From 96ed5a6789939933c3c7a736a327972bb9c7321f Mon Sep 17 00:00:00 2001 From: Windpicker-owo <3431391539@qq.com> Date: Mon, 8 Dec 2025 15:40:29 +0800 Subject: [PATCH] =?UTF-8?q?feat(embedding):=20=E4=BC=98=E5=8C=96embedding?= =?UTF-8?q?=E7=BC=93=E5=AD=98=E7=AE=A1=E7=90=86=EF=BC=8C=E9=81=BF=E5=85=8D?= =?UTF-8?q?=E6=97=A0=E5=85=B3=E6=A0=87=E7=AD=BE=E7=9A=84=E7=BC=93=E5=AD=98?= =?UTF-8?q?=E5=BD=B1=E5=93=8D=E5=86=85=E5=AD=98=E4=BD=BF=E7=94=A8?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../interest_system/bot_interest_manager.py | 34 ++++++++++++++----- 1 file changed, 25 insertions(+), 9 deletions(-) diff --git a/src/chat/interest_system/bot_interest_manager.py b/src/chat/interest_system/bot_interest_manager.py index 52e71ed84..c8ab8dbb8 100644 --- a/src/chat/interest_system/bot_interest_manager.py +++ b/src/chat/interest_system/bot_interest_manager.py @@ -331,7 +331,12 @@ class BotInterestManager: # 尝试从文件加载缓存 file_cache = await self._load_embedding_cache_from_file(interests.personality_id) if file_cache: - self.embedding_cache.update(file_cache) + allowed_keys = {tag.tag_name for tag in interests.interest_tags} + filtered_cache = {key: value for key, value in file_cache.items() if key in allowed_keys} + dropped_cache = len(file_cache) - len(filtered_cache) + if dropped_cache > 0: + logger.debug(f"🧹 跳过 {dropped_cache} 个与当前兴趣标签无关的缓存embedding") + self.embedding_cache.update(filtered_cache) memory_cached_count = 0 file_cached_count = 0 @@ -371,13 +376,16 @@ class BotInterestManager: interests.last_updated = datetime.now() - async def _get_embedding(self, text: str) -> list[float]: - """获取文本的embedding向量""" + async def _get_embedding(self, text: str, cache: bool = True) -> list[float]: + """获取文本的embedding向量 + + cache=False 用于消息内容，避免在 embedding_cache 中长期保留大文本导致内存膨胀。 + """ if not hasattr(self, "embedding_request"): raise RuntimeError("❌ Embedding请求客户端未初始化") # 检查缓存 - if text in self.embedding_cache: + if cache and text in self.embedding_cache: return self.embedding_cache[text] # 使用LLMRequest获取embedding @@ -389,10 +397,12 @@ class BotInterestManager: if isinstance(embedding[0], list): # If it's a list of lists, take the first one (though get_embedding(str) should return list[float]) embedding = embedding[0] - + # Now we can safely cast to list[float] as we've handled the nested list case embedding_float = cast(list[float], embedding) - self.embedding_cache[text] = embedding_float + + if cache: + self.embedding_cache[text] = embedding_float current_dim = len(embedding_float) if self._detected_embedding_dimension is None: @@ -424,7 +434,7 @@ class BotInterestManager: combined_text = message_text # 生成embedding - embedding = await self._get_embedding(combined_text) + embedding = await self._get_embedding(combined_text, cache=False) return embedding async def generate_embeddings_for_texts( @@ -531,7 +541,8 @@ class BotInterestManager: # 生成消息的embedding logger.debug("正在生成消息 embedding...") if not message_embedding: - message_embedding = await self._get_embedding(message_text) + # 消息文本embedding不入全局缓存，避免缓存随着对话历史无限增长 + message_embedding = await self._get_embedding(message_text, cache=False) logger.debug(f"消息 embedding 生成成功, 维度: {len(message_embedding)}") # 计算与每个兴趣标签的相似度（使用扩展标签） @@ -1104,12 +1115,17 @@ class BotInterestManager: if self.embedding_config and hasattr(self.embedding_config, "model_list") and self.embedding_config.model_list: current_embedding_model = self.embedding_config.model_list[0] + tag_embeddings = self.embedding_cache + if self.current_interests: + allowed_keys = {tag.tag_name for tag in self.current_interests.interest_tags} + tag_embeddings = {key: value for key, value in self.embedding_cache.items() if key in allowed_keys} + cache_data = { "version": 1, "personality_id": personality_id, "embedding_model": current_embedding_model, "last_updated": datetime.now().isoformat(), - "embeddings": self.embedding_cache, + "embeddings": tag_embeddings, "expanded_embeddings": self.expanded_embedding_cache, # 同时保存扩展标签的embedding }