增加代码知识库；修复文档处理内容；增加API设置

rag-web-ui/backend/app/api/api_v1/knowledge_base.py

@@ -28,6 +28,7 @@ from app.core.minio import get_minio_client
 from minio.error import MinioException
 from app.services.vector_store import VectorStoreFactory
 from app.services.embedding.embedding_factory import EmbeddingsFactory
+from app.services.model_config import ModelConfigService
 
 router = APIRouter()
 
@@ -163,19 +164,23 @@ async def delete_knowledge_base(
         # Get all document file paths before deletion
         document_paths = [doc.file_path for doc in kb.documents]
         
+        cleanup_errors = []
+
         # Initialize services
         minio_client = get_minio_client()
-        embeddings = EmbeddingsFactory.create()
-
-        vector_store = VectorStoreFactory.create(
-            store_type=settings.VECTOR_STORE_TYPE,
-            collection_name=f"kb_{kb_id}",
-            embedding_function=embeddings,
-        )
+        vector_store = None
+        try:
+            model_profile = ModelConfigService.get_active_config(db, current_user.id)
+            embeddings = EmbeddingsFactory.create(model_profile=model_profile)
+            vector_store = VectorStoreFactory.create(
+                store_type=settings.VECTOR_STORE_TYPE,
+                collection_name=f"kb_{kb_id}",
+                embedding_function=embeddings,
+            )
+        except Exception as e:
+            cleanup_errors.append(f"Failed to initialize vector store cleanup: {str(e)}")
         
         # Clean up external resources first
-        cleanup_errors = []
-        
         # 1. Clean up MinIO files
         try:
             # Delete all objects with prefix kb_{kb_id}/
@@ -188,12 +193,13 @@ async def delete_knowledge_base(
             logger.error(f"MinIO cleanup error for kb {kb_id}: {str(e)}")
         
         # 2. Clean up vector store
-        try:
-            vector_store._store.delete_collection(f"kb_{kb_id}")
-            logger.info(f"Cleaned up vector store for knowledge base {kb_id}")
-        except Exception as e:
-            cleanup_errors.append(f"Failed to clean up vector store: {str(e)}")
-            logger.error(f"Vector store cleanup error for kb {kb_id}: {str(e)}")
+        if vector_store is not None:
+            try:
+                vector_store._store.delete_collection(f"kb_{kb_id}")
+                logger.info(f"Cleaned up vector store for knowledge base {kb_id}")
+            except Exception as e:
+                cleanup_errors.append(f"Failed to clean up vector store: {str(e)}")
+                logger.error(f"Vector store cleanup error for kb {kb_id}: {str(e)}")
         
         # Finally, delete database records in a single transaction
         db.delete(kb)
@@ -366,6 +372,11 @@ async def process_kb_documents(
     
     if not upload_ids:
         return {"tasks": []}
+
+    try:
+        ModelConfigService.require_active_config(db, current_user.id)
+    except ValueError as exc:
+        raise HTTPException(status_code=400, detail=str(exc)) from exc
     
     uploads = db.query(DocumentUpload).filter(DocumentUpload.id.in_(upload_ids)).all()
     uploads_dict = {upload.id: upload for upload in uploads}
@@ -411,12 +422,13 @@ async def process_kb_documents(
     background_tasks.add_task(
         add_processing_tasks_to_queue,
         task_data,
-        kb_id
+        kb_id,
+        current_user.id,
     )
     
     return {"tasks": task_info}
 
-async def add_processing_tasks_to_queue(task_data, kb_id):
+async def add_processing_tasks_to_queue(task_data, kb_id, user_id):
     """Helper function to add document processing tasks to the queue without blocking the main response."""
     for data in task_data:
         asyncio.create_task(
@@ -425,7 +437,8 @@ async def add_processing_tasks_to_queue(task_data, kb_id):
                 data["file_name"],
                 kb_id,
                 data["task_id"],
-                None
+                None,
+                user_id=user_id,
             )
         )
     logger.info(f"Added {len(task_data)} document processing tasks to queue")
@@ -551,7 +564,11 @@ async def test_retrieval(
                 detail=f"Knowledge base {request.kb_id} not found",
             )
         
-        embeddings = EmbeddingsFactory.create()
+        try:
+            model_profile = ModelConfigService.require_active_config(db, current_user.id)
+        except ValueError as exc:
+            raise HTTPException(status_code=400, detail=str(exc)) from exc
+        embeddings = EmbeddingsFactory.create(model_profile=model_profile)
         
         vector_store = VectorStoreFactory.create(
             store_type=settings.VECTOR_STORE_TYPE,
@@ -571,5 +588,7 @@ async def test_retrieval(
             
         return {"results": response}
         
+    except HTTPException:
+        raise
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))