feat: KnowledgeDocServiceWithCreator, optimize

app/model/KnowledgeDoc.py

@@ -87,25 +87,28 @@ KnowledgeDocService = create_model_service(
 
 # /*---------------------------------------KnowledgeDocServiceWithCreator-------------------------------------------*/
 
+def _create_document(cls: KnowledgeDocModel):
+  metadata = cls.model_dump()
+  # 删除metadata中的content属性值，避免将原文内容保存到元信息中
+  del metadata['content']
+  return Document(
+    doc_id=cls.id,
+    text=cls.content,
+    metadata=metadata
+  )
+
+
 # 新建文档之后，要给对应的文档创建向量索引
 async def handle_after_insert(insert_cls: KnowledgeDocModel, row_dict: dict, session: AsyncSessionDep):
-  await milvus_service.async_create_index_from_documents([Document(
+  metadata = insert_cls.model_dump()
-    doc_id=insert_cls.id,
+  # 删除metadata中的content属性值，避免将原文内容保存到元信息中
-    text=insert_cls.content,
+  del metadata['content']
-    metadata=insert_cls.model_dump()
+  await milvus_service.async_create_index_from_documents([_create_document(insert_cls)])
-  )])
 
 
 # 批量新建之后，给所有对应的文档创建向量索引
 async def handle_after_batch_insert(insert_cls_list: list[KnowledgeDocModel], row_dict_list: list[dict], session: AsyncSessionDep):
-  document_list = [
+  document_list = [_create_document(insert_cls) for insert_cls in insert_cls_list]
-    Document(
-      doc_id=insert_cls.id,
-      text=insert_cls.content,
-      metadata=insert_cls.model_dump()
-    )
-    for insert_cls in insert_cls_list
-  ]
   await milvus_service.async_create_index_from_documents(document_list)