[add] Create the attribute values of the community nodes

api/app/repositories/neo4j/community_repository.py

@@ -17,6 +17,7 @@ from app.repositories.neo4j.cypher_queries import (
     GET_ALL_COMMUNITY_MEMBERS_BATCH,
     CHECK_USER_HAS_COMMUNITIES,
     UPDATE_COMMUNITY_MEMBER_COUNT,
+    UPDATE_COMMUNITY_METADATA,
 )
 
 logger = logging.getLogger(__name__)
@@ -147,3 +148,26 @@ class CommunityRepository:
         except Exception as e:
             logger.error(f"refresh_member_count failed: {e}")
             return 0
+
+    async def update_community_metadata(
+        self,
+        community_id: str,
+        end_user_id: str,
+        name: str,
+        summary: str,
+        core_entities: List[str],
+    ) -> bool:
+        """更新社区的名称、摘要和核心实体列表。"""
+        try:
+            result = await self.connector.execute_query(
+                UPDATE_COMMUNITY_METADATA,
+                community_id=community_id,
+                end_user_id=end_user_id,
+                name=name,
+                summary=summary,
+                core_entities=core_entities,
+            )
+            return bool(result)
+        except Exception as e:
+            logger.error(f"update_community_metadata failed: {e}")
+            return False

api/app/repositories/neo4j/cypher_queries.py

@@ -1150,3 +1150,12 @@ WITH c, count(e) AS cnt
 SET c.member_count = cnt
 RETURN c.community_id AS community_id, cnt AS member_count
 """
+
+UPDATE_COMMUNITY_METADATA = """
+MATCH (c:Community {community_id: $community_id, end_user_id: $end_user_id})
+SET c.name         = $name,
+    c.summary      = $summary,
+    c.core_entities = $core_entities,
+    c.updated_at   = datetime()
+RETURN c.community_id AS community_id
+"""

api/app/repositories/neo4j/graph_saver.py

@@ -1,5 +1,6 @@
 import asyncio
-from typing import List
+import os
+from typing import List, Optional
 
 # 使用新的仓储层
 from app.repositories.neo4j.neo4j_connector import Neo4jConnector
@@ -156,7 +157,9 @@ async def save_dialog_and_statements_to_neo4j(
         entity_edges: List[EntityEntityEdge],
         statement_chunk_edges: List[StatementChunkEdge],
         statement_entity_edges: List[StatementEntityEdge],
-        connector: Neo4jConnector
+        connector: Neo4jConnector,
+        config_id: Optional[str] = None,
+        llm_model_id: Optional[str] = None,
 ) -> bool:
     """Save dialogue nodes, chunk nodes, statement nodes, entities, and all relationships to Neo4j using graph models.
 
@@ -290,12 +293,15 @@ async def save_dialog_and_statements_to_neo4j(
         logger.info("Transaction completed. Summary: %s", summary)
         logger.debug("Full transaction results: %r", results)
 
-        # 写入成功后，触发聚类
-        if entity_nodes:
+        # 写入成功后，触发聚类（可通过环境变量 CLUSTERING_ENABLED=false 禁用，用于基准测试对比）
+        clustering_enabled = os.getenv("CLUSTERING_ENABLED", "true").lower() != "false"
+        if entity_nodes and clustering_enabled:
             end_user_id = entity_nodes[0].end_user_id
             new_entity_ids = [e.id for e in entity_nodes]
             logger.info(f"[Clustering] 准备触发聚类，实体数: {len(new_entity_ids)}, end_user_id: {end_user_id}")
-            await _trigger_clustering(new_entity_ids, end_user_id)
+            asyncio.create_task(_trigger_clustering(new_entity_ids, end_user_id, config_id=config_id, llm_model_id=llm_model_id))
+        elif entity_nodes and not clustering_enabled:
+            logger.info("[Clustering] 聚类已禁用（CLUSTERING_ENABLED=false），跳过聚类触发")
 
         return True
 
@@ -309,6 +315,8 @@ async def save_dialog_and_statements_to_neo4j(
 async def _trigger_clustering(
     new_entity_ids: List[str],
     end_user_id: str,
+    config_id: Optional[str] = None,
+    llm_model_id: Optional[str] = None,
 ) -> None:
     """
     聚类触发函数，自动判断全量初始化还是增量更新。
@@ -318,7 +326,7 @@ async def _trigger_clustering(
         from app.core.memory.storage_services.clustering_engine import LabelPropagationEngine
         logger.info(f"[Clustering] 开始聚类，end_user_id={end_user_id}, 实体数={len(new_entity_ids)}")
         connector = Neo4jConnector()
-        engine = LabelPropagationEngine(connector)
+        engine = LabelPropagationEngine(connector, config_id=config_id, llm_model_id=llm_model_id)
         await engine.run(end_user_id=end_user_id, new_entity_ids=new_entity_ids)
         logger.info(f"[Clustering] 聚类完成，end_user_id={end_user_id}")
     except Exception as e: