feat: 准备实现知识库管理的功能

app/controller/add_knowledge_route.py

@@ -21,7 +21,7 @@ def add_knowledge_route(app):
       Document(
         text=text,
         id=id_list[index],
-        metadata={"kb_id": "abc"}
+        metadata={"kb_id": "cde"}
       )
       for index, text in enumerate(text_list)
     ]

app/utils/milvus_utils.py

@@ -2,6 +2,7 @@ import asyncio
 from typing import List, Optional
 
 from llama_index.core import Document, VectorStoreIndex
+from llama_index.core.vector_stores import MetadataFilters, MetadataFilter, FilterOperator
 from llama_index.vector_stores.milvus import MilvusVectorStore
 
 from app.config.env import env
@@ -65,24 +66,29 @@ class MilvusService:
       vector_store=await self.get_vector_store(),
       embed_model=self.embeddings,
     )
-    query_engine = vector_index.as_query_engine(
-      streaming=False,
-      similarity_top_k=top_k,  # 这里正确使用top_k参数
-      llm=create_llama_index_llm()
-    )
+
+    filters = MetadataFilters(filters=[MetadataFilter(key="kb_id", value="abc", operator=FilterOperator.EQ)])
+
+    # query_engine = vector_index.as_query_engine(
+    #   streaming=False,
+    #   similarity_top_k=top_k,  # 这里正确使用top_k参数
+    #   llm=create_llama_index_llm()
+    # )
+    retriever = vector_index.as_retriever(filters=filters, similarity_top_k=5)
+    result_nodes = await retriever.aretrieve(query)
 
     # 异步执行查询
-    response = await query_engine.aquery(query)
+    # response = await query_engine.aquery(query)
 
     return {
-      "answer": str(response),
+      # "answer": str(response),
       "sources": [
         {
           "text": node.node.text,
           "metadata": node.node.metadata,
           "score": node.score
         }
-        for node in response.source_nodes
+        for node in result_nodes
       ]
     }