diff --git a/api/app/core/memory/agent/services/search_service.py b/api/app/core/memory/agent/services/search_service.py index 4fc4256e..c9346c16 100644 --- a/api/app/core/memory/agent/services/search_service.py +++ b/api/app/core/memory/agent/services/search_service.py @@ -120,7 +120,7 @@ class SearchService: raw_results is None if return_raw_results=False """ if include is None: - include = ["statements", "chunks", "entities", "summaries"] + include = ["statements", "chunks", "entities", "summaries", "communities"] # Clean query cleaned_query = self.clean_query(question) @@ -146,8 +146,8 @@ class SearchService: if search_type == "hybrid": reranked_results = answer.get('reranked_results', {}) - # Priority order: summaries first (most contextual), then statements, chunks, entities - priority_order = ['summaries', 'statements', 'chunks', 'entities'] + # Priority order: summaries first (most contextual), then communities, statements, chunks, entities + priority_order = ['summaries', 'communities', 'statements', 'chunks', 'entities'] for category in priority_order: if category in include and category in reranked_results: @@ -157,13 +157,43 @@ class SearchService: else: # For keyword or embedding search, results are directly in answer dict # Apply same priority order - priority_order = ['summaries', 'statements', 'chunks', 'entities'] + priority_order = ['summaries', 'communities', 'statements', 'chunks', 'entities'] for category in priority_order: if category in include and category in answer: category_results = answer[category] if isinstance(category_results, list): answer_list.extend(category_results) + + # 对命中的 community 节点展开其成员 statements + if "communities" in include: + community_results = ( + answer.get('reranked_results', {}).get('communities', []) + if search_type == "hybrid" + else answer.get('communities', []) + ) + community_ids = [ + r.get("id") for r in community_results if r.get("id") + ] + if community_ids and end_user_id: + from app.repositories.neo4j.graph_search import search_graph_community_expand + from app.repositories.neo4j.neo4j_connector import Neo4jConnector + expand_connector = Neo4jConnector() + try: + expand_result = await search_graph_community_expand( + connector=expand_connector, + community_ids=community_ids, + end_user_id=end_user_id, + limit=10, + ) + expanded_stmts = expand_result.get("expanded_statements", []) + if expanded_stmts: + answer_list.extend(expanded_stmts) + logger.info(f"社区展开检索追加 {len(expanded_stmts)} 条 statements") + except Exception as e: + logger.warning(f"社区展开检索失败，跳过: {e}") + finally: + await expand_connector.close() # Extract clean content from all results content_list = [ diff --git a/api/app/core/memory/agent/utils/write_tools.py b/api/app/core/memory/agent/utils/write_tools.py index b3707083..02aa1b44 100644 --- a/api/app/core/memory/agent/utils/write_tools.py +++ b/api/app/core/memory/agent/utils/write_tools.py @@ -19,7 +19,7 @@ from app.core.memory.utils.log.logging_utils import log_time from app.db import get_db_context from app.repositories.neo4j.add_edges import add_memory_summary_statement_edges from app.repositories.neo4j.add_nodes import add_memory_summary_nodes -from app.repositories.neo4j.graph_saver import save_dialog_and_statements_to_neo4j +from app.repositories.neo4j.graph_saver import save_dialog_and_statements_to_neo4j, schedule_clustering_after_write from app.repositories.neo4j.neo4j_connector import Neo4jConnector from app.schemas.memory_config_schema import MemoryConfig @@ -171,6 +171,13 @@ async def write( ) if success: logger.info("Successfully saved all data to Neo4j") + # 写入成功后，异步触发聚类（不阻塞写入响应） + schedule_clustering_after_write( + all_entity_nodes, + config_id=config_id, + llm_model_id=str(memory_config.llm_model_id) if memory_config.llm_model_id else None, + embedding_model_id=str(memory_config.embedding_model_id) if memory_config.embedding_model_id else None, + ) break else: logger.warning("Failed to save some data to Neo4j") diff --git a/api/app/core/memory/src/search.py b/api/app/core/memory/src/search.py index 0e1d8424..3570d707 100644 --- a/api/app/core/memory/src/search.py +++ b/api/app/core/memory/src/search.py @@ -238,7 +238,7 @@ def rerank_with_activation( reranked: Dict[str, List[Dict[str, Any]]] = {} - for category in ["statements", "chunks", "entities", "summaries"]: + for category in ["statements", "chunks", "entities", "summaries", "communities"]: keyword_items = keyword_results.get(category, []) embedding_items = embedding_results.get(category, []) @@ -281,21 +281,23 @@ def rerank_with_activation( for item in items_list: item_id = item.get("id") or item.get("uuid") or item.get("chunk_id") if item_id and item_id in combined_items: - combined_items[item_id]["normalized_activation_value"] = item.get("normalized_activation_value", 0) + combined_items[item_id]["normalized_activation_value"] = item.get("normalized_activation_value") # 步骤 4: 计算基础分数和最终分数 for item_id, item in combined_items.items(): bm25_norm = float(item.get("bm25_score", 0) or 0) emb_norm = float(item.get("embedding_score", 0) or 0) - act_norm = float(item.get("normalized_activation_value", 0) or 0) + # normalized_activation_value 为 None 表示该节点无激活值，保留 None 语义 + raw_act_norm = item.get("normalized_activation_value") + act_norm = float(raw_act_norm) if raw_act_norm is not None else None # 第一阶段：只考虑内容相关性（BM25 + Embedding） # alpha 控制 BM25 权重，(1-alpha) 控制 Embedding 权重 content_score = alpha * bm25_norm + (1 - alpha) * emb_norm base_score = content_score # 第一阶段用内容分数 - # 存储激活度分数供第二阶段使用 - item["activation_score"] = act_norm + # 存储激活度分数供第二阶段使用（None 表示无激活值，不参与激活值排序） + item["activation_score"] = act_norm # 可能为 None item["content_score"] = content_score item["base_score"] = base_score diff --git a/api/app/core/memory/storage_services/clustering_engine/label_propagation.py b/api/app/core/memory/storage_services/clustering_engine/label_propagation.py index cbc303b1..21257f2e 100644 --- a/api/app/core/memory/storage_services/clustering_engine/label_propagation.py +++ b/api/app/core/memory/storage_services/clustering_engine/label_propagation.py @@ -7,6 +7,7 @@ - 增量更新（incremental_update）：新实体到达时，只处理新实体及其邻居 """ +import asyncio import logging import uuid from math import sqrt @@ -19,8 +20,9 @@ logger = logging.getLogger(__name__) # 全量迭代最大轮数，防止不收敛 MAX_ITERATIONS = 10 -# 社区摘要核心实体数量 -CORE_ENTITY_LIMIT = 5 + +# 社区核心实体取 top-N 数量 +CORE_ENTITY_LIMIT = 10 def _cosine_similarity(v1: Optional[List[float]], v2: Optional[List[float]]) -> float: @@ -69,11 +71,13 @@ class LabelPropagationEngine: connector: Neo4jConnector, config_id: Optional[str] = None, llm_model_id: Optional[str] = None, + embedding_model_id: Optional[str] = None, ): self.connector = connector self.repo = CommunityRepository(connector) self.config_id = config_id self.llm_model_id = llm_model_id + self.embedding_model_id = embedding_model_id # ────────────────────────────────────────────────────────────────────────── # 公开接口 @@ -103,58 +107,81 @@ class LabelPropagationEngine: async def full_clustering(self, end_user_id: str) -> None: """ - 全量标签传播初始化。 + 全量标签传播初始化（分批处理，控制内存峰值）。 - 1. 拉取所有实体，初始化每个实体为独立社区 - 2. 迭代：每轮对所有实体做邻居投票，更新社区标签 - 3. 直到标签不再变化或达到 MAX_ITERATIONS - 4. 将最终标签写入 Neo4j + 策略： + - 每次只加载 BATCH_SIZE 个实体及其邻居进内存 + - labels 字典跨批次共享（只存 id→community_id，内存极小） + - 每批独立跑 MAX_ITERATIONS 轮 LPA，批次间通过 labels 传递社区信息 + - 所有批次完成后统一 flush 和 merge """ - entities = await self.repo.get_all_entities(end_user_id) - if not entities: + BATCH_SIZE = 888 # 每批实体数，可按需调整 + + # 轻量查询：只获取总数和 ID 列表，不加载 embedding 等大字段 + total_count = await self.repo.get_entity_count(end_user_id) + if not total_count: logger.info(f"[Clustering] 用户 {end_user_id} 无实体，跳过全量聚类") return - # 初始化：每个实体持有自己 id 作为社区标签 - labels: Dict[str, str] = {e["id"]: e["id"] for e in entities} - embeddings: Dict[str, Optional[List[float]]] = { - e["id"]: e.get("name_embedding") for e in entities - } + all_entity_ids = await self.repo.get_all_entity_ids(end_user_id) + logger.info(f"[Clustering] 用户 {end_user_id} 共 {total_count} 个实体，" + f"分批大小 {BATCH_SIZE}，共 {(total_count + BATCH_SIZE - 1) // BATCH_SIZE} 批") - # 预加载所有实体的邻居，避免迭代内 O(iterations * |E|) 次 Neo4j 往返 - logger.info(f"[Clustering] 预加载 {len(entities)} 个实体的邻居图...") - neighbors_cache: Dict[str, List[Dict]] = await self.repo.get_all_entity_neighbors_batch(end_user_id) - logger.info(f"[Clustering] 邻居预加载完成，覆盖实体数: {len(neighbors_cache)}") + # labels 跨批次共享：只存 id→community_id，内存极小 + labels: Dict[str, str] = {eid: eid for eid in all_entity_ids} + del all_entity_ids # 释放 ID 列表，后续按批次加载完整数据 - for iteration in range(MAX_ITERATIONS): - changed = 0 - # 随机顺序（Python dict 在 3.7+ 保持插入顺序，这里直接遍历） - for entity in entities: - eid = entity["id"] - # 直接从缓存取邻居，不再发起 Neo4j 查询 - neighbors = neighbors_cache.get(eid, []) - - # 将邻居的当前内存标签注入（覆盖 Neo4j 中的旧值） - enriched = [] - for nb in neighbors: - nb_copy = dict(nb) - nb_copy["community_id"] = labels.get(nb["id"], nb.get("community_id")) - enriched.append(nb_copy) - - new_label = _weighted_vote(enriched, embeddings.get(eid)) - if new_label and new_label != labels[eid]: - labels[eid] = new_label - changed += 1 - - logger.info( - f"[Clustering] 全量迭代 {iteration + 1}/{MAX_ITERATIONS}，" - f"标签变化数: {changed}" + for batch_start in range(0, total_count, BATCH_SIZE): + batch_entities = await self.repo.get_entities_page( + end_user_id, skip=batch_start, limit=BATCH_SIZE ) - if changed == 0: - logger.info("[Clustering] 标签已收敛，提前结束迭代") + if not batch_entities: break - # 将最终标签写入 Neo4j + batch_ids = [e["id"] for e in batch_entities] + batch_embeddings: Dict[str, Optional[List[float]]] = { + e["id"]: e.get("name_embedding") for e in batch_entities + } + + logger.info( + f"[Clustering] 批次 {batch_start // BATCH_SIZE + 1}：" + f"加载 {len(batch_entities)} 个实体的邻居图..." + ) + neighbors_cache = await self.repo.get_entity_neighbors_for_ids( + batch_ids, end_user_id + ) + logger.info(f"[Clustering] 邻居预加载完成，覆盖实体数: {len(neighbors_cache)}") + + for iteration in range(MAX_ITERATIONS): + changed = 0 + for entity in batch_entities: + eid = entity["id"] + neighbors = neighbors_cache.get(eid, []) + + # 注入跨批次的最新标签（邻居可能在其他批次，labels 里有其最新值） + enriched = [] + for nb in neighbors: + nb_copy = dict(nb) + nb_copy["community_id"] = labels.get(nb["id"], nb.get("community_id")) + enriched.append(nb_copy) + + new_label = _weighted_vote(enriched, batch_embeddings.get(eid)) + if new_label and new_label != labels[eid]: + labels[eid] = new_label + changed += 1 + + logger.info( + f"[Clustering] 批次 {batch_start // BATCH_SIZE + 1} " + f"迭代 {iteration + 1}/{MAX_ITERATIONS}，标签变化数: {changed}" + ) + if changed == 0: + logger.info("[Clustering] 标签已收敛，提前结束本批迭代") + break + + # 释放本批次的大对象 + del neighbors_cache, batch_embeddings, batch_entities + + # 所有批次完成，统一写入 Neo4j await self._flush_labels(labels, end_user_id) pre_merge_count = len(set(labels.values())) logger.info( @@ -162,7 +189,6 @@ class LabelPropagationEngine: f"{len(labels)} 个实体，开始后处理合并" ) - # 全量初始化后做一轮社区合并（基于 name_embedding 余弦相似度） all_community_ids = list(set(labels.values())) await self._evaluate_merge(all_community_ids, end_user_id) @@ -170,17 +196,15 @@ class LabelPropagationEngine: f"[Clustering] 全量聚类完成，合并前 {pre_merge_count} 个社区，" f"{len(labels)} 个实体" ) - # 为所有社区生成元数据 - # 注意：_evaluate_merge 后部分社区已被合并消解，需重新从 Neo4j 查询实际存活的社区 - # 不能复用 labels.values()，那里包含已被 dissolve 的旧社区 ID + + # 查询存活社区并生成元数据 surviving_communities = await self.repo.get_all_entities(end_user_id) surviving_community_ids = list({ e.get("community_id") for e in surviving_communities if e.get("community_id") }) logger.info(f"[Clustering] 合并后实际存活社区数: {len(surviving_community_ids)}") - for cid in surviving_community_ids: - await self._generate_community_metadata(cid, end_user_id) + await self._generate_community_metadata(surviving_community_ids, end_user_id) async def incremental_update( self, new_entity_ids: List[str], end_user_id: str @@ -237,7 +261,7 @@ class LabelPropagationEngine: logger.debug( f"[Clustering] 新实体 {entity_id} 与 {len(neighbors)} 个无社区邻居 → 新社区 {new_cid}" ) - await self._generate_community_metadata(new_cid, end_user_id) + await self._generate_community_metadata([new_cid], end_user_id) else: # 加入得票最多的社区 await self.repo.assign_entity_to_community(entity_id, target_cid, end_user_id) @@ -249,7 +273,7 @@ class LabelPropagationEngine: await self._evaluate_merge( list(community_ids_in_neighbors), end_user_id ) - await self._generate_community_metadata(target_cid, end_user_id) + await self._generate_community_metadata([target_cid], end_user_id) async def _evaluate_merge( self, community_ids: List[str], end_user_id: str @@ -413,71 +437,122 @@ class LabelPropagationEngine: except Exception: return None + @staticmethod + def _build_entity_lines(members: List[Dict]) -> List[str]: + """将实体列表格式化为 prompt 行，包含 name、aliases、description。""" + lines = [] + for m in members: + m_name = m.get("name", "") + aliases = m.get("aliases") or [] + description = m.get("description") or "" + aliases_str = f"（别名：{'、'.join(aliases)}）" if aliases else "" + desc_str = f"：{description}" if description else "" + lines.append(f"- {m_name}{aliases_str}{desc_str}") + return lines + async def _generate_community_metadata( - self, community_id: str, end_user_id: str + self, community_ids: List[str], end_user_id: str ) -> None: """ - 为社区生成并写入元数据：名称、摘要、核心实体。 + 为一个或多个社区生成并写入元数据。 - - core_entities：按 activation_value 排序取 top-N 实体名称列表（无需 LLM） - - name / summary：若有 llm_model_id 则调用 LLM 生成，否则用实体名称拼接兜底 + 流程： + 1. 逐个社区调 LLM 生成 name / summary（串行） + 2. 收集所有 summary，一次性批量 embed + 3. 单个社区用 update_community_metadata，多个用 batch_update_community_metadata """ - try: - members = await self.repo.get_community_members(community_id, end_user_id) - if not members: - return + if not community_ids: + return + + from app.db import get_db_context + from app.core.memory.utils.llm.llm_utils import MemoryClientFactory + + # --- 阶段1：并发调 LLM 生成每个社区的 name / summary --- + async def _build_one(cid: str): + members = await self.repo.get_community_members(cid, end_user_id) + if not members: + return None - # 核心实体：按 activation_value 降序取 top-N sorted_members = sorted( members, key=lambda m: m.get("activation_value") or 0, reverse=True, ) core_entities = [m["name"] for m in sorted_members[:CORE_ENTITY_LIMIT] if m.get("name")] - all_names = [m["name"] for m in members if m.get("name")] - name = "、".join(core_entities[:3]) if core_entities else community_id[:8] - summary = f"包含实体：{', '.join(all_names)}" - - # 若有 LLM 配置，调用 LLM 生成更好的名称和摘要 - if self.llm_model_id: - try: - from app.db import get_db_context - from app.core.memory.utils.llm.llm_utils import MemoryClientFactory - - entity_list_str = "、".join(all_names) - prompt = ( - f"以下是一组语义相关的实体：{entity_list_str}\n\n" - f"请为这组实体所代表的主题：\n" - f"1. 起一个简洁的中文名称（不超过10个字）\n" - f"2. 写一句话摘要（不超过50个字）\n\n" - f"严格按以下格式输出，不要有其他内容：\n" - f"名称：<名称>\n摘要：<摘要>" - ) - with get_db_context() as db: - factory = MemoryClientFactory(db) - llm_client = factory.get_llm_client(self.llm_model_id) - response = await llm_client.chat([{"role": "user", "content": prompt}]) - text = response.content if hasattr(response, "content") else str(response) - - for line in text.strip().splitlines(): - if line.startswith("名称："): - name = line[3:].strip() - elif line.startswith("摘要："): - summary = line[3:].strip() - except Exception as e: - logger.warning(f"[Clustering] LLM 生成社区元数据失败，使用兜底值: {e}") - - await self.repo.update_community_metadata( - community_id=community_id, - end_user_id=end_user_id, - name=name, - summary=summary, - core_entities=core_entities, + entity_list_str = "\n".join(self._build_entity_lines(members)) + prompt = ( + f"以下是一组语义相关的实体：\n{entity_list_str}\n\n" + f"请为这组实体所代表的主题：\n" + f"1. 起一个简洁的中文名称（不超过10个字）\n" + f"2. 写一句话摘要（不超过50个字）\n\n" + f"严格按以下格式输出，不要有其他内容：\n" + f"名称：<名称>\n摘要：<摘要>" ) - logger.debug(f"[Clustering] 社区 {community_id} 元数据已更新: name={name}") - except Exception as e: - logger.error(f"[Clustering] _generate_community_metadata failed for {community_id}: {e}") + with get_db_context() as db: + llm_client = MemoryClientFactory(db).get_llm_client(self.llm_model_id) + response = await llm_client.chat([{"role": "user", "content": prompt}]) + text = response.content if hasattr(response, "content") else str(response) + + name, summary = "", "" + for line in text.strip().splitlines(): + if line.startswith("名称："): + name = line[3:].strip() + elif line.startswith("摘要："): + summary = line[3:].strip() + + return { + "community_id": cid, + "end_user_id": end_user_id, + "name": name, + "summary": summary, + "core_entities": core_entities, + "summary_embedding": None, + } + + results = await asyncio.gather( + *[_build_one(cid) for cid in community_ids], + return_exceptions=True, + ) + metadata_list = [] + for cid, res in zip(community_ids, results): + if isinstance(res, Exception): + logger.error(f"[Clustering] 社区 {cid} 元数据准备失败: {res}", exc_info=res) + elif res is not None: + metadata_list.append(res) + + if not metadata_list: + return + + # --- 阶段2：批量生成 summary_embedding --- + summaries = [m["summary"] for m in metadata_list] + with get_db_context() as db: + embedder = MemoryClientFactory(db).get_embedder_client(self.embedding_model_id) + embeddings = await embedder.response(summaries) + for i, meta in enumerate(metadata_list): + meta["summary_embedding"] = embeddings[i] if i < len(embeddings) else None + + # --- 阶段3：写入（单个 or 批量）--- + if len(metadata_list) == 1: + m = metadata_list[0] + result = await self.repo.update_community_metadata( + community_id=m["community_id"], + end_user_id=m["end_user_id"], + name=m["name"], + summary=m["summary"], + core_entities=m["core_entities"], + summary_embedding=m["summary_embedding"], + ) + if result: + logger.info(f"[Clustering] 社区 {m['community_id']} 元数据写入成功: name={m['name']}, summary={m['summary'][:30]}...") + else: + logger.warning(f"[Clustering] 社区 {m['community_id']} 元数据写入返回 False") + else: + ok = await self.repo.batch_update_community_metadata(metadata_list) + if ok: + logger.info(f"[Clustering] 批量写入 {len(metadata_list)} 个社区元数据成功") + else: + logger.warning(f"[Clustering] 批量写入社区元数据失败") @staticmethod def _new_community_id() -> str: diff --git a/api/app/models/memory_perceptual_model.py b/api/app/models/memory_perceptual_model.py index 9fed7c5d..ae8cc1bd 100644 --- a/api/app/models/memory_perceptual_model.py +++ b/api/app/models/memory_perceptual_model.py @@ -9,7 +9,6 @@ from sqlalchemy.dialects.postgresql import JSONB from app.db import Base from app.schemas import FileType - class PerceptualType(IntEnum): VISION = 1 AUDIO = 2 diff --git a/api/app/repositories/neo4j/community_repository.py b/api/app/repositories/neo4j/community_repository.py index f2f11f76..f9c4bd92 100644 --- a/api/app/repositories/neo4j/community_repository.py +++ b/api/app/repositories/neo4j/community_repository.py @@ -13,12 +13,17 @@ from app.repositories.neo4j.cypher_queries import ( ENTITY_LEAVE_ALL_COMMUNITIES, GET_ENTITY_NEIGHBORS, GET_ALL_ENTITIES_FOR_USER, + GET_ENTITY_COUNT_FOR_USER, + GET_ALL_ENTITY_IDS_FOR_USER, + GET_ENTITIES_PAGE, GET_COMMUNITY_MEMBERS, GET_ALL_COMMUNITY_MEMBERS_BATCH, GET_ALL_ENTITY_NEIGHBORS_BATCH, + GET_ENTITY_NEIGHBORS_BATCH_FOR_IDS, CHECK_USER_HAS_COMMUNITIES, UPDATE_COMMUNITY_MEMBER_COUNT, UPDATE_COMMUNITY_METADATA, + BATCH_UPDATE_COMMUNITY_METADATA, ) logger = logging.getLogger(__name__) @@ -110,6 +115,65 @@ class CommunityRepository: logger.error(f"get_all_entities failed: {e}") return [] + async def get_entity_count(self, end_user_id: str) -> int: + """仅返回用户实体总数，不加载实体数据。""" + try: + result = await self.connector.execute_query( + GET_ENTITY_COUNT_FOR_USER, + end_user_id=end_user_id, + ) + return result[0]["entity_count"] if result else 0 + except Exception as e: + logger.error(f"get_entity_count failed: {e}") + return 0 + + async def get_all_entity_ids(self, end_user_id: str) -> List[str]: + """仅返回用户所有实体 ID 列表，不加载 embedding 等大字段。""" + try: + result = await self.connector.execute_query( + GET_ALL_ENTITY_IDS_FOR_USER, + end_user_id=end_user_id, + ) + return [r["id"] for r in result] + except Exception as e: + logger.error(f"get_all_entity_ids failed: {e}") + return [] + + async def get_entities_page( + self, end_user_id: str, skip: int, limit: int + ) -> List[Dict]: + """分页拉取实体，用于全量聚类分批处理。""" + try: + return await self.connector.execute_query( + GET_ENTITIES_PAGE, + end_user_id=end_user_id, + skip=skip, + limit=limit, + ) + except Exception as e: + logger.error(f"get_entities_page failed: {e}") + return [] + + async def get_entity_neighbors_for_ids( + self, entity_ids: List[str], end_user_id: str + ) -> Dict[str, List[Dict]]: + """批量拉取指定实体列表的邻居，返回 {entity_id: [neighbors]}。""" + try: + rows = await self.connector.execute_query( + GET_ENTITY_NEIGHBORS_BATCH_FOR_IDS, + entity_ids=entity_ids, + end_user_id=end_user_id, + ) + result: Dict[str, List[Dict]] = {} + for row in rows: + eid = row["entity_id"] + neighbor = {k: v for k, v in row.items() if k != "entity_id"} + result.setdefault(eid, []).append(neighbor) + return result + except Exception as e: + logger.error(f"get_entity_neighbors_for_ids failed: {e}") + return {} + async def get_community_members( self, community_id: str, end_user_id: str ) -> List[Dict]: @@ -177,8 +241,9 @@ class CommunityRepository: name: str, summary: str, core_entities: List[str], + summary_embedding: Optional[List[float]] = None, ) -> bool: - """更新社区的名称、摘要和核心实体列表。""" + """更新社区的名称、摘要、核心实体列表和摘要向量。""" try: result = await self.connector.execute_query( UPDATE_COMMUNITY_METADATA, @@ -187,8 +252,31 @@ class CommunityRepository: name=name, summary=summary, core_entities=core_entities, + summary_embedding=summary_embedding, ) return bool(result) except Exception as e: logger.error(f"update_community_metadata failed: {e}") return False + + async def batch_update_community_metadata( + self, + communities: List[Dict], + ) -> bool: + """批量更新多个社区的元数据。 + + Args: + communities: 每项包含 community_id, end_user_id, name, summary, + core_entities, summary_embedding + """ + if not communities: + return True + try: + await self.connector.execute_query( + BATCH_UPDATE_COMMUNITY_METADATA, + communities=communities, + ) + return True + except Exception as e: + logger.error(f"batch_update_community_metadata failed: {e}") + return False diff --git a/api/app/repositories/neo4j/create_indexes.py b/api/app/repositories/neo4j/create_indexes.py index 55dead1b..29f60fdd 100644 --- a/api/app/repositories/neo4j/create_indexes.py +++ b/api/app/repositories/neo4j/create_indexes.py @@ -42,6 +42,13 @@ async def create_fulltext_indexes(): OPTIONS { indexConfig: { `fulltext.analyzer`: 'cjk' } } """) print("✓ Created: summariesFulltext") + + # 创建 Community 索引 + await connector.execute_query(""" + CREATE FULLTEXT INDEX communitiesFulltext IF NOT EXISTS FOR (c:Community) ON EACH [c.name, c.summary] + OPTIONS { indexConfig: { `fulltext.analyzer`: 'cjk' } } + """) + print("✓ Created: communitiesFulltext") print("\nFull-text indexes created successfully with BM25 support.") except Exception as e: @@ -124,6 +131,18 @@ async def create_vector_indexes(): }} """) print("✓ Created: dialogue_embedding_index") + + # Community summary embedding index + await connector.execute_query(""" + CREATE VECTOR INDEX community_summary_embedding_index IF NOT EXISTS + FOR (c:Community) + ON c.summary_embedding + OPTIONS {indexConfig: { + `vector.dimensions`: 1024, + `vector.similarity_function`: 'cosine' + }} + """) + print("✓ Created: community_summary_embedding_index") print("\nVector indexes created successfully!") print("\nExpected performance improvement:") diff --git a/api/app/repositories/neo4j/cypher_queries.py b/api/app/repositories/neo4j/cypher_queries.py index 48a5ac87..16a26b3b 100644 --- a/api/app/repositories/neo4j/cypher_queries.py +++ b/api/app/repositories/neo4j/cypher_queries.py @@ -1122,21 +1122,33 @@ RETURN e.id AS id, CASE WHEN c IS NOT NULL THEN c.community_id ELSE null END AS community_id """ +GET_ENTITY_COUNT_FOR_USER = """ +MATCH (e:ExtractedEntity {end_user_id: $end_user_id}) +RETURN count(e) AS entity_count +""" + +GET_ALL_ENTITY_IDS_FOR_USER = """ +MATCH (e:ExtractedEntity {end_user_id: $end_user_id}) +RETURN e.id AS id +""" + GET_COMMUNITY_MEMBERS = """ MATCH (e:ExtractedEntity {end_user_id: $end_user_id})-[:BELONGS_TO_COMMUNITY]->(c:Community {community_id: $community_id}) RETURN e.id AS id, e.name AS name, e.entity_type AS entity_type, e.importance_score AS importance_score, e.activation_value AS activation_value, - e.name_embedding AS name_embedding + e.name_embedding AS name_embedding, + e.aliases AS aliases, e.description AS description ORDER BY coalesce(e.activation_value, 0) DESC """ GET_ALL_COMMUNITY_MEMBERS_BATCH = """ MATCH (e:ExtractedEntity {end_user_id: $end_user_id})-[:BELONGS_TO_COMMUNITY]->(c:Community) -WHERE c.community_id IN $community_ids RETURN c.community_id AS community_id, - e.id AS id, + e.id AS id, e.name AS name, e.entity_type AS entity_type, + e.importance_score AS importance_score, e.activation_value AS activation_value, e.name_embedding AS name_embedding, - e.activation_value AS activation_value + e.aliases AS aliases, e.description AS description +ORDER BY c.community_id, coalesce(e.activation_value, 0) DESC """ CHECK_USER_HAS_COMMUNITIES = """ @@ -1153,13 +1165,58 @@ 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() +SET c.name = $name, + c.summary = $summary, + c.core_entities = $core_entities, + c.summary_embedding = $summary_embedding, + c.updated_at = datetime() RETURN c.community_id AS community_id """ +BATCH_UPDATE_COMMUNITY_METADATA = """ +UNWIND $communities AS row +MATCH (c:Community {community_id: row.community_id, end_user_id: row.end_user_id}) +SET c.name = row.name, + c.summary = row.summary, + c.core_entities = row.core_entities, + c.summary_embedding = row.summary_embedding, + c.updated_at = datetime() +RETURN c.community_id AS community_id +""" + +GET_ENTITIES_PAGE = """ +MATCH (e:ExtractedEntity {end_user_id: $end_user_id}) +OPTIONAL MATCH (e)-[:BELONGS_TO_COMMUNITY]->(c:Community) +RETURN e.id AS id, + e.name AS name, + e.name_embedding AS name_embedding, + e.activation_value AS activation_value, + CASE WHEN c IS NOT NULL THEN c.community_id ELSE null END AS community_id +ORDER BY e.id +SKIP $skip LIMIT $limit +""" + +GET_ENTITY_NEIGHBORS_BATCH_FOR_IDS = """ +// 批量拉取指定实体列表的邻居（用于分批全量聚类） +MATCH (e:ExtractedEntity {end_user_id: $end_user_id}) +WHERE e.id IN $entity_ids +OPTIONAL MATCH (e)-[:EXTRACTED_RELATIONSHIP]-(nb1:ExtractedEntity {end_user_id: $end_user_id}) +OPTIONAL MATCH (s:Statement)-[:REFERENCES_ENTITY]->(e) +OPTIONAL MATCH (s)-[:REFERENCES_ENTITY]->(nb2:ExtractedEntity {end_user_id: $end_user_id}) +WHERE nb2.id <> e.id +WITH e, collect(DISTINCT nb1) + collect(DISTINCT nb2) AS all_neighbors +UNWIND all_neighbors AS nb +WITH e, nb WHERE nb IS NOT NULL +OPTIONAL MATCH (nb)-[:BELONGS_TO_COMMUNITY]->(c:Community) +RETURN DISTINCT + e.id AS entity_id, + nb.id AS id, + nb.name AS name, + nb.name_embedding AS name_embedding, + nb.activation_value AS activation_value, + CASE WHEN c IS NOT NULL THEN c.community_id ELSE null END AS community_id +""" + GET_ALL_ENTITY_NEIGHBORS_BATCH = """ // 批量拉取某用户下所有实体的邻居（用于全量聚类预加载） MATCH (e:ExtractedEntity {end_user_id: $end_user_id}) @@ -1185,20 +1242,59 @@ RETURN DISTINCT CASE WHEN c IS NOT NULL THEN c.community_id ELSE null END AS community_id """ -GET_COMMUNITY_GRAPH_DATA = """ -MATCH (c:Community {end_user_id: $end_user_id}) -MATCH (e:ExtractedEntity {end_user_id: $end_user_id})-[b:BELONGS_TO_COMMUNITY]->(c) -OPTIONAL MATCH (e)-[r:EXTRACTED_RELATIONSHIP]-(e2:ExtractedEntity {end_user_id: $end_user_id}) -RETURN - elementId(c) AS c_id, - properties(c) AS c_props, - elementId(e) AS e_id, - properties(e) AS e_props, - elementId(b) AS b_id, - elementId(e2) AS e2_id, - properties(e2) AS e2_props, - elementId(r) AS r_id, - type(r) AS r_type, - properties(r) AS r_props, - startNode(r) = e AS r_from_e + +# Community keyword search: matches name or summary via fulltext index +SEARCH_COMMUNITIES_BY_KEYWORD = """ +CALL db.index.fulltext.queryNodes("communitiesFulltext", $q) YIELD node AS c, score +WHERE ($end_user_id IS NULL OR c.end_user_id = $end_user_id) +RETURN c.community_id AS id, + c.name AS name, + c.summary AS content, + c.core_entities AS core_entities, + c.member_count AS member_count, + c.end_user_id AS end_user_id, + c.updated_at AS updated_at, + score +ORDER BY score DESC +LIMIT $limit +""" + +# Community 向量检索 ────────────────────────────────────────────────── +# Community embedding-based search: cosine similarity on Community.summary_embedding +COMMUNITY_EMBEDDING_SEARCH = """ +CALL db.index.vector.queryNodes('community_summary_embedding_index', $limit * 100, $embedding) +YIELD node AS c, score +WHERE c.summary_embedding IS NOT NULL + AND ($end_user_id IS NULL OR c.end_user_id = $end_user_id) +RETURN c.community_id AS id, + c.name AS name, + c.summary AS content, + c.core_entities AS core_entities, + c.member_count AS member_count, + c.end_user_id AS end_user_id, + c.updated_at AS updated_at, + score +ORDER BY score DESC +LIMIT $limit +""" + +# Community 展开检索 ────────────────────────────────────────────────── +# 命中社区后，拉取该社区所有成员实体关联的 Statement 节点（主题→细节两级检索） +EXPAND_COMMUNITY_STATEMENTS = """ +MATCH (c:Community {community_id: $community_id}) +MATCH (e:ExtractedEntity)-[:BELONGS_TO_COMMUNITY]->(c) +MATCH (s:Statement)-[:REFERENCES_ENTITY]->(e) +WHERE s.end_user_id = $end_user_id +RETURN s.statement AS statement, + s.id AS id, + s.end_user_id AS end_user_id, + s.created_at AS created_at, + s.valid_at AS valid_at, + s.invalid_at AS invalid_at, + COALESCE(s.activation_value, s.importance_score, 0.5) AS activation_value, + COALESCE(s.importance_score, 0.5) AS importance_score, + e.name AS source_entity, + c.name AS community_name +ORDER BY COALESCE(s.activation_value, 0) DESC +LIMIT $limit """ diff --git a/api/app/repositories/neo4j/graph_saver.py b/api/app/repositories/neo4j/graph_saver.py index cbd2b532..29e337f1 100644 --- a/api/app/repositories/neo4j/graph_saver.py +++ b/api/app/repositories/neo4j/graph_saver.py @@ -1,5 +1,4 @@ import asyncio -import os from typing import List, Optional # 使用新的仓储层 @@ -158,11 +157,12 @@ async def save_dialog_and_statements_to_neo4j( statement_chunk_edges: List[StatementChunkEdge], statement_entity_edges: List[StatementEntityEdge], 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. + 只负责数据写入，不触发聚类。聚类由调用方在写入成功后通过 + schedule_clustering_after_write() 显式触发。 + Args: dialogue_nodes: List of DialogueNode objects to save chunk_nodes: List of ChunkNode objects to save @@ -293,9 +293,6 @@ 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，异步触发聚类（不阻塞写入响应） - schedule_clustering_after_write(entity_nodes, config_id=config_id, llm_model_id=llm_model_id) - return True except Exception as e: @@ -309,6 +306,7 @@ def schedule_clustering_after_write( entity_nodes: List, config_id: Optional[str] = None, llm_model_id: Optional[str] = None, + embedding_model_id: Optional[str] = None, ) -> None: """ 写入 Neo4j 成功后，调度后台聚类任务。 @@ -327,7 +325,7 @@ def schedule_clustering_after_write( 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}") - asyncio.create_task(_trigger_clustering(new_entity_ids, end_user_id, config_id=config_id, llm_model_id=llm_model_id)) + asyncio.create_task(_trigger_clustering(new_entity_ids, end_user_id, config_id=config_id, llm_model_id=llm_model_id, embedding_model_id=embedding_model_id)) async def _trigger_clustering( @@ -335,6 +333,7 @@ async def _trigger_clustering( end_user_id: str, config_id: Optional[str] = None, llm_model_id: Optional[str] = None, + embedding_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。 @@ -344,7 +343,7 @@ 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, config_id=config_id, llm_model_id=llm_model_id) + engine = LabelPropagationEngine(connector, config_id=config_id, llm_model_id=llm_model_id, embedding_model_id=embedding_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: diff --git a/api/app/repositories/neo4j/graph_search.py b/api/app/repositories/neo4j/graph_search.py index e8f52535..19e40a82 100644 --- a/api/app/repositories/neo4j/graph_search.py +++ b/api/app/repositories/neo4j/graph_search.py @@ -4,10 +4,13 @@ from typing import Any, Dict, List, Optional from app.repositories.neo4j.cypher_queries import ( CHUNK_EMBEDDING_SEARCH, + COMMUNITY_EMBEDDING_SEARCH, ENTITY_EMBEDDING_SEARCH, + EXPAND_COMMUNITY_STATEMENTS, MEMORY_SUMMARY_EMBEDDING_SEARCH, SEARCH_CHUNK_BY_CHUNK_ID, SEARCH_CHUNKS_BY_CONTENT, + SEARCH_COMMUNITIES_BY_KEYWORD, SEARCH_DIALOGUE_BY_DIALOG_ID, SEARCH_ENTITIES_BY_NAME, SEARCH_MEMORY_SUMMARIES_BY_KEYWORD, @@ -285,6 +288,15 @@ async def search_graph( limit=limit, )) task_keys.append("summaries") + + if "communities" in include: + tasks.append(connector.execute_query( + SEARCH_COMMUNITIES_BY_KEYWORD, + q=q, + end_user_id=end_user_id, + limit=limit, + )) + task_keys.append("communities") # Execute all queries in parallel task_results = await asyncio.gather(*tasks, return_exceptions=True) @@ -396,6 +408,16 @@ async def search_graph_by_embedding( )) task_keys.append("summaries") + # Communities (向量语义匹配) + if "communities" in include: + tasks.append(connector.execute_query( + COMMUNITY_EMBEDDING_SEARCH, + embedding=embedding, + end_user_id=end_user_id, + limit=limit, + )) + task_keys.append("communities") + # Execute all queries in parallel query_start = time.time() task_results = await asyncio.gather(*tasks, return_exceptions=True) @@ -408,6 +430,7 @@ async def search_graph_by_embedding( "chunks": [], "entities": [], "summaries": [], + "communities": [], } for key, result in zip(task_keys, task_results): @@ -661,6 +684,62 @@ async def search_graph_by_chunk_id( return {"chunks": chunks} +async def search_graph_community_expand( + connector: Neo4jConnector, + community_ids: List[str], + end_user_id: str, + limit: int = 10, +) -> Dict[str, List[Dict[str, Any]]]: + """ + 三期：社区展开检索 —— 主题 → 细节两级检索。 + + 命中 Community 节点后，沿 BELONGS_TO_COMMUNITY 关系拉取成员实体， + 再沿 REFERENCES_ENTITY 关系拉取关联的 Statement 节点， + 按 activation_value 降序返回，实现"主题摘要 → 具体记忆"的深度召回。 + + Args: + connector: Neo4j 连接器 + community_ids: 已命中的社区 ID 列表 + end_user_id: 用户 ID，用于数据隔离 + limit: 每个社区最多返回的 Statement 数量 + + Returns: + {"expanded_statements": [Statement 列表，含 community_name / source_entity 字段]} + """ + if not community_ids or not end_user_id: + return {"expanded_statements": []} + + tasks = [ + connector.execute_query( + EXPAND_COMMUNITY_STATEMENTS, + community_id=cid, + end_user_id=end_user_id, + limit=limit, + ) + for cid in community_ids + ] + + task_results = await asyncio.gather(*tasks, return_exceptions=True) + + expanded: List[Dict[str, Any]] = [] + for cid, result in zip(community_ids, task_results): + if isinstance(result, Exception): + logger.warning(f"社区展开检索失败 community_id={cid}: {result}") + else: + expanded.extend(result) + + # 按 activation_value 全局排序后去重 + from app.core.memory.src.search import _deduplicate_results + expanded.sort( + key=lambda x: float(x.get("activation_value") or 0), + reverse=True, + ) + expanded = _deduplicate_results(expanded) + + logger.info(f"社区展开检索完成: community_ids={community_ids}, 展开 statements={len(expanded)}") + return {"expanded_statements": expanded} + + async def search_graph_by_created_at( connector: Neo4jConnector, end_user_id: Optional[str] = None, diff --git a/redbear-mem-benchmark b/redbear-mem-benchmark index c3bbc693..89053e48 160000 --- a/redbear-mem-benchmark +++ b/redbear-mem-benchmark @@ -1 +1 @@ -Subproject commit c3bbc6931c570e6fac88c0b00658b4f08dc2ac77 +Subproject commit 89053e48e932332d2a0f17760034ee2bce75ea43