diff --git a/docs/docs/Components/bundles-docling.mdx b/docs/docs/Components/bundles-docling.mdx index 16d750d186..df675756a2 100644 --- a/docs/docs/Components/bundles-docling.mdx +++ b/docs/docs/Components/bundles-docling.mdx @@ -25,8 +25,9 @@ The Docling dependency is required to use the Docling components in Langflow. * **Langflow Desktop**: Set `LANGFLOW_DOCLING=True` in your `.env` file to enable Docling dependency installation. For more information, see [Set environment variables for Langflow Desktop](/environment-variables#set-environment-variables-for-langflow-desktop). * **Langflow OSS**: - * If you installed `langflow` (`uv pip install langflow`), Docling is included automatically through bundled extras. + * If you installed `langflow` (`uv pip install langflow`), Docling is included automatically through bundled extras, without Docling's optional chunking dependencies. * If you installed `langflow-base` directly, install Docling with an extra, for example `uv pip install "langflow-base[docling]"`. + * To use the **Chunk DoclingDocument** component, install the chunking extra with `uv pip install "langflow[docling-chunking]"` or `uv pip install "langflow-base[docling-chunking]"`. * **macOS Intel (x86_64)**: Use the [Docling installation guide](https://docling-project.github.io/docling/installation/) to install the Docling dependency. @@ -103,6 +104,8 @@ For more information, see the [Docling serve project repository](https://github. The **Chunk DoclingDocument** component splits `DoclingDocument` objects into chunks. +This component requires Docling's optional chunking dependencies. Install them with `uv pip install "langflow[docling-chunking]"` or `uv pip install "langflow-base[docling-chunking]"`. + It outputs the chunked documents as a [`Table`](/data-types#table). For more information, see the [Docling core project repository](https://github.com/docling-project/docling-core). diff --git a/pyproject.toml b/pyproject.toml index 2131f60390..df52949ac9 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -92,7 +92,7 @@ Documentation = "https://docs.langflow.org" [project.optional-dependencies] docling = [ - "langchain-docling>=1.1.0", + "langflow-base[docling]>=0.9.5", "tesserocr>=2.8.0", "rapidocr-onnxruntime>=1.4.4", "ocrmac>=1.0.0; sys_platform == 'darwin'", @@ -103,6 +103,11 @@ docling = [ "torchvision>=0.21.0", # torchvision 0.21+ is compatible with PyTorch 2.6+ ] +docling-chunking = [ + "langflow-base[docling-chunking]>=0.9.5", + "langchain-docling>=1.1.0", +] + audio = [ "webrtcvad>=2.0.10", ] diff --git a/src/backend/base/pyproject.toml b/src/backend/base/pyproject.toml index 4773730faf..992d8d3219 100644 --- a/src/backend/base/pyproject.toml +++ b/src/backend/base/pyproject.toml @@ -327,7 +327,12 @@ openlayer = ["openlayer>=0.9.0,<1.0.0"] # Document processing / OCR docling = [ "docling-core>=2.77.0,<3.0.0", - "docling>=2.36.1,<3.0.0; sys_platform != 'darwin' or platform_machine != 'x86_64'", + "docling-slim[cli,convert-core,extract-core,feat-ocr-rapidocr,format-latex,format-office,format-pdf,format-web,models-local,service-client]>=2.36.1,<3.0.0; sys_platform != 'darwin' or platform_machine != 'x86_64'", +] +docling-chunking = [ + "langflow-base[docling]", + "docling-core[chunking]>=2.77.0,<3.0.0", + "tiktoken>=0.7.0", ] easyocr = ["easyocr>=1.7.2,<2.0.0; sys_platform != 'darwin' or platform_machine != 'x86_64'"] diff --git a/src/backend/tests/unit/components/docling/test_chunk_docling_document_component.py b/src/backend/tests/unit/components/docling/test_chunk_docling_document_component.py index ca2928f7b0..5d13fa805b 100644 --- a/src/backend/tests/unit/components/docling/test_chunk_docling_document_component.py +++ b/src/backend/tests/unit/components/docling/test_chunk_docling_document_component.py @@ -1,14 +1,14 @@ """Tests for ChunkDoclingDocumentComponent HybridChunker parameters.""" +import builtins import sys import types import pytest - -pytest.importorskip("tiktoken") -pytest.importorskip("docling_core") - -from lfx.components.docling.chunk_docling_document import ChunkDoclingDocumentComponent +from lfx.components.docling.chunk_docling_document import ( + ChunkDoclingDocumentComponent, + _load_docling_chunker_dependencies, +) def _base_build_config(): @@ -95,6 +95,11 @@ class TestChunkDoclingDocumentComponentHybridChunker: captured["max_tokens"] = max_tokens return "tokenizer" + class DummyDocMeta: + @classmethod + def model_validate(cls, meta): + return meta + hybrid_chunker_module = types.ModuleType("docling_core.transforms.chunker.hybrid_chunker") hybrid_chunker_module.HybridChunker = DummyHybridChunker monkeypatch.setitem(sys.modules, "docling_core.transforms.chunker.hybrid_chunker", hybrid_chunker_module) @@ -110,8 +115,8 @@ class TestChunkDoclingDocumentComponentHybridChunker: huggingface_tokenizer_module, ) monkeypatch.setattr( - "lfx.components.docling.chunk_docling_document.HierarchicalChunker", - DummyHierarchicalChunker, + "lfx.components.docling.chunk_docling_document._load_docling_chunker_dependencies", + lambda: (DummyDocMeta, DummyHierarchicalChunker), ) monkeypatch.setattr( "lfx.components.docling.chunk_docling_document.extract_docling_documents", @@ -182,3 +187,20 @@ class TestChunkDoclingDocumentComponentHybridChunker: always_emit_headings_input=True, ) assert captured["hierarchical_called"] is True + + def test_missing_chunking_extra_has_actionable_error(self, monkeypatch): + original_import = builtins.__import__ + + def fake_import(name, globals_=None, locals_=None, fromlist=(), level=0): + if name in { + "docling_core.transforms.chunker.doc_chunk", + "docling_core.transforms.chunker.hierarchical_chunker", + }: + msg = "Module requires 'chunking' extra" + raise RuntimeError(msg) + return original_import(name, globals_, locals_, fromlist, level) + + monkeypatch.setattr(builtins, "__import__", fake_import) + + with pytest.raises(ImportError, match=r"langflow-base\[docling-chunking\]"): + _load_docling_chunker_dependencies() diff --git a/src/lfx/src/lfx/components/docling/chunk_docling_document.py b/src/lfx/src/lfx/components/docling/chunk_docling_document.py index 4137bf75f3..0f3aa83dad 100644 --- a/src/lfx/src/lfx/components/docling/chunk_docling_document.py +++ b/src/lfx/src/lfx/components/docling/chunk_docling_document.py @@ -1,14 +1,27 @@ import json - -import tiktoken -from docling_core.transforms.chunker import BaseChunker, DocMeta -from docling_core.transforms.chunker.hierarchical_chunker import HierarchicalChunker +from typing import Any from lfx.base.data.docling_utils import extract_docling_documents from lfx.custom import Component from lfx.io import BoolInput, DropdownInput, HandleInput, IntInput, MessageTextInput, Output, StrInput from lfx.schema import Data, DataFrame +_CHUNKING_INSTALL_HINT = ( + "Install them with `uv pip install 'langflow[docling-chunking]'`, " + "`uv pip install 'langflow-base[docling-chunking]'`, or " + "`uv pip install 'docling-core[chunking]'`." +) + + +def _load_docling_chunker_dependencies() -> tuple[type[Any], type[Any]]: + try: + from docling_core.transforms.chunker.doc_chunk import DocMeta as DocMetaCls + from docling_core.transforms.chunker.hierarchical_chunker import HierarchicalChunker as HierarchicalChunkerCls + except (ImportError, RuntimeError) as e: + msg = f"Docling chunking dependencies are not installed. {_CHUNKING_INSTALL_HINT}" + raise ImportError(msg) from e + return DocMetaCls, HierarchicalChunkerCls + class ChunkDoclingDocumentComponent(Component): display_name: str = "Chunk DoclingDocument" @@ -144,25 +157,20 @@ class ChunkDoclingDocumentComponent(Component): if warning: self.status = warning - chunker: BaseChunker + doc_meta_cls, hierarchical_chunker_cls = _load_docling_chunker_dependencies() + chunker: Any if self.chunker == "HybridChunker": try: from docling_core.transforms.chunker.hybrid_chunker import HybridChunker - except ImportError as e: - msg = ( - "HybridChunker is not installed. Please install it with `uv pip install docling-core[chunking] " - "or `uv pip install transformers`" - ) + except (ImportError, RuntimeError) as e: + msg = f"HybridChunker is not installed. {_CHUNKING_INSTALL_HINT}" raise ImportError(msg) from e max_tokens: int | None = self.max_tokens if self.max_tokens else None if self.provider == "Hugging Face": try: from docling_core.transforms.chunker.tokenizer.huggingface import HuggingFaceTokenizer - except ImportError as e: - msg = ( - "HuggingFaceTokenizer is not installed." - " Please install it with `uv pip install docling-core[chunking]`" - ) + except (ImportError, RuntimeError) as e: + msg = f"HuggingFaceTokenizer is not installed. {_CHUNKING_INSTALL_HINT}" raise ImportError(msg) from e tokenizer = HuggingFaceTokenizer.from_pretrained( model_name=self.hf_model_name, @@ -170,13 +178,10 @@ class ChunkDoclingDocumentComponent(Component): ) elif self.provider == "OpenAI": try: + import tiktoken from docling_core.transforms.chunker.tokenizer.openai import OpenAITokenizer - except ImportError as e: - msg = ( - "OpenAITokenizer is not installed." - " Please install it with `uv pip install docling-core[chunking]`" - " or `uv pip install transformers`" - ) + except (ImportError, RuntimeError) as e: + msg = f"OpenAITokenizer is not installed. {_CHUNKING_INSTALL_HINT}" raise ImportError(msg) from e if max_tokens is None: max_tokens = 128 * 1024 # context window length required for OpenAI tokenizers @@ -190,7 +195,7 @@ class ChunkDoclingDocumentComponent(Component): ) elif self.chunker == "HierarchicalChunker": - chunker = HierarchicalChunker() + chunker = hierarchical_chunker_cls() else: msg = f"Unknown chunker: {self.chunker}" raise ValueError(msg) @@ -200,7 +205,7 @@ class ChunkDoclingDocumentComponent(Component): for doc in documents: for chunk in chunker.chunk(dl_doc=doc): enriched_text = chunker.contextualize(chunk=chunk) - meta = DocMeta.model_validate(chunk.meta) + meta = doc_meta_cls.model_validate(chunk.meta) results.append( Data( diff --git a/uv.lock b/uv.lock index 52cc5b63f1..245a8b9e56 100644 --- a/uv.lock +++ b/uv.lock @@ -3023,6 +3023,61 @@ wheels = [ ] [package.optional-dependencies] +cli = [ + { name = "rich" }, + { name = "typer" }, +] +convert-core = [ + { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" }, + { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" }, + { name = "pillow" }, + { name = "rtree" }, + { name = "scipy", version = "1.15.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" }, + { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" }, +] +extract-core = [ + { name = "numpy", version = "2.2.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" }, + { name = "numpy", version = "2.4.6", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" }, + { name = "pillow" }, + { name = "polyfactory" }, + { name = "rtree" }, + { name = "scipy", version = "1.15.3", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.11'" }, + { name = "scipy", version = "1.17.1", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version >= '3.11'" }, +] +feat-ocr-rapidocr = [ + { name = "rapidocr" }, +] +format-latex = [ + { name = "pylatexenc" }, +] +format-office = [ + { name = "openpyxl" }, + { name = "python-docx" }, + { name = "python-pptx" }, +] +format-pdf = [ + { name = "docling-parse" }, + { name = "pypdfium2" }, +] +format-web = [ + { name = "beautifulsoup4" }, + { name = "lxml" }, + { name = "marko" }, +] +models-local = [ + { name = "accelerate" }, + { name = "defusedxml" }, + { name = "docling-ibm-models" }, + { name = "huggingface-hub" }, + { name = "torch", version = "2.12.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(python_full_version >= '3.14' and sys_platform == 'darwin') or (platform_machine == 'arm64' and sys_platform == 'darwin')" }, + { name = "torch", version = "2.12.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(python_full_version < '3.14' and platform_machine != 'arm64') or sys_platform != 'darwin'" }, + { name = "torchvision", version = "0.27.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(python_full_version >= '3.14' and sys_platform == 'darwin') or (platform_machine == 'arm64' and sys_platform == 'darwin')" }, + { name = "torchvision", version = "0.27.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(python_full_version < '3.14' and platform_machine != 'arm64') or sys_platform != 'darwin'" }, +] +service-client = [ + { name = "httpx" }, + { name = "websockets" }, +] standard = [ { name = "accelerate" }, { name = "beautifulsoup4" }, @@ -6934,7 +6989,7 @@ couchbase = [ { name = "couchbase" }, ] docling = [ - { name = "langchain-docling" }, + { name = "langflow-base", extra = ["docling"] }, { name = "ocrmac", marker = "sys_platform == 'darwin'" }, { name = "rapidocr-onnxruntime" }, { name = "tesserocr" }, @@ -6943,6 +6998,10 @@ docling = [ { name = "torchvision", version = "0.27.0", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(python_full_version >= '3.14' and sys_platform == 'darwin') or (platform_machine == 'arm64' and sys_platform == 'darwin')" }, { name = "torchvision", version = "0.27.0+cpu", source = { registry = "https://download.pytorch.org/whl/cpu" }, marker = "(python_full_version < '3.14' and platform_machine != 'arm64') or sys_platform != 'darwin'" }, ] +docling-chunking = [ + { name = "langchain-docling" }, + { name = "langflow-base", extra = ["docling-chunking"] }, +] local = [ { name = "ctransformers" }, { name = "sentence-transformers" }, @@ -7003,8 +7062,10 @@ requires-dist = [ { name = "clickhouse-connect", marker = "extra == 'clickhouse-connect'", specifier = "==0.7.19" }, { name = "couchbase", marker = "extra == 'couchbase'", specifier = ">=4.2.1" }, { name = "ctransformers", marker = "extra == 'local'", specifier = ">=0.2.10" }, - { name = "langchain-docling", marker = "extra == 'docling'", specifier = ">=1.1.0" }, + { name = "langchain-docling", marker = "extra == 'docling-chunking'", specifier = ">=1.1.0" }, { name = "langflow-base", extras = ["complete"], editable = "src/backend/base" }, + { name = "langflow-base", extras = ["docling"], marker = "extra == 'docling'", editable = "src/backend/base" }, + { name = "langflow-base", extras = ["docling-chunking"], marker = "extra == 'docling-chunking'", editable = "src/backend/base" }, { name = "nv-ingest-api", marker = "python_full_version >= '3.12' and extra == 'nv-ingest'", specifier = ">=26.1.0,<27.0.0" }, { name = "nv-ingest-client", marker = "python_full_version >= '3.12' and extra == 'nv-ingest'", specifier = ">=26.1.0,<27.0.0" }, { name = "ocrmac", marker = "sys_platform == 'darwin' and extra == 'docling'", specifier = ">=1.0.0" }, @@ -7017,7 +7078,7 @@ requires-dist = [ { name = "torchvision", marker = "extra == 'docling'", specifier = ">=0.21.0", index = "https://download.pytorch.org/whl/cpu" }, { name = "webrtcvad", marker = "extra == 'audio'", specifier = ">=2.0.10" }, ] -provides-extras = ["docling", "audio", "couchbase", "cassio", "local", "clickhouse-connect", "nv-ingest", "postgresql"] +provides-extras = ["docling", "docling-chunking", "audio", "couchbase", "cassio", "local", "clickhouse-connect", "nv-ingest", "postgresql"] [package.metadata.requires-dev] dev = [ @@ -7190,8 +7251,8 @@ all = [ { name = "cuga", marker = "(python_full_version < '3.14' and platform_machine == 'arm64') or (python_full_version < '3.14' and sys_platform != 'darwin')" }, { name = "datasets" }, { name = "ddgs" }, - { name = "docling", marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, { name = "docling-core" }, + { name = "docling-slim", extra = ["cli", "convert-core", "extract-core", "feat-ocr-rapidocr", "format-latex", "format-office", "format-pdf", "format-web", "models-local", "service-client"], marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, { name = "easyocr", marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, { name = "elasticsearch" }, { name = "faiss-cpu", version = "1.9.0.post1", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.14'" }, @@ -7359,8 +7420,8 @@ complete = [ { name = "cuga", marker = "(python_full_version < '3.14' and platform_machine == 'arm64') or (python_full_version < '3.14' and sys_platform != 'darwin')" }, { name = "datasets" }, { name = "ddgs" }, - { name = "docling", marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, { name = "docling-core" }, + { name = "docling-slim", extra = ["cli", "convert-core", "extract-core", "feat-ocr-rapidocr", "format-latex", "format-office", "format-pdf", "format-web", "models-local", "service-client"], marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, { name = "easyocr", marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, { name = "elasticsearch" }, { name = "faiss-cpu", version = "1.9.0.post1", source = { registry = "https://pypi.org/simple" }, marker = "python_full_version < '3.14'" }, @@ -7472,8 +7533,13 @@ datasets = [ { name = "datasets" }, ] docling = [ - { name = "docling", marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, { name = "docling-core" }, + { name = "docling-slim", extra = ["cli", "convert-core", "extract-core", "feat-ocr-rapidocr", "format-latex", "format-office", "format-pdf", "format-web", "models-local", "service-client"], marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, +] +docling-chunking = [ + { name = "docling-core", extra = ["chunking"] }, + { name = "docling-slim", extra = ["cli", "convert-core", "extract-core", "feat-ocr-rapidocr", "format-latex", "format-office", "format-pdf", "format-web", "models-local", "service-client"], marker = "platform_machine != 'x86_64' or sys_platform != 'darwin'" }, + { name = "tiktoken" }, ] docx = [ { name = "python-docx" }, @@ -7794,8 +7860,9 @@ requires-dist = [ { name = "datasets", marker = "extra == 'datasets'", specifier = ">2.14.7,<5.0.0" }, { name = "ddgs", marker = "extra == 'duckduckgo'", specifier = ">=9.0.0" }, { name = "defusedxml", specifier = ">=0.7.1,<1.0.0" }, - { name = "docling", marker = "(platform_machine != 'x86_64' and extra == 'docling') or (sys_platform != 'darwin' and extra == 'docling')", specifier = ">=2.36.1,<3.0.0" }, { name = "docling-core", marker = "extra == 'docling'", specifier = ">=2.77.0,<3.0.0" }, + { name = "docling-core", extras = ["chunking"], marker = "extra == 'docling-chunking'", specifier = ">=2.77.0,<3.0.0" }, + { name = "docling-slim", extras = ["cli", "convert-core", "extract-core", "feat-ocr-rapidocr", "format-latex", "format-office", "format-pdf", "format-web", "models-local", "service-client"], marker = "(platform_machine != 'x86_64' and extra == 'docling') or (sys_platform != 'darwin' and extra == 'docling')", specifier = ">=2.36.1,<3.0.0" }, { name = "docstring-parser", specifier = ">=0.16,<1.0.0" }, { name = "duckdb", specifier = ">=1.0.0,<2.0.0" }, { name = "dynaconf", specifier = ">=3.2.13,<4.0.0" }, @@ -7896,6 +7963,7 @@ requires-dist = [ { name = "langflow-base", extras = ["cuga"], marker = "extra == 'complete'", editable = "src/backend/base" }, { name = "langflow-base", extras = ["datasets"], marker = "extra == 'complete'", editable = "src/backend/base" }, { name = "langflow-base", extras = ["docling"], marker = "extra == 'complete'", editable = "src/backend/base" }, + { name = "langflow-base", extras = ["docling"], marker = "extra == 'docling-chunking'", editable = "src/backend/base" }, { name = "langflow-base", extras = ["docx"], marker = "extra == 'complete'", editable = "src/backend/base" }, { name = "langflow-base", extras = ["duckduckgo"], marker = "extra == 'complete'", editable = "src/backend/base" }, { name = "langflow-base", extras = ["easyocr"], marker = "extra == 'complete'", editable = "src/backend/base" }, @@ -8052,6 +8120,7 @@ requires-dist = [ { name = "sseclient-py", marker = "extra == 'sseclient'", specifier = "==1.8.0" }, { name = "structlog", specifier = ">=25.4.0,<26.0.0" }, { name = "supabase", marker = "extra == 'supabase'", specifier = ">=2.6.0,<3.0.0" }, + { name = "tiktoken", marker = "extra == 'docling-chunking'", specifier = ">=0.7.0" }, { name = "toolguard", marker = "extra == 'toolguard'", specifier = ">=0.2.16,<1.0.0" }, { name = "traceloop-sdk", marker = "extra == 'traceloop'", specifier = ">=0.43.1,<1.0.0" }, { name = "transformers", specifier = ">=5.6.0,<6.0.0" }, @@ -8072,7 +8141,7 @@ requires-dist = [ { name = "youtube-transcript-api", marker = "extra == 'youtube'", specifier = ">=1.0.0,<2.0.0" }, { name = "zep-python", marker = "extra == 'zep'", specifier = "==2.0.2" }, ] -provides-extras = ["audio", "postgresql", "local", "couchbase", "cassandra", "clickhouse", "mongodb", "supabase", "redis", "elasticsearch", "faiss", "qdrant", "qdrant-vectors", "weaviate", "chroma", "upstash", "pinecone", "milvus", "mongodb-vectors", "astradb", "openai", "anthropic", "cohere", "google", "ollama", "nvidia", "mistral", "sambanova", "groq", "sentence-transformers", "ctransformers", "langfuse", "langwatch", "langsmith", "arize", "pypdf", "docx", "pytube", "smolagents", "beautifulsoup", "serpapi", "google-api", "wikipedia", "fake-useragent", "duckduckgo", "yfinance", "wolframalpha", "pyarrow", "fastavro", "gitpython", "nltk", "lark", "huggingface", "metaphor", "metal", "markup", "aws", "numexpr", "qianfan", "pgvector", "litellm", "zep", "youtube", "markdown", "kubernetes", "json-repair", "composio", "ragstack", "opensearch", "atlassian", "mem0", "needle", "sseclient", "openinference", "mcp", "ag2", "scrapegraph", "apify", "spider", "altk", "toolguard", "langchain-huggingface", "langchain-unstructured", "graph-retriever", "langchain-mcp-adapters", "opik", "traceloop", "openlayer", "docling", "easyocr", "cleanlab", "twelvelabs", "jigsawstack", "assemblyai", "datasets", "fastparquet", "vlmrun", "cuga", "mlx", "fastmcp", "aioboto3", "astrapy", "gassist", "ibm-watsonx-clients", "complete", "all"] +provides-extras = ["audio", "postgresql", "local", "couchbase", "cassandra", "clickhouse", "mongodb", "supabase", "redis", "elasticsearch", "faiss", "qdrant", "qdrant-vectors", "weaviate", "chroma", "upstash", "pinecone", "milvus", "mongodb-vectors", "astradb", "openai", "anthropic", "cohere", "google", "ollama", "nvidia", "mistral", "sambanova", "groq", "sentence-transformers", "ctransformers", "langfuse", "langwatch", "langsmith", "arize", "pypdf", "docx", "pytube", "smolagents", "beautifulsoup", "serpapi", "google-api", "wikipedia", "fake-useragent", "duckduckgo", "yfinance", "wolframalpha", "pyarrow", "fastavro", "gitpython", "nltk", "lark", "huggingface", "metaphor", "metal", "markup", "aws", "numexpr", "qianfan", "pgvector", "litellm", "zep", "youtube", "markdown", "kubernetes", "json-repair", "composio", "ragstack", "opensearch", "atlassian", "mem0", "needle", "sseclient", "openinference", "mcp", "ag2", "scrapegraph", "apify", "spider", "altk", "toolguard", "langchain-huggingface", "langchain-unstructured", "graph-retriever", "langchain-mcp-adapters", "opik", "traceloop", "openlayer", "docling", "docling-chunking", "easyocr", "cleanlab", "twelvelabs", "jigsawstack", "assemblyai", "datasets", "fastparquet", "vlmrun", "cuga", "mlx", "fastmcp", "aioboto3", "astrapy", "gassist", "ibm-watsonx-clients", "complete", "all"] [package.metadata.requires-dev] dev = [