From 453e31765c66a85e9df7d9986db8224ee178871d Mon Sep 17 00:00:00 2001 From: minecraft1024a Date: Sat, 25 Oct 2025 20:13:32 +0800 Subject: [PATCH] =?UTF-8?q?refactor(memory):=20=E5=BC=95=E5=85=A5=20Memory?= =?UTF-8?q?Manager=20=E7=BB=9F=E4=B8=80=E7=AE=A1=E7=90=86=E7=9E=AC?= =?UTF-8?q?=E6=97=B6=E8=AE=B0=E5=BF=86?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 引入 `MemoryManager` 类来封装和统一处理瞬时记忆（消息集合）的检索和管理逻辑。此举将瞬时记忆的相关操作从 `MemorySystem` 中解耦，提高了代码的模块化和可维护性。 主要变更： - 创建 `MemoryManager` 类，负责消息集合的初始化、上下文检索等。 - `MemorySystem` 现在通过 `MemoryManager` 实例来获取瞬时记忆，简化了其内部实现。 - 移除了 `MemorySystem` 中原有的、分散的瞬时记忆检索代码，使其职责更单一。 --- src/chat/memory_system/memory_system.py | 83 +++++++++---------- .../message_collection_storage.py | 8 +- 2 files changed, 41 insertions(+), 50 deletions(-) diff --git a/src/chat/memory_system/memory_system.py b/src/chat/memory_system/memory_system.py index 4c52c26e8..f43aa5fb9 100644 --- a/src/chat/memory_system/memory_system.py +++ b/src/chat/memory_system/memory_system.py @@ -18,6 +18,7 @@ import orjson from src.chat.memory_system.memory_builder import MemoryBuilder, MemoryExtractionError from src.chat.memory_system.memory_chunk import MemoryChunk from src.chat.memory_system.memory_fusion import MemoryFusionEngine +from src.chat.memory_system.memory_manager import MemoryManager from src.chat.memory_system.memory_query_planner import MemoryQueryPlanner from src.chat.memory_system.message_collection_storage import MessageCollectionStorage @@ -146,6 +147,7 @@ class MemorySystem: self.message_collection_storage: MessageCollectionStorage | None = None self.query_planner: MemoryQueryPlanner | None = None self.forgetting_engine: MemoryForgettingEngine | None = None + self.memory_manager: MemoryManager | None = None # LLM模型 self.value_assessment_model: LLMRequest | None = None @@ -171,6 +173,10 @@ class MemorySystem: async def initialize(self): """异步初始化记忆系统""" try: + # 初始化 MemoryManager + self.memory_manager = MemoryManager() + await self.memory_manager.initialize() + # 初始化LLM模型 fallback_task = getattr(self.llm_model, "model_for_task", None) if self.llm_model else None @@ -936,58 +942,43 @@ class MemorySystem: async def _retrieve_instant_memories(self, query_text: str, chat_id: str | None) -> list[MemoryChunk]: """检索瞬时记忆（消息集合）并转换为MemoryChunk""" - if not self.message_collection_storage: + if not self.memory_manager: return [] - # 1. 优先检索当前聊天的消息集合 - collections = [] - if chat_id: - collections = await self.memory_manger.get_message_collection_context(query_text, chat_id=chat_id, n_results=1) - - # 2. 如果当前聊天没有，或者不需要区分聊天，则进行全局检索 - if not collections: - collections = await self.message_collection_storage.get_relevant_collection(query_text, chat_id=None, n_results=1) - - if not collections: + if not chat_id: return [] - # 3. 将 MessageCollection 转换为 MemoryChunk - instant_memories = [] - for collection in collections: - from src.chat.memory_system.memory_chunk import ( - ContentStructure, - ImportanceLevel, - MemoryMetadata, - MemoryType, - ) - - header = f"[来自群/聊 {collection.chat_id} 的近期对话]" - if collection.chat_id == chat_id: - header = f"[🔥 来自当前聊天的近期对话]" + context_text = await self.memory_manager.get_message_collection_context(query_text, chat_id=chat_id) + if not context_text: + return [] - display_text = f"{header}\n---\n{collection.combined_text}" + from src.chat.memory_system.memory_chunk import ( + ContentStructure, + ImportanceLevel, + MemoryMetadata, + MemoryType, + ) - metadata = MemoryMetadata( - memory_id=f"instant_{collection.collection_id}", - user_id=GLOBAL_MEMORY_SCOPE, - chat_id=collection.chat_id, - created_at=collection.created_at, - importance=ImportanceLevel.HIGH, # 瞬时记忆通常具有高重要性 - ) - content = ContentStructure( - subject="对话上下文", - predicate="包含", - object=collection.combined_text, - display=display_text - ) - chunk = MemoryChunk( - metadata=metadata, - content=content, - memory_type=MemoryType.CONTEXTUAL, - ) - instant_memories.append(chunk) - - return instant_memories + metadata = MemoryMetadata( + memory_id=f"instant_{chat_id}_{time.time()}", + user_id=GLOBAL_MEMORY_SCOPE, + chat_id=chat_id, + created_at=time.time(), + importance=ImportanceLevel.HIGH, # 瞬时记忆通常具有高重要性 + ) + content = ContentStructure( + subject="近期对话上下文", + predicate="相关内容", + object=context_text, + display=context_text, + ) + chunk = MemoryChunk( + metadata=metadata, + content=content, + memory_type=MemoryType.CONTEXTUAL, + ) + + return [chunk] @staticmethod def _extract_json_payload(response: str) -> str | None: diff --git a/src/chat/memory_system/message_collection_storage.py b/src/chat/memory_system/message_collection_storage.py index 92c42dcc7..d55d3ea92 100644 --- a/src/chat/memory_system/message_collection_storage.py +++ b/src/chat/memory_system/message_collection_storage.py @@ -81,22 +81,22 @@ class MessageCollectionStorage: current_count = self.vector_db_service.count(self.collection_name) if current_count > self.config.instant_memory_max_collections: num_to_delete = current_count - self.config.instant_memory_max_collections - + # 获取所有文档的元数据以进行排序 all_docs = self.vector_db_service.get( collection_name=self.collection_name, include=["metadatas"] ) - + if all_docs and all_docs.get("ids"): # 在内存中排序找到最旧的文档 sorted_docs = sorted( zip(all_docs["ids"], all_docs["metadatas"]), key=lambda item: item[1].get("created_at", 0), ) - + ids_to_delete = [doc[0] for doc in sorted_docs[:num_to_delete]] - + if ids_to_delete: self.vector_db_service.delete(collection_name=self.collection_name, ids=ids_to_delete) logger.info(f"消息集合已满，删除最旧的 {len(ids_to_delete)} 个集合")