Merge branch 'feat/milvus-support' into deploy/dev

# Conflicts:
#	web/app/components/base/icons/src/vender/line/general/index.ts

api/core/indexing_runner.py

@@ -10,6 +10,7 @@ from concurrent.futures import ThreadPoolExecutor, as_completed
 from typing import Optional, List, cast
 
 import openai
+from billiard.pool import Pool
 from flask import current_app, Flask
 from flask_login import current_user
 from langchain.embeddings import OpenAIEmbeddings
@@ -538,6 +539,11 @@ class IndexingRunner:
             #     document_format_thread.start()
             # for thread in threads:
             #     thread.join()
+            # with Pool(5) as pool:
+            #     for doc in documents:
+            #         result = pool.apply_async(format_document, kwds={'flask_app': current_app._get_current_object(), 'document_node': doc, 'split_documents': split_documents})
+            #         if result.ready():
+            #             split_documents.extend(result.get())
             with ThreadPoolExecutor(max_workers=10) as executor:
                 future_to_doc = {executor.submit(format_document, current_app._get_current_object(), doc): doc for doc in documents}
                 for future in concurrent.futures.as_completed(future_to_doc):

api/core/tool/dataset_retriever_tool.py

@@ -12,7 +12,7 @@ from core.index.keyword_table_index.keyword_table_index import KeywordTableIndex
 from core.index.vector_index.vector_index import VectorIndex
 from core.llm.llm_builder import LLMBuilder
 from extensions.ext_database import db
-from models.dataset import Dataset
+from models.dataset import Dataset, DocumentSegment
 
 
 class DatasetRetrieverToolInput(BaseModel):
@@ -68,6 +68,7 @@ class DatasetRetrieverTool(BaseTool):
             )
 
             documents = kw_table_index.search(query, search_kwargs={'k': self.k})
+            return str("\n".join([document.page_content for document in documents]))
         else:
             model_credentials = LLMBuilder.get_model_credentials(
                 tenant_id=dataset.tenant_id,
@@ -98,8 +99,22 @@ class DatasetRetrieverTool(BaseTool):
 
             hit_callback = DatasetIndexToolCallbackHandler(dataset.id)
             hit_callback.on_tool_end(documents)
+            document_context_list = []
+            index_node_ids = [document.metadata['doc_id'] for document in documents]
+            segments = DocumentSegment.query.filter(DocumentSegment.completed_at.isnot(None),
+                                                    DocumentSegment.status == 'completed',
+                                                    DocumentSegment.enabled == True,
+                                                    DocumentSegment.index_node_id.in_(index_node_ids)
+                                                    ).all()
+
+            if segments:
+                for segment in segments:
+                    if segment.answer:
+                        document_context_list.append(segment.answer)
+                    else:
+                        document_context_list.append(segment.content)
 
-        return str("\n".join([document.page_content for document in documents]))
+            return str("\n".join(document_context_list))
 
     async def _arun(self, tool_input: str) -> str:
         raise NotImplementedError()

web/app/components/base/icons/src/vender/line/general/index.ts

 export { default as Check } from './Check'
-export { default as LinkExternal02 } from './LinkExternal02'
+<<<<<<<<< Temporary merge branch 1
 export { default as Edit03 } from './Edit03'
 export { default as Hash02 } from './Hash02'
+export { default as LinkExternal02 } from './LinkExternal02'
 export { default as Loading02 } from './Loading02'
 export { default as LogOut01 } from './LogOut01'
 export { default as Trash03 } from './Trash03'