add auto keywords and auto-question (#2965)

### What problem does this PR solve?

#2687

### Type of change

- [x] New Feature (non-breaking change which adds functionality)

api/apps/api_app.py

@@ -25,7 +25,7 @@ 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
+from api.db.services.dialog_service import DialogService, chat, keyword_extraction
 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
@@ -38,7 +38,6 @@ from api.utils.api_utils import server_error_response, get_data_error_result, ge
     generate_confirmation_token
 
 from api.utils.file_utils import filename_type, thumbnail
-from rag.nlp import keyword_extraction
 from rag.utils.storage_factory import STORAGE_IMPL
 
 from api.db.services.canvas_service import UserCanvasService

api/apps/chunk_app.py

@@ -21,8 +21,9 @@ 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 rag.app.qa import rmPrefix, beAdoc
-from rag.nlp import search, rag_tokenizer, keyword_extraction
+from rag.nlp import search, rag_tokenizer
 from rag.utils.es_conn import ELASTICSEARCH
 from rag.utils import rmSpace
 from api.db import LLMType, ParserType

api/apps/sdk/chat.py

@@ -16,16 +16,15 @@
 from flask import request
 
 from api.db import StatusEnum
-from api.db.db_models import TenantLLM
 from api.db.services.dialog_service import DialogService
 from api.db.services.knowledgebase_service import KnowledgebaseService
-from api.db.services.llm_service import LLMService, TenantLLMService
+from api.db.services.llm_service import TenantLLMService
 from api.db.services.user_service import TenantService
-from api.settings import RetCode
 from api.utils import get_uuid
 from api.utils.api_utils import get_error_data_result, token_required
 from api.utils.api_utils import get_result
 
+
 @manager.route('/chat', methods=['POST'])
 @token_required
 def create(tenant_id):

api/apps/sdk/dify_retrieval.py

@@ -1,10 +1,25 @@
+#
+#  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 flask import request, jsonify
 
-from db import LLMType, ParserType
-from db.services.knowledgebase_service import KnowledgebaseService
-from db.services.llm_service import LLMBundle
-from settings import retrievaler, kg_retrievaler, RetCode
-from utils.api_utils import validate_request, build_error_result, apikey_required
+from api.db import LLMType, ParserType
+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.utils.api_utils import validate_request, build_error_result, apikey_required
 
 
 @manager.route('/dify/retrieval', methods=['POST'])

api/apps/sdk/doc.py

@@ -1,48 +1,37 @@
+#
+#  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 pathlib
-import re
 import datetime
-import json
-import traceback
-
-from botocore.docs.method import document_model_driven_method
-from flask import request
-from flask_login import login_required, current_user
-from elasticsearch_dsl import Q
-from pygments import highlight
-from sphinx.addnodes import document
 
+from api.db.services.dialog_service import keyword_extraction
 from rag.app.qa import rmPrefix, beAdoc
-from rag.nlp import search, rag_tokenizer, keyword_extraction
-from rag.utils.es_conn import ELASTICSEARCH
-from rag.utils import rmSpace
+from rag.nlp import rag_tokenizer
 from api.db import LLMType, ParserType
-from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.llm_service import TenantLLMService
-from api.db.services.user_service import UserTenantService
-from api.utils.api_utils import server_error_response, get_error_data_result, validate_request
-from api.db.services.document_service import DocumentService
-from api.settings import RetCode, retrievaler, kg_retrievaler
-from api.utils.api_utils import get_result
+from api.settings import kg_retrievaler
 import hashlib
 import re
-from api.utils.api_utils import get_result, token_required, get_error_data_result
-
-from api.db.db_models import Task, File
-
+from api.utils.api_utils import token_required
+from api.db.db_models import Task
 from api.db.services.task_service import TaskService, queue_tasks
-from api.db.services.user_service import TenantService, UserTenantService
-
-from api.utils.api_utils import server_error_response, get_error_data_result, validate_request
-
-from api.utils.api_utils import get_result, get_result, get_error_data_result
-
-from functools import partial
+from api.utils.api_utils import server_error_response
+from api.utils.api_utils import get_result, get_error_data_result
 from io import BytesIO
-
 from elasticsearch_dsl import Q
 from flask import request, send_file
-from flask_login import login_required
-
 from api.db import FileSource, TaskStatus, FileType
 from api.db.db_models import File
 from api.db.services.document_service import DocumentService
@@ -50,8 +39,7 @@ 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.settings import RetCode, retrievaler
-from api.utils.api_utils import construct_json_result, construct_error_response
-from rag.app import book, laws, manual, naive, one, paper, presentation, qa, resume, table, picture, audio, email
+from api.utils.api_utils import construct_json_result
 from rag.nlp import search
 from rag.utils import rmSpace
 from rag.utils.es_conn import ELASTICSEARCH
@@ -365,7 +353,6 @@ def list_chunks(tenant_id,dataset_id,document_id):
     return get_result(data=res)
 
 
-
 @manager.route('/dataset/<dataset_id>/document/<document_id>/chunk', methods=['POST'])
 @token_required
 def create(tenant_id,dataset_id,document_id):
@@ -454,7 +441,6 @@ def rm_chunk(tenant_id,dataset_id,document_id):
     return get_result()
 
 
-
 @manager.route('/dataset/<dataset_id>/document/<document_id>/chunk/<chunk_id>', methods=['PUT'])
 @token_required
 def update_chunk(tenant_id,dataset_id,document_id,chunk_id):
@@ -512,7 +498,6 @@ def update_chunk(tenant_id,dataset_id,document_id,chunk_id):
     return get_result()
 
 
-
 @manager.route('/retrieval', methods=['POST'])
 @token_required
 def retrieval_test(tenant_id):

api/db/services/dialog_service.py

@@ -28,7 +28,6 @@ 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 rag.app.resume import forbidden_select_fields4resume
-from rag.nlp import keyword_extraction
 from rag.nlp.search import index_name
 from rag.utils import rmSpace, num_tokens_from_string, encoder
 from api.utils.file_utils import get_project_base_directory
@@ -80,6 +79,7 @@ class ConversationService(CommonService):
 
         return list(sessions.dicts())
 
+
 def message_fit_in(msg, max_length=4000):
     def count():
         nonlocal msg
@@ -456,6 +456,58 @@ def rewrite(tenant_id, llm_id, question):
     return ans
 
 
+def keyword_extraction(chat_mdl, content, topn=3):
+    prompt = f"""
+Role: You're a text analyzer. 
+Task: extract the most important keywords/phrases of a given piece of text content.
+Requirements: 
+  - Summarize the text content, and give top {topn} important keywords/phrases.
+  - The keywords MUST be in language of the given piece of text content.
+  - The keywords are delimited by ENGLISH COMMA.
+  - Keywords ONLY in output.
+
+### 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": 0.2})
+    if isinstance(kwd, tuple): kwd = kwd[0]
+    if kwd.find("**ERROR**") >=0: return ""
+    return kwd
+
+
+def question_proposal(chat_mdl, content, topn=3):
+    prompt = f"""
+Role: You're a text analyzer. 
+Task:  propose {topn} questions about a given piece of text content.
+Requirements: 
+  - Understand and summarize the text content, and propose top {topn} important questions.
+  - The questions SHOULD NOT have overlapping meanings.
+  - The questions SHOULD cover the main content of the text as much as possible.
+  - The questions MUST be in language of the given piece of text content.
+  - One question per line.
+  - Question ONLY in output.
+
+### 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": 0.2})
+    if isinstance(kwd, tuple): kwd = kwd[0]
+    if kwd.find("**ERROR**") >= 0: return ""
+    return kwd
+
+
 def full_question(tenant_id, llm_id, messages):
     if llm_id2llm_type(llm_id) == "image2text":
         chat_mdl = LLMBundle(tenant_id, LLMType.IMAGE2TEXT, llm_id)