
         if answer.lower().find("invalid key") >= 0 or answer.lower().find("invalid api") >= 0:
-            answer += " Please set LLM API-Key in 'User Setting -> Model Providers -> API-Key'"
-        done_tm = timer()
-        prompt += "\n\n### Elapsed\n  - Retrieval: %.1f ms\n  - LLM: %.1f ms"%((retrieval_tm-st)*1000, (done_tm-st)*1000)
-        return {"answer": answer, "reference": refs, "prompt": prompt}
+            answer += " Please set LLM API-Key in 'User Setting -> Model providers -> API-Key'"
+        finish_chat_ts = timer()
+
+        total_time_cost = (finish_chat_ts - chat_start_ts) * 1000
+        check_llm_time_cost = (check_llm_ts - chat_start_ts) * 1000
+        create_retriever_time_cost = (create_retriever_ts - check_llm_ts) * 1000
+        bind_embedding_time_cost = (bind_embedding_ts - create_retriever_ts) * 1000
+        bind_llm_time_cost = (bind_llm_ts - bind_embedding_ts) * 1000
+        refine_question_time_cost = (refine_question_ts - bind_llm_ts) * 1000
+        bind_reranker_time_cost = (bind_reranker_ts - refine_question_ts) * 1000
+        generate_keyword_time_cost = (generate_keyword_ts - bind_reranker_ts) * 1000
+        retrieval_time_cost = (retrieval_ts - generate_keyword_ts) * 1000
+        generate_result_time_cost = (finish_chat_ts - retrieval_ts) * 1000
+
+        prompt = f"{prompt}\n\n - Total: {total_time_cost:.1f}ms\n  - Check LLM: {check_llm_time_cost:.1f}ms\n  - Create retriever: {create_retriever_time_cost:.1f}ms\n  - Bind embedding: {bind_embedding_time_cost:.1f}ms\n  - Bind LLM: {bind_llm_time_cost:.1f}ms\n  - Tune question: {refine_question_time_cost:.1f}ms\n  - Bind reranker: {bind_reranker_time_cost:.1f}ms\n  - Generate keyword: {generate_keyword_time_cost:.1f}ms\n  - Retrieval: {retrieval_time_cost:.1f}ms\n  - Generate answer: {generate_result_time_cost:.1f}ms"
+        return {"answer": answer, "reference": refs, "prompt": re.sub(r"\n", "  \n", prompt)}
 
     if stream:
         last_ans = ""
@@ -283,7 +372,7 @@ def chat(dialog, messages, stream=True, **kwargs):
         yield decorate_answer(answer)
     else:
         answer = chat_mdl.chat(prompt, msg[1:], gen_conf)
-        chat_logger.info("User: {}|Assistant: {}".format(
+        logging.debug("User: {}|Assistant: {}".format(
             msg[-1]["content"], answer))
         res = decorate_answer(answer)
         res["audio_binary"] = tts(tts_mdl, answer)
@@ -291,15 +380,15 @@ def chat(dialog, messages, stream=True, **kwargs):
 
 
 def use_sql(question, field_map, tenant_id, chat_mdl, quota=True):
-    sys_prompt = "你是一个DBA。你需要这对以下表的字段结构，根据用户的问题列表，写出最后一个问题对应的SQL。"
-    user_promt = """
-表名：{}；
-数据库表字段说明如下：
+    sys_prompt = "You are a Database Administrator. You need to check the fields of the following tables based on the user's list of questions and write the SQL corresponding to the last question."
+    user_prompt = """
+Table name: {};
+Table of database fields are as follows:
 {}
 
-问题如下：
+Question are as follows:
 {}
-请写出SQL, 且只要SQL，不要有其他说明及文字。
+Please write the SQL, only SQL, without any other explanations or text.
 """.format(
         index_name(tenant_id),
         "\n".join([f"{k}: {v}" for k, v in field_map.items()]),
@@ -308,11 +397,10 @@ def use_sql(question, field_map, tenant_id, chat_mdl, quota=True):
     tried_times = 0
 
     def get_table():
-        nonlocal sys_prompt, user_promt, question, tried_times
-        sql = chat_mdl.chat(sys_prompt, [{"role": "user", "content": user_promt}], {
+        nonlocal sys_prompt, user_prompt, question, tried_times
+        sql = chat_mdl.chat(sys_prompt, [{"role": "user", "content": user_prompt}], {
             "temperature": 0.06})
-        print(user_promt, sql)
-        chat_logger.info(f"“{question}”==>{user_promt} get SQL: {sql}")
+        logging.debug(f"{question} ==> {user_prompt} get SQL: {sql}")
         sql = re.sub(r"[\r\n]+", " ", sql.lower())
         sql = re.sub(r".*select ", "select ", sql.lower())
         sql = re.sub(r" +", " ", sql)
@@ -332,86 +420,86 @@ def use_sql(question, field_map, tenant_id, chat_mdl, quota=True):
                     flds.append(k)
                 sql = "select doc_id,docnm_kwd," + ",".join(flds) + sql[8:]
 
-        print(f"“{question}” get SQL(refined): {sql}")
-
-        chat_logger.info(f"“{question}” get SQL(refined): {sql}")
+        logging.debug(f"{question} get SQL(refined): {sql}")
         tried_times += 1
-        return retrievaler.sql_retrieval(sql, format="json"), sql
+        return settings.retrievaler.sql_retrieval(sql, format="json"), sql
 
     tbl, sql = get_table()
     if tbl is None:
         return None
     if tbl.get("error") and tried_times <= 2:
-        user_promt = """
-        表名：{}；
-        数据库表字段说明如下：
+        user_prompt = """
+        Table name: {};
+        Table of database fields are as follows:
         {}
         
-        问题如下：
+        Question are as follows:
+        {}
+        Please write the SQL, only SQL, without any other explanations or text.
+        
+
+        The SQL error you provided last time is as follows:
         {}
 
-        你上一次给出的错误SQL如下：
+        Error issued by database as follows:
         {}
 
-        后台报错如下：
-        {}
-
-        请纠正SQL中的错误再写一遍，且只要SQL，不要有其他说明及文字。
+        Please correct the error and write SQL again, only SQL, without any other explanations or text.
         """.format(
             index_name(tenant_id),
             "\n".join([f"{k}: {v}" for k, v in field_map.items()]),
             question, sql, tbl["error"]
         )
         tbl, sql = get_table()
-        chat_logger.info("TRY it again: {}".format(sql))
+        logging.debug("TRY it again: {}".format(sql))
 
-    chat_logger.info("GET table: {}".format(tbl))
-    print(tbl)
+    logging.debug("GET table: {}".format(tbl))
     if tbl.get("error") or len(tbl["rows"]) == 0:
         return None
 
     docid_idx = set([ii for ii, c in enumerate(
         tbl["columns"]) if c["name"] == "doc_id"])
-    docnm_idx = set([ii for ii, c in enumerate(
+    doc_name_idx = set([ii for ii, c in enumerate(
         tbl["columns"]) if c["name"] == "docnm_kwd"])
-    clmn_idx = [ii for ii in range(
-        len(tbl["columns"])) if ii not in (docid_idx | docnm_idx)]
+    column_idx = [ii for ii in range(
+        len(tbl["columns"])) if ii not in (docid_idx | doc_name_idx)]
 
-    # compose markdown table
-    clmns = "|" + "|".join([re.sub(r"(/.*|（[^（）]+）)", "", field_map.get(tbl["columns"][i]["name"],
+    # compose Markdown table
+    columns = "|" + "|".join([re.sub(r"(/.*|（[^（）]+）)", "", field_map.get(tbl["columns"][i]["name"],
                                                                           tbl["columns"][i]["name"])) for i in
-                            clmn_idx]) + ("|Source|" if docid_idx and docid_idx else "|")
+                              column_idx]) + ("|Source|" if docid_idx and docid_idx else "|")
 
-    line = "|" + "|".join(["------" for _ in range(len(clmn_idx))]) + \
+    line = "|" + "|".join(["------" for _ in range(len(column_idx))]) + \
            ("|------|" if docid_idx and docid_idx else "")
 
     rows = ["|" +
-            "|".join([rmSpace(str(r[i])) for i in clmn_idx]).replace("None", " ") +
+            "|".join([rmSpace(str(r[i])) for i in column_idx]).replace("None", " ") +
             "|" for r in tbl["rows"]]
+    rows = [r for r in rows if re.sub(r"[ |]+", "", r)]
     if quota:
         rows = "\n".join([r + f" ##{ii}$$ |" for ii, r in enumerate(rows)])
     else:
         rows = "\n".join([r + f" ##{ii}$$ |" for ii, r in enumerate(rows)])
     rows = re.sub(r"T[0-9]{2}:[0-9]{2}:[0-9]{2}(\.[0-9]+Z)?\|", "|", rows)
 
-    if not docid_idx or not docnm_idx:
-        chat_logger.warning("SQL missing field: " + sql)
+    if not docid_idx or not doc_name_idx:
+        logging.warning("SQL missing field: " + sql)
         return {
-            "answer": "\n".join([clmns, line, rows]),
+            "answer": "\n".join([columns, line, rows]),
             "reference": {"chunks": [], "doc_aggs": []},
             "prompt": sys_prompt
         }
 
     docid_idx = list(docid_idx)[0]
-    docnm_idx = list(docnm_idx)[0]
+    doc_name_idx = list(doc_name_idx)[0]
     doc_aggs = {}
     for r in tbl["rows"]:
         if r[docid_idx] not in doc_aggs:
-            doc_aggs[r[docid_idx]] = {"doc_name": r[docnm_idx], "count": 0}
+            doc_aggs[r[docid_idx]] = {"doc_name": r[doc_name_idx], "count": 0}
         doc_aggs[r[docid_idx]]["count"] += 1
     return {
-        "answer": "\n".join([clmns, line, rows]),
-        "reference": {"chunks": [{"doc_id": r[docid_idx], "docnm_kwd": r[docnm_idx]} for r in tbl["rows"]],
+        "answer": "\n".join([columns, line, rows]),
+        "reference": {"chunks": [{"doc_id": r[docid_idx], "docnm_kwd": r[doc_name_idx]} for r in tbl["rows"]],
                       "doc_aggs": [{"doc_id": did, "doc_name": d["doc_name"], "count": d["count"]} for did, d in
                                    doc_aggs.items()]},
         "prompt": sys_prompt
@@ -430,13 +518,15 @@ def relevant(tenant_id, llm_id, question, contents: list):
         Give a binary score 'yes' or 'no' score to indicate whether the document is relevant to the question.
         No other words needed except 'yes' or 'no'.
     """
-    if not contents:return False
+    if not contents:
+        return False
     contents = "Documents: \n" + "   - ".join(contents)
     contents = f"Question: {question}\n" + contents
     if num_tokens_from_string(contents) >= chat_mdl.max_length - 4:
         contents = encoder.decode(encoder.encode(contents)[:chat_mdl.max_length - 4])
     ans = chat_mdl.chat(prompt, [{"role": "user", "content": contents}], {"temperature": 0.01})
-    if ans.lower().find("yes") >= 0: return True
+    if ans.lower().find("yes") >= 0:
+        return True
     return False
 
 
@@ -478,8 +568,10 @@ Requirements:
     ]
     _, msg = message_fit_in(msg, chat_mdl.max_length)
     kwd = chat_mdl.chat(prompt, msg[1:], {"temperature": 0.2})
-    if isinstance(kwd, tuple): kwd = kwd[0]
-    if kwd.find("**ERROR**") >=0: return ""
+    if isinstance(kwd, tuple):
+        kwd = kwd[0]
+    if kwd.find("**ERROR**") >= 0:
+        return ""
     return kwd
 
 
@@ -505,8 +597,10 @@ Requirements:
     ]
     _, msg = message_fit_in(msg, chat_mdl.max_length)
     kwd = chat_mdl.chat(prompt, msg[1:], {"temperature": 0.2})
-    if isinstance(kwd, tuple): kwd = kwd[0]
-    if kwd.find("**ERROR**") >= 0: return ""
+    if isinstance(kwd, tuple):
+        kwd = kwd[0]
+    if kwd.find("**ERROR**") >= 0:
+        return ""
     return kwd
 
 
@@ -517,12 +611,20 @@ def full_question(tenant_id, llm_id, messages):
         chat_mdl = LLMBundle(tenant_id, LLMType.CHAT, llm_id)
     conv = []
     for m in messages:
-        if m["role"] not in ["user", "assistant"]: continue
+        if m["role"] not in ["user", "assistant"]:
+            continue
         conv.append("{}: {}".format(m["role"].upper(), m["content"]))
     conv = "\n".join(conv)
+    today = datetime.date.today().isoformat()
+    yesterday = (datetime.date.today() - timedelta(days=1)).isoformat()
+    tomorrow = (datetime.date.today() + timedelta(days=1)).isoformat()
     prompt = f"""
 Role: A helpful assistant
-Task: Generate a full user question that would follow the conversation.
+
+Task and steps: 
+    1. Generate a full user question that would follow the conversation.
+    2. If the user's question involves relative date, you need to convert it into absolute date based on the current date, which is {today}. For example: 'yesterday' would be converted to {yesterday}.
+    
 Requirements & Restrictions:
   - Text generated MUST be in the same language of the original user's question.
   - If the user's latest question is completely, don't do anything, just return the original question.
@@ -551,6 +653,14 @@ User: What's her full name?
 ###############
 Output: What's the full name of Donald Trump's mother Mary Trump?
 
+------------
+# Example 3
+## Conversation
+USER: What's the weather today in London?
+ASSISTANT:  Cloudy.
+USER: What's about tomorrow in Rochester?
+###############
+Output: What's the weather in Rochester on {tomorrow}?
 ######################
 
 # Real Data
@@ -563,7 +673,8 @@ Output: What's the full name of Donald Trump's mother Mary Trump?
 
 
 def tts(tts_mdl, text):
-    if not tts_mdl or not text: return
+    if not tts_mdl or not text:
+        return
     bin = b""
     for chunk in tts_mdl.tts(text):
         bin += chunk
@@ -572,25 +683,20 @@ def tts(tts_mdl, text):
 
 def ask(question, kb_ids, tenant_id):
     kbs = KnowledgebaseService.get_by_ids(kb_ids)
-    embd_nms = list(set([kb.embd_id for kb in kbs]))
+    embedding_list = list(set([kb.embd_id for kb in kbs]))
 
-    is_kg = all([kb.parser_id == ParserType.KG for kb in kbs])
-    retr = retrievaler if not is_kg else kg_retrievaler
+    is_knowledge_graph = all([kb.parser_id == ParserType.KG for kb in kbs])
+    retriever = settings.retrievaler if not is_knowledge_graph else settings.kg_retrievaler
 
-    embd_mdl = LLMBundle(tenant_id, LLMType.EMBEDDING, embd_nms[0])
+    embd_mdl = LLMBundle(tenant_id, LLMType.EMBEDDING, embedding_list[0])
     chat_mdl = LLMBundle(tenant_id, LLMType.CHAT)
     max_tokens = chat_mdl.max_length
-
-    kbinfos = retr.retrieval(question, embd_mdl, tenant_id, kb_ids, 1, 12, 0.1, 0.3, aggs=False)
-    knowledges = [ck["content_with_weight"] for ck in kbinfos["chunks"]]
-
-    used_token_count = 0
-    for i, c in enumerate(knowledges):
-        used_token_count += num_tokens_from_string(c)
-        if max_tokens * 0.97 < used_token_count:
-            knowledges = knowledges[:i]
-            break
-
+    tenant_ids = list(set([kb.tenant_id for kb in kbs]))
+    kbinfos = retriever.retrieval(question, embd_mdl, tenant_ids, kb_ids,
+                                  1, 12, 0.1, 0.3, aggs=False,
+                                  rank_feature=label_question(question, kbs)
+                                  )
+    knowledges = kb_prompt(kbinfos, max_tokens)
     prompt = """
     Role: You're a smart assistant. Your name is Miss R.
     Task: Summarize the information from knowledge bases and answer user's question.
@@ -611,7 +717,7 @@ def ask(question, kb_ids, tenant_id):
 
     def decorate_answer(answer):
         nonlocal knowledges, kbinfos, prompt
-        answer, idx = retr.insert_citations(answer,
+        answer, idx = retriever.insert_citations(answer,
                                                  [ck["content_ltks"]
                                                   for ck in kbinfos["chunks"]],
                                                  [ck["vector"]
@@ -622,7 +728,8 @@ def ask(question, kb_ids, tenant_id):
         idx = set([kbinfos["chunks"][int(i)]["doc_id"] for i in idx])
         recall_docs = [
             d for d in kbinfos["doc_aggs"] if d["doc_id"] in idx]
-        if not recall_docs: recall_docs = kbinfos["doc_aggs"]
+        if not recall_docs:
+            recall_docs = kbinfos["doc_aggs"]
         kbinfos["doc_aggs"] = recall_docs
         refs = deepcopy(kbinfos)
         for c in refs["chunks"]:
@@ -639,3 +746,63 @@ def ask(question, kb_ids, tenant_id):
         yield {"answer": answer, "reference": {}}
     yield decorate_answer(answer)
 
+
+def content_tagging(chat_mdl, content, all_tags, examples, topn=3):
+    prompt = f"""
+Role: You're a text analyzer. 
+
+Task: Tag (put on some labels) to a given piece of text content based on the examples and the entire tag set.
+
+Steps:: 
+  - Comprehend the tag/label set.
+  - Comprehend examples which all consist of both text content and assigned tags with relevance score in format of JSON.
+  - Summarize the text content, and tag it with top {topn} most relevant tags from the set of tag/label and the corresponding relevance score.
+
+Requirements
+  - The tags MUST be from the tag set.
+  - The output MUST be in JSON format only, the key is tag and the value is its relevance score.
+  - The relevance score must be range from 1 to 10.
+  - Keywords ONLY in output.
+
+# TAG SET
+{", ".join(all_tags)}
+
+"""
+    for i, ex in enumerate(examples):
+        prompt += """
+# Examples {}
+### Text Content
+{}
+
+Output:
+{}
+
+        """.format(i, ex["content"], json.dumps(ex[TAG_FLD], indent=2, ensure_ascii=False))
+
+    prompt += f"""
+# Real Data
+### Text Content
+{content}
+
+"""
+    msg = [
+        {"role": "system", "content": prompt},
+        {"role": "user", "content": "Output: "}
+    ]
+    _, msg = message_fit_in(msg, chat_mdl.max_length)
+    kwd = chat_mdl.chat(prompt, msg[1:], {"temperature": 0.5})
+    if isinstance(kwd, tuple):
+        kwd = kwd[0]
+    if kwd.find("**ERROR**") >= 0:
+        raise Exception(kwd)
+
+    try:
+        return json_repair.loads(kwd)
+    except json_repair.JSONDecodeError:
+        try:
+            result = kwd.replace(prompt[:-1], '').replace('user', '').replace('model', '').strip()
+            result = '{' + result.split('{')[1].split('}')[0] + '}'
+            return json_repair.loads(result)
+        except Exception as e:
+            logging.exception(f"JSON parsing error: {result} -> {e}")
+            raise e

api/db/services/document_service.py

@@ -13,27 +13,23 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
-import hashlib
+import logging
+import xxhash
 import json
-import os
 import random
 import re
-import traceback
 from concurrent.futures import ThreadPoolExecutor
 from copy import deepcopy
 from datetime import datetime
 from io import BytesIO
 
-from elasticsearch_dsl import Q
 from peewee import fn
 
 from api.db.db_utils import bulk_insert_into_db
-from api.settings import stat_logger
+from api import settings
 from api.utils import current_timestamp, get_format_time, get_uuid
-from api.utils.file_utils import get_project_base_directory
-from graphrag.mind_map_extractor import MindMapExtractor
+from graphrag.general.mind_map_extractor import MindMapExtractor
 from rag.settings import SVR_QUEUE_NAME
-from rag.utils.es_conn import ELASTICSEARCH
 from rag.utils.storage_factory import STORAGE_IMPL
 from rag.nlp import search, rag_tokenizer
 
@@ -52,11 +48,15 @@ class DocumentService(CommonService):
     @classmethod
     @DB.connection_context()
     def get_list(cls, kb_id, page_number, items_per_page,
-                     orderby, desc, keywords, id):
+                 orderby, desc, keywords, id, name):
         docs = cls.model.select().where(cls.model.kb_id == kb_id)
         if id:
             docs = docs.where(
                 cls.model.id == id)
+        if name:
+            docs = docs.where(
+                cls.model.name == name
+            )
         if keywords:
             docs = docs.where(
                 fn.LOWER(cls.model.name).contains(keywords.lower())
@@ -66,11 +66,10 @@ class DocumentService(CommonService):
         else:
             docs = docs.order_by(cls.model.getter_by(orderby).asc())
 
-        docs = docs.paginate(page_number, items_per_page)
         count = docs.count()
+        docs = docs.paginate(page_number, items_per_page)
         return list(docs.dicts()), count
 
-
     @classmethod
     @DB.connection_context()
     def get_by_kb_id(cls, kb_id, page_number, items_per_page,
@@ -92,55 +91,33 @@ class DocumentService(CommonService):
 
         return list(docs.dicts()), count
 
-    @classmethod
-    @DB.connection_context()
-    def list_documents_in_dataset(cls, dataset_id, offset, count, order_by, descend, keywords):
-        if keywords:
-            docs = cls.model.select().where(
-                (cls.model.kb_id == dataset_id),
-                (fn.LOWER(cls.model.name).contains(keywords.lower()))
-            )
-        else:
-            docs = cls.model.select().where(cls.model.kb_id == dataset_id)
-
-        total = docs.count()
-
-        if descend == 'True':
-            docs = docs.order_by(cls.model.getter_by(order_by).desc())
-        if descend == 'False':
-            docs = docs.order_by(cls.model.getter_by(order_by).asc())
-
-        docs = list(docs.dicts())
-        docs_length = len(docs)
-
-        if offset < 0 or offset > docs_length:
-            raise IndexError("Offset is out of the valid range.")
-
-        if count == -1:
-            return docs[offset:], total
-
-        return docs[offset:offset + count], total
-
     @classmethod
     @DB.connection_context()
     def insert(cls, doc):
         if not cls.save(**doc):
             raise RuntimeError("Database error (Document)!")
-        e, doc = cls.get_by_id(doc["id"])
-        if not e:
-            raise RuntimeError("Database error (Document retrieval)!")
-        e, kb = KnowledgebaseService.get_by_id(doc.kb_id)
+        e, kb = KnowledgebaseService.get_by_id(doc["kb_id"])
         if not KnowledgebaseService.update_by_id(
                 kb.id, {"doc_num": kb.doc_num + 1}):
             raise RuntimeError("Database error (Knowledgebase)!")
-        return doc
+        return Document(**doc)
 
     @classmethod
     @DB.connection_context()
     def remove_document(cls, doc, tenant_id):
-        ELASTICSEARCH.deleteByQuery(
-                Q("match", doc_id=doc.id), idxnm=search.index_name(tenant_id))
         cls.clear_chunk_num(doc.id)
+        try:
+            settings.docStoreConn.delete({"doc_id": doc.id}, search.index_name(tenant_id), doc.kb_id)
+            settings.docStoreConn.update({"kb_id": doc.kb_id, "knowledge_graph_kwd": ["entity", "relation", "graph", "community_report"], "source_id": doc.id},
+                                         {"remove": {"source_id": doc.id}},
+                                         search.index_name(tenant_id), doc.kb_id)
+            settings.docStoreConn.update({"kb_id": doc.kb_id, "knowledge_graph_kwd": ["graph"]},
+                                         {"removed_kwd": "Y"},
+                                         search.index_name(tenant_id), doc.kb_id)
+            settings.docStoreConn.delete({"kb_id": doc.kb_id, "knowledge_graph_kwd": ["entity", "relation", "graph", "community_report"], "must_not": {"exists": "source_id"}},
+                                         search.index_name(tenant_id), doc.kb_id)
+        except Exception:
+            pass
         return cls.delete_by_id(doc.id)
 
     @classmethod
@@ -175,7 +152,8 @@ class DocumentService(CommonService):
     @classmethod
     @DB.connection_context()
     def get_unfinished_docs(cls):
-        fields = [cls.model.id, cls.model.process_begin_at, cls.model.parser_config, cls.model.progress_msg, cls.model.run]
+        fields = [cls.model.id, cls.model.process_begin_at, cls.model.parser_config, cls.model.progress_msg,
+                  cls.model.run, cls.model.parser_id]
         docs = cls.model.select(*fields) \
             .where(
             cls.model.status == StatusEnum.VALID.value,
@@ -250,6 +228,15 @@ class DocumentService(CommonService):
             return
         return docs[0]["tenant_id"]
 
+    @classmethod
+    @DB.connection_context()
+    def get_knowledgebase_id(cls, doc_id):
+        docs = cls.model.select(cls.model.kb_id).where(cls.model.id == doc_id)
+        docs = docs.dicts()
+        if not docs:
+            return
+        return docs[0]["kb_id"]
+
     @classmethod
     @DB.connection_context()
     def get_tenant_id_by_name(cls, name):
@@ -303,6 +290,31 @@ class DocumentService(CommonService):
             return
         return docs[0]["embd_id"]
 
+    @classmethod
+    @DB.connection_context()
+    def get_chunking_config(cls, doc_id):
+        configs = (
+            cls.model.select(
+                cls.model.id,
+                cls.model.kb_id,
+                cls.model.parser_id,
+                cls.model.parser_config,
+                Knowledgebase.language,
+                Knowledgebase.embd_id,
+                Tenant.id.alias("tenant_id"),
+                Tenant.img2txt_id,
+                Tenant.asr_id,
+                Tenant.llm_id,
+            )
+                .join(Knowledgebase, on=(cls.model.kb_id == Knowledgebase.id))
+                .join(Tenant, on=(Knowledgebase.tenant_id == Tenant.id))
+                .where(cls.model.id == doc_id)
+        )
+        configs = configs.dicts()
+        if not configs:
+            return None
+        return configs[0]
+
     @classmethod
     @DB.connection_context()
     def get_doc_id_by_doc_name(cls, doc_name):
@@ -338,7 +350,10 @@ class DocumentService(CommonService):
                     dfs_update(old[k], v)
                 else:
                     old[k] = v
+
         dfs_update(d.parser_config, config)
+        if not config.get("raptor") and d.parser_config.get("raptor"):
+            del d.parser_config["raptor"]
         cls.update_by_id(id, {"parser_config": d.parser_config})
 
     @classmethod
@@ -354,13 +369,19 @@ class DocumentService(CommonService):
     def begin2parse(cls, docid):
         cls.update_by_id(
             docid, {"progress": random.random() * 1 / 100.,
-                    "progress_msg": "Task dispatched...",
+                    "progress_msg": "Task is queued...",
                     "process_begin_at": get_format_time()
                     })
 
     @classmethod
     @DB.connection_context()
     def update_progress(cls):
+        MSG = {
+            "raptor": "Start RAPTOR (Recursive Abstractive Processing for Tree-Organized Retrieval).",
+            "graphrag": "Start Graph Extraction",
+            "graph_resolution": "Start Graph Resolution",
+            "graph_community": "Start Graph Community Reports Generation"
+        }
         docs = cls.get_unfinished_docs()
         for d in docs:
             try:
@@ -386,14 +407,27 @@ class DocumentService(CommonService):
                     prg = -1
                     status = TaskStatus.FAIL.value
                 elif finished:
-                    if d["parser_config"].get("raptor", {}).get("use_raptor") and d["progress_msg"].lower().find(" raptor")<0:
-                        queue_raptor_tasks(d)
+                    m = "\n".join(sorted(msg))
+                    if d["parser_config"].get("raptor", {}).get("use_raptor") and m.find(MSG["raptor"]) < 0:
+                        queue_raptor_o_graphrag_tasks(d, "raptor", MSG["raptor"])
+                        prg = 0.98 * len(tsks) / (len(tsks) + 1)
+                    elif d["parser_config"].get("graphrag", {}).get("use_graphrag") and m.find(MSG["graphrag"]) < 0:
+                        queue_raptor_o_graphrag_tasks(d, "graphrag", MSG["graphrag"])
+                        prg = 0.98 * len(tsks) / (len(tsks) + 1)
+                    elif d["parser_config"].get("graphrag", {}).get("use_graphrag") \
+                        and d["parser_config"].get("graphrag", {}).get("resolution") \
+                        and m.find(MSG["graph_resolution"]) < 0:
+                        queue_raptor_o_graphrag_tasks(d, "graph_resolution", MSG["graph_resolution"])
+                        prg = 0.98 * len(tsks) / (len(tsks) + 1)
+                    elif d["parser_config"].get("graphrag", {}).get("use_graphrag") \
+                        and d["parser_config"].get("graphrag", {}).get("community") \
+                        and m.find(MSG["graph_community"]) < 0:
+                        queue_raptor_o_graphrag_tasks(d, "graph_community", MSG["graph_community"])
                         prg = 0.98 * len(tsks) / (len(tsks) + 1)
-                        msg.append("------ RAPTOR -------")
                     else:
                         status = TaskStatus.DONE.value
 
-                msg = "\n".join(msg)
+                msg = "\n".join(sorted(msg))
                 info = {
                     "process_duation": datetime.timestamp(
                         datetime.now()) -
@@ -406,7 +440,7 @@ class DocumentService(CommonService):
                 cls.update_by_id(d["id"], info)
             except Exception as e:
                 if str(e).find("'0'") < 0:
-                    stat_logger.error("fetch task exception:" + str(e))
+                    logging.exception("fetch task exception")
 
     @classmethod
     @DB.connection_context()
@@ -414,42 +448,51 @@ class DocumentService(CommonService):
         return len(cls.model.select(cls.model.id).where(
             cls.model.kb_id == kb_id).dicts())
 
-
     @classmethod
     @DB.connection_context()
     def do_cancel(cls, doc_id):
         try:
             _, doc = DocumentService.get_by_id(doc_id)
             return doc.run == TaskStatus.CANCEL.value or doc.progress < 0
-        except Exception as e:
+        except Exception:
             pass
         return False
 
 
-def queue_raptor_tasks(doc):
+def queue_raptor_o_graphrag_tasks(doc, ty, msg):
+    chunking_config = DocumentService.get_chunking_config(doc["id"])
+    hasher = xxhash.xxh64()
+    for field in sorted(chunking_config.keys()):
+        hasher.update(str(chunking_config[field]).encode("utf-8"))
+
     def new_task():
         nonlocal doc
         return {
             "id": get_uuid(),
             "doc_id": doc["id"],
-            "from_page": 0,
-            "to_page": -1,
-            "progress_msg": "Start to do RAPTOR (Recursive Abstractive Processing For Tree-Organized Retrieval)."
+            "from_page": 100000000,
+            "to_page": 100000000,
+            "progress_msg":  datetime.now().strftime("%H:%M:%S") + " " + msg
         }
 
     task = new_task()
+    for field in ["doc_id", "from_page", "to_page"]:
+        hasher.update(str(task.get(field, "")).encode("utf-8"))
+    hasher.update(ty.encode("utf-8"))
+    task["digest"] = hasher.hexdigest()
     bulk_insert_into_db(Task, [task], True)
-    task["type"] = "raptor"
+    task["task_type"] = ty
     assert REDIS_CONN.queue_product(SVR_QUEUE_NAME, message=task), "Can't access Redis. Please check the Redis' status."
 
 
 def doc_upload_and_parse(conversation_id, file_objs, user_id):
     from rag.app import presentation, picture, naive, audio, email
-    from api.db.services.dialog_service import ConversationService, DialogService
+    from api.db.services.dialog_service import DialogService
     from api.db.services.file_service import FileService
     from api.db.services.llm_service import LLMBundle
     from api.db.services.user_service import TenantService
     from api.db.services.api_service import API4ConversationService
+    from api.db.services.conversation_service import ConversationService
 
     e, conv = ConversationService.get_by_id(conversation_id)
     if not e:
@@ -462,11 +505,6 @@ def doc_upload_and_parse(conversation_id, file_objs, user_id):
     if not e:
         raise LookupError("Can't find this knowledgebase!")
 
-    idxnm = search.index_name(kb.tenant_id)
-    if not ELASTICSEARCH.indexExist(idxnm):
-        ELASTICSEARCH.createIdx(idxnm, json.load(
-            open(os.path.join(get_project_base_directory(), "conf", "mapping.json"), "r")))
-
     embd_mdl = LLMBundle(kb.tenant_id, LLMType.EMBEDDING, llm_name=kb.embd_id, lang=kb.language)
 
     err, files = FileService.upload_document(kb, file_objs, user_id)
@@ -481,7 +519,7 @@ def doc_upload_and_parse(conversation_id, file_objs, user_id):
         ParserType.AUDIO.value: audio,
         ParserType.EMAIL.value: email
     }
-    parser_config = {"chunk_token_num": 4096, "delimiter": "\n!?;。；！？", "layout_recognize": False}
+    parser_config = {"chunk_token_num": 4096, "delimiter": "\n!?;。；！？", "layout_recognize": "Plain Text"}
     exe = ThreadPoolExecutor(max_workers=12)
     threads = []
     doc_nm = {}
@@ -507,10 +545,7 @@ def doc_upload_and_parse(conversation_id, file_objs, user_id):
         for ck in th.result():
             d = deepcopy(doc)
             d.update(ck)
-            md5 = hashlib.md5()
-            md5.update((ck["content_with_weight"] +
-                        str(d["doc_id"])).encode("utf-8"))
-            d["_id"] = md5.hexdigest()
+            d["id"] = xxhash.xxh64((ck["content_with_weight"] + str(d["doc_id"])).encode("utf-8")).hexdigest()
             d["create_time"] = str(datetime.now()).replace("T", " ")[:19]
             d["create_timestamp_flt"] = datetime.now().timestamp()
             if not d.get("image"):
@@ -523,9 +558,9 @@ def doc_upload_and_parse(conversation_id, file_objs, user_id):
             else:
                 d["image"].save(output_buffer, format='JPEG')
 
-            STORAGE_IMPL.put(kb.id, d["_id"], output_buffer.getvalue())
-            d["img_id"] = "{}-{}".format(kb.id, d["_id"])
-            del d["image"]
+            STORAGE_IMPL.put(kb.id, d["id"], output_buffer.getvalue())
+            d["img_id"] = "{}-{}".format(kb.id, d["id"])
+            d.pop("image", None)
             docs.append(d)
 
     parser_ids = {d["id"]: d["parser_id"] for d, _ in files}
@@ -544,6 +579,9 @@ def doc_upload_and_parse(conversation_id, file_objs, user_id):
             token_counts[doc_id] += c
         return vects
 
+    idxnm = search.index_name(kb.tenant_id)
+    try_create_idx = True
+
     _, tenant = TenantService.get_by_id(kb.tenant_id)
     llm_bdl = LLMBundle(kb.tenant_id, LLMType.CHAT, tenant.llm_id)
     for doc_id in docids:
@@ -554,19 +592,20 @@ def doc_upload_and_parse(conversation_id, file_objs, user_id):
             try:
                 mind_map = json.dumps(mindmap([c["content_with_weight"] for c in docs if c["doc_id"] == doc_id]).output,
                                       ensure_ascii=False, indent=2)
-                if len(mind_map) < 32: raise Exception("Few content: " + mind_map)
+                if len(mind_map) < 32:
+                    raise Exception("Few content: " + mind_map)
                 cks.append({
                     "id": get_uuid(),
                     "doc_id": doc_id,
                     "kb_id": [kb.id],
                     "docnm_kwd": doc_nm[doc_id],
                     "title_tks": rag_tokenizer.tokenize(re.sub(r"\.[a-zA-Z]+$", "", doc_nm[doc_id])),
-                    "content_ltks": "",
+                    "content_ltks": rag_tokenizer.tokenize("summary summarize 总结 概况 file 文件 概括"),
                     "content_with_weight": mind_map,
                     "knowledge_graph_kwd": "mind_map"
                 })
             except Exception as e:
-                stat_logger.error("Mind map generation error:", traceback.format_exc())
+                logging.exception("Mind map generation error")
 
         vects = embedding(doc_id, [c["content_with_weight"] for c in cks])
         assert len(cks) == len(vects)
@@ -574,7 +613,11 @@ def doc_upload_and_parse(conversation_id, file_objs, user_id):
             v = vects[i]
             d["q_%d_vec" % len(v)] = v
         for b in range(0, len(cks), es_bulk_size):
-            ELASTICSEARCH.bulk(cks[b:b + es_bulk_size], idxnm)
+            if try_create_idx:
+                if not settings.docStoreConn.indexExist(idxnm, kb_id):
+                    settings.docStoreConn.createIdx(idxnm, kb_id, len(vects[0]))
+                try_create_idx = False
+            settings.docStoreConn.insert(cks[b:b + es_bulk_size], idxnm, kb_id)
 
         DocumentService.increment_chunk_num(
             doc_id, kb.id, token_counts[doc_id], chunk_counts[doc_id], 0)

api/db/services/file2document_service.py

@@ -20,7 +20,7 @@ from api.db.db_models import DB
 from api.db.db_models import File, File2Document
 from api.db.services.common_service import CommonService
 from api.db.services.document_service import DocumentService
-from api.utils import current_timestamp, datetime_format, get_uuid
+from api.utils import current_timestamp, datetime_format
 
 
 class File2DocumentService(CommonService):
@@ -43,10 +43,7 @@ class File2DocumentService(CommonService):
     def insert(cls, obj):
         if not cls.save(**obj):
             raise RuntimeError("Database error (File)!")
-        e, obj = cls.get_by_id(obj["id"])
-        if not e:
-            raise RuntimeError("Database error (File retrieval)!")
-        return obj
+        return File2Document(**obj)
 
     @classmethod
     @DB.connection_context()
@@ -63,9 +60,8 @@ class File2DocumentService(CommonService):
     def update_by_file_id(cls, file_id, obj):
         obj["update_time"] = current_timestamp()
         obj["update_date"] = datetime_format(datetime.now())
-        num = cls.model.update(obj).where(cls.model.id == file_id).execute()
-        e, obj = cls.get_by_id(cls.model.id)
-        return obj
+        cls.model.update(obj).where(cls.model.id == file_id).execute()
+        return File2Document(**obj)
 
     @classmethod
     @DB.connection_context()