From dd7f88d6fa0231ba1003ceff0625a18eb69c58d2 Mon Sep 17 00:00:00 2001
From: martsforever <martsforever@aliyun.com>
Date: Sun, 7 Sep 2025 16:21:11 +0800
Subject: [PATCH] =?UTF-8?q?feat:=20=E5=87=86=E5=A4=87=E5=AE=9E=E7=8E=B0?=
 =?UTF-8?q?=E7=9F=A5=E8=AF=86=E5=BA=93=E7=AE=A1=E7=90=86=E7=9A=84=E5=8A=9F?=
 =?UTF-8?q?=E8=83=BD?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 app/controller/add_knowledge_route.py |  2 +-
 app/utils/milvus_utils.py             | 22 ++++++++++++++--------
 2 files changed, 15 insertions(+), 9 deletions(-)

diff --git a/app/controller/add_knowledge_route.py b/app/controller/add_knowledge_route.py
index 2972f29..33cd544 100644
--- a/app/controller/add_knowledge_route.py
+++ b/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)
     ]
diff --git a/app/utils/milvus_utils.py b/app/utils/milvus_utils.py
index 3e3cc28..10044d6 100644
--- a/app/utils/milvus_utils.py
+++ b/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
       ]
     }