diff --git a/src/chat/affinity_flow/interest_scoring.py b/src/chat/affinity_flow/interest_scoring.py index 380037880..1e6a4c6a2 100644 --- a/src/chat/affinity_flow/interest_scoring.py +++ b/src/chat/affinity_flow/interest_scoring.py @@ -49,49 +49,36 @@ class InterestScoringSystem: self, messages: List[DatabaseMessages], bot_nickname: str ) -> List[InterestScore]: """计算消息的兴趣度评分""" - logger.info("🚀 开始计算消息兴趣度评分...") - logger.info(f"📨 收到 {len(messages)} 条消息") - - # 通过 user_id 判断是否是用户消息（非机器人发送的消息） + logger.info(f"开始为 {len(messages)} 条消息计算兴趣度...") user_messages = [msg for msg in messages if str(msg.user_info.user_id) != str(global_config.bot.qq_account)] - logger.info(f"👤 过滤出 {len(user_messages)} 条用户消息") + logger.info(f"正在处理 {len(user_messages)} 条用户消息。") scores = [] for i, msg in enumerate(user_messages, 1): - logger.info(f"📋 [{i}/{len(user_messages)}] 处理消息 ID: {msg.message_id}") + logger.debug(f"[{i}/{len(user_messages)}] 处理消息 ID: {msg.message_id}") score = await self._calculate_single_message_score(msg, bot_nickname) scores.append(score) - logger.info(f"✅ 兴趣度评分计算完成，生成 {len(scores)} 个评分") + logger.info(f"兴趣度计算完成，共生成 {len(scores)} 个评分。") return scores async def _calculate_single_message_score(self, message: DatabaseMessages, bot_nickname: str) -> InterestScore: """计算单条消息的兴趣度评分""" - logger.info(f"🎯 计算消息 {message.message_id} 的兴趣度评分...") - logger.debug(f"📝 消息长度: {len(message.processed_plain_text)} 字符") + logger.info(f"计算消息 {message.message_id} 的分数...") + logger.debug(f"消息长度: {len(message.processed_plain_text)} 字符") - # 提取关键词（从数据库的反序列化字段） - logger.debug("🔍 提取关键词...") keywords = self._extract_keywords_from_database(message) - logger.debug(f"🏷️ 提取到 {len(keywords)} 个关键词") + logger.debug(f"提取到 {len(keywords)} 个关键词。") - # 1. 计算兴趣匹配度（现在是异步的） - logger.debug("🧠 计算兴趣匹配度...") interest_match_score = await self._calculate_interest_match_score(message.processed_plain_text, keywords) - logger.debug(f"📊 兴趣匹配度: {interest_match_score:.3f}") + logger.debug(f"兴趣匹配度: {interest_match_score:.3f}") - # 2. 计算关系分 - logger.debug("🤝 计算关系分...") relationship_score = self._calculate_relationship_score(message.user_info.user_id) - logger.debug(f"💝 关系分: {relationship_score:.3f}") + logger.debug(f"关系分数: {relationship_score:.3f}") - # 3. 计算提及分数 - logger.debug("📢 计算提及分数...") mentioned_score = self._calculate_mentioned_score(message, bot_nickname) - logger.debug(f"📣 提及分数: {mentioned_score:.3f}") + logger.debug(f"提及分数: {mentioned_score:.3f}") - # 4. 计算总分 - logger.debug("🧮 计算加权总分...") total_score = ( interest_match_score * self.score_weights["interest_match"] + relationship_score * self.score_weights["relationship"] @@ -99,14 +86,14 @@ class InterestScoringSystem: ) details = { - "interest_match": f"兴趣匹配度: {interest_match_score:.3f}", - "relationship": f"关系分: {relationship_score:.3f}", - "mentioned": f"提及分数: {mentioned_score:.3f}", + "interest_match": f"兴趣匹配: {interest_match_score:.3f}", + "relationship": f"关系: {relationship_score:.3f}", + "mentioned": f"提及: {mentioned_score:.3f}", } - logger.info(f"📈 消息 {message.message_id} 最终评分: {total_score:.3f}") - logger.debug(f"⚖️ 评分权重: {self.score_weights}") - logger.debug(f"📋 评分详情: {details}") + logger.info(f"消息 {message.message_id} 最终得分: {total_score:.3f}") + logger.debug(f"Score weights: {self.score_weights}") + logger.debug(f"Score details: {details}") return InterestScore( message_id=message.message_id, @@ -279,59 +266,41 @@ class InterestScoringSystem: def should_reply(self, score: InterestScore) -> bool: """判断是否应该回复""" - logger.info("🤔 评估是否应该回复...") - logger.debug("📊 评分详情:") - logger.debug(f" 📝 消息ID: {score.message_id}") - logger.debug(f" 💯 总分: {score.total_score:.3f}") - logger.debug(f" 🧠 兴趣匹配: {score.interest_match_score:.3f}") - logger.debug(f" 🤝 关系分: {score.relationship_score:.3f}") - logger.debug(f" 📢 提及分: {score.mentioned_score:.3f}") - + logger.info(f"评估消息 {score.message_id} (得分: {score.total_score:.3f}) 是否回复...") base_threshold = self.reply_threshold - logger.debug(f"📋 基础阈值: {base_threshold:.3f}") # 如果被提及，降低阈值 - if ( - score.mentioned_score >= global_config.affinity_flow.mention_bot_adjustment_threshold - ): # 使用提及bot兴趣分的一半作为判断阈值 + if score.mentioned_score >= global_config.affinity_flow.mention_bot_adjustment_threshold: base_threshold = self.mention_threshold - logger.debug(f"📣 消息提及了机器人，使用降低阈值: {base_threshold:.3f}") + logger.debug(f"机器人被提及, 使用较低阈值: {base_threshold:.3f}") # 计算连续不回复的概率提升 probability_boost = min(self.no_reply_count * self.probability_boost_per_no_reply, 0.8) effective_threshold = base_threshold - probability_boost - - logger.debug("📈 连续不回复统计:") - logger.debug(f" 🚫 不回复次数: {self.no_reply_count}") - logger.debug(f" 📈 概率提升: {probability_boost:.3f}") - logger.debug(f" 🎯 有效阈值: {effective_threshold:.3f}") + logger.debug( + f"基础阈值: {base_threshold:.3f}, 不回复提升: {probability_boost:.3f}, 有效阈值: {effective_threshold:.3f}" + ) # 做出决策 - score.total_score = score.total_score * 1 should_reply = score.total_score >= effective_threshold - decision = "✅ 应该回复" if should_reply else "❌ 不回复" - - logger.info(f"🎯 回复决策: {decision}") - logger.info(f"📊 决策依据: {score.total_score:.3f} {'>=' if should_reply else '<'} {effective_threshold:.3f}") + decision = "✅ 回复" if should_reply else "❌ 不回复" + logger.info(f"回复决策: {decision} (分数: {score.total_score:.3f} {' >=' if should_reply else ' <'} 阈值: {effective_threshold:.3f})") return should_reply, score.total_score def record_reply_action(self, did_reply: bool): """记录回复动作""" old_count = self.no_reply_count - if did_reply: self.no_reply_count = max(0, self.no_reply_count - global_config.affinity_flow.reply_cooldown_reduction) - action = "✅ reply动作可用" + action = "回复" else: self.no_reply_count += 1 - action = "❌ reply动作不可用" + action = "不回复" # 限制最大计数 self.no_reply_count = min(self.no_reply_count, self.max_no_reply_count) - - logger.info(f"📊 记录回复动作: {action}") - logger.info(f"📈 连续不回复次数: {old_count} → {self.no_reply_count}") + logger.info(f"记录动作: {action} | 连续不回复次数: {old_count} -> {self.no_reply_count}") logger.debug(f"📋 最大限制: {self.max_no_reply_count} 次") def update_user_relationship(self, user_id: str, relationship_change: float): @@ -370,23 +339,22 @@ class InterestScoringSystem: async def initialize_smart_interests(self, personality_description: str, personality_id: str = "default"): """初始化智能兴趣系统""" try: - logger.info("🚀 开始初始化智能兴趣系统...") - logger.info(f"📋 人设ID: {personality_id}") - logger.info(f"📝 人设描述长度: {len(personality_description)} 字符") + logger.info("开始初始化智能兴趣系统...") + logger.info(f"人设ID: {personality_id}, 描述长度: {len(personality_description)}") await bot_interest_manager.initialize(personality_description, personality_id) - logger.info("✅ 智能兴趣系统初始化完成") + logger.info("智能兴趣系统初始化完成。") # 显示初始化后的统计信息 stats = bot_interest_manager.get_interest_stats() - logger.info("📊 兴趣系统统计:") - logger.info(f" 🏷️ 总标签数: {stats.get('total_tags', 0)}") - logger.info(f" 💾 缓存大小: {stats.get('cache_size', 0)}") - logger.info(f" 🧠 模型: {stats.get('embedding_model', '未知')}") + logger.info( + f"兴趣系统统计: 总标签={stats.get('total_tags', 0)}, " + f"缓存大小={stats.get('cache_size', 0)}, " + f"模型='{stats.get('embedding_model', '未知')}'" + ) except Exception as e: - logger.error(f"❌ 初始化智能兴趣系统失败: {e}") - logger.error("🔍 错误详情:") + logger.error(f"初始化智能兴趣系统失败: {e}") traceback.print_exc() def get_matching_config(self) -> Dict[str, Any]: diff --git a/src/chat/interest_system/bot_interest_manager.py b/src/chat/interest_system/bot_interest_manager.py index abdc3563d..4e7702487 100644 --- a/src/chat/interest_system/bot_interest_manager.py +++ b/src/chat/interest_system/bot_interest_manager.py @@ -37,22 +37,17 @@ class BotInterestManager: async def initialize(self, personality_description: str, personality_id: str = "default"): """初始化兴趣标签系统""" try: - logger.info("=" * 60) - logger.info("🚀 开始初始化机器人兴趣标签系统") - logger.info(f"📋 人设ID: {personality_id}") - logger.info(f"📝 人设描述长度: {len(personality_description)} 字符") - logger.info("=" * 60) + logger.info("机器人兴趣系统开始初始化...") + logger.info(f"人设ID: {personality_id}, 描述长度: {len(personality_description)}") # 初始化embedding模型 - logger.info("🧠 正在初始化embedding模型...") await self._initialize_embedding_model() # 检查embedding客户端是否成功初始化 if not self.embedding_request: - raise RuntimeError("❌ Embedding客户端初始化失败，无法继续") + raise RuntimeError("Embedding客户端初始化失败") # 生成或加载兴趣标签 - logger.info("🎯 正在生成或加载兴趣标签...") await self._load_or_generate_interests(personality_description, personality_id) self._initialized = True @@ -60,18 +55,13 @@ class BotInterestManager: # 检查是否成功获取兴趣标签 if self.current_interests and len(self.current_interests.get_active_tags()) > 0: active_tags_count = len(self.current_interests.get_active_tags()) - logger.info("=" * 60) - logger.info("✅ 机器人兴趣标签系统初始化完成!") - logger.info(f"📊 活跃兴趣标签数量: {active_tags_count}") - logger.info(f"💾 Embedding缓存大小: {len(self.embedding_cache)}") - logger.info("=" * 60) + logger.info("机器人兴趣系统初始化完成！") + logger.info(f"当前已激活 {active_tags_count} 个兴趣标签, Embedding缓存 {len(self.embedding_cache)} 个") else: - raise RuntimeError("❌ 未能成功生成或加载兴趣标签") + raise RuntimeError("未能成功加载或生成兴趣标签") except Exception as e: - logger.error("=" * 60) - logger.error(f"❌ 初始化机器人兴趣标签系统失败: {e}") - logger.error("=" * 60) + logger.error(f"机器人兴趣系统初始化失败: {e}") traceback.print_exc() raise # 重新抛出异常，不允许降级初始化 @@ -113,19 +103,19 @@ class BotInterestManager: logger.info(f"📚 正在为 '{personality_id}' 加载或生成兴趣标签...") # 首先尝试从数据库加载 - logger.info("💾 尝试从数据库加载现有兴趣标签...") + logger.info("尝试从数据库加载兴趣标签...") loaded_interests = await self._load_interests_from_database(personality_id) if loaded_interests: self.current_interests = loaded_interests active_count = len(loaded_interests.get_active_tags()) - logger.info(f"✅ 成功从数据库加载 {active_count} 个兴趣标签") - logger.info(f"📅 最后更新时间: {loaded_interests.last_updated}") - logger.info(f"🔄 版本号: {loaded_interests.version}") + logger.info(f"成功从数据库加载 {active_count} 个兴趣标签 (版本: {loaded_interests.version})") + tags_info = [f" - '{tag.tag_name}' (权重: {tag.weight:.2f})" for tag in loaded_interests.get_active_tags()] + tags_str = "\n".join(tags_info) + logger.info(f"当前兴趣标签:\n{tags_str}") else: # 生成新的兴趣标签 - logger.info("🆕 数据库中未找到兴趣标签，开始生成新的...") - logger.info("🤖 正在调用LLM生成个性化兴趣标签...") + logger.info("数据库中未找到兴趣标签，开始生成...") generated_interests = await self._generate_interests_from_personality( personality_description, personality_id ) @@ -133,10 +123,13 @@ class BotInterestManager: if generated_interests: self.current_interests = generated_interests active_count = len(generated_interests.get_active_tags()) - logger.info(f"✅ 成功生成 {active_count} 个兴趣标签") + logger.info(f"成功生成 {active_count} 个新兴趣标签。") + tags_info = [f" - '{tag.tag_name}' (权重: {tag.weight:.2f})" for tag in generated_interests.get_active_tags()] + tags_str = "\n".join(tags_info) + logger.info(f"当前兴趣标签:\n{tags_str}") # 保存到数据库 - logger.info("💾 正在保存兴趣标签到数据库...") + logger.info("正在保存至数据库...") await self._save_interests_to_database(generated_interests) else: raise RuntimeError("❌ 兴趣标签生成失败") @@ -411,10 +404,8 @@ class BotInterestManager: if not self.current_interests or not self._initialized: raise RuntimeError("❌ 兴趣标签系统未初始化") - logger.info("🎯 开始计算兴趣匹配度...") - logger.debug(f"💬 消息长度: {len(message_text)} 字符") - if keywords: - logger.debug(f"🏷️ 关键词数量: {len(keywords)}") + logger.info("开始计算兴趣匹配度...") + logger.debug(f"消息长度: {len(message_text)}, 关键词: {len(keywords) if keywords else 0}") message_id = f"msg_{datetime.now().timestamp()}" result = InterestMatchResult(message_id=message_id) @@ -422,14 +413,14 @@ class BotInterestManager: # 获取活跃的兴趣标签 active_tags = self.current_interests.get_active_tags() if not active_tags: - raise RuntimeError("❌ 没有活跃的兴趣标签") + raise RuntimeError("没有检测到活跃的兴趣标签") - logger.info(f"📊 有 {len(active_tags)} 个活跃兴趣标签参与匹配") + logger.info(f"正在与 {len(active_tags)} 个兴趣标签进行匹配...") # 生成消息的embedding - logger.debug("🔄 正在生成消息embedding...") + logger.debug("正在生成消息 embedding...") message_embedding = await self._get_embedding(message_text) - logger.debug(f"✅ 消息embedding生成成功，维度: {len(message_embedding)}") + 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 match_count = 0 @@ -483,10 +474,12 @@ class BotInterestManager: f" 🏷️ '{tag.tag_name}': 相似度={similarity:.3f}, 权重={tag.weight:.2f}, 基础分数={weighted_score:.3f}, 增强分数={enhanced_score:.3f} [低匹配]" ) - logger.info(f"📈 匹配统计: {match_count}/{len(active_tags)} 个标签超过阈值") - logger.info(f"🔥 高相似度匹配(>{high_threshold}): {high_similarity_count} 个") - logger.info(f"⚡ 中相似度匹配(>{medium_threshold}): {medium_similarity_count} 个") - logger.info(f"🌊 低相似度匹配(>{low_threshold}): {low_similarity_count} 个") + logger.info( + f"匹配统计: {match_count}/{len(active_tags)} 个标签命中 | " + f"高(>{high_threshold}): {high_similarity_count}, " + f"中(>{medium_threshold}): {medium_similarity_count}, " + f"低(>{low_threshold}): {low_similarity_count}" + ) # 添加直接关键词匹配奖励 keyword_bonus = self._calculate_keyword_match_bonus(keywords, result.matched_tags) @@ -509,10 +502,10 @@ class BotInterestManager: if result.matched_tags: top_tag_name = max(result.match_scores.items(), key=lambda x: x[1])[0] result.top_tag = top_tag_name - logger.info(f"🏆 最佳匹配标签: '{top_tag_name}' (分数: {result.match_scores[top_tag_name]:.3f})") + logger.info(f"最佳匹配: '{top_tag_name}' (分数: {result.match_scores[top_tag_name]:.3f})") logger.info( - f"📊 最终结果: 总分={result.overall_score:.3f}, 置信度={result.confidence:.3f}, 匹配标签数={len(result.matched_tags)}" + f"最终结果: 总分={result.overall_score:.3f}, 置信度={result.confidence:.3f}, 匹配标签数={len(result.matched_tags)}" ) return result @@ -620,7 +613,7 @@ class BotInterestManager: async def _load_interests_from_database(self, personality_id: str) -> Optional[BotPersonalityInterests]: """从数据库加载兴趣标签""" try: - logger.info(f"💾 正在从数据库加载兴趣标签，personality_id: {personality_id}") + logger.debug(f"从数据库加载兴趣标签, personality_id: {personality_id}") # 导入SQLAlchemy相关模块 from src.common.database.sqlalchemy_models import BotPersonalityInterests as DBBotPersonalityInterests @@ -637,7 +630,7 @@ class BotInterestManager: ) if db_interests: - logger.info(f"✅ 找到数据库中的兴趣标签配置，版本: {db_interests.version}") + logger.debug(f"在数据库中找到兴趣标签配置, 版本: {db_interests.version}") logger.debug(f"📅 最后更新时间: {db_interests.last_updated}") logger.debug(f"🧠 使用的embedding模型: {db_interests.embedding_model}") @@ -671,7 +664,7 @@ class BotInterestManager: ) interests.interest_tags.append(tag) - logger.info(f"✅ 成功从数据库加载 {len(interests.interest_tags)} 个兴趣标签") + logger.debug(f"成功解析 {len(interests.interest_tags)} 个兴趣标签") return interests except (orjson.JSONDecodeError, Exception) as e: diff --git a/src/chat/planner_actions/plan_executor.py b/src/chat/planner_actions/plan_executor.py index e8bf9fdf8..acd9b376f 100644 --- a/src/chat/planner_actions/plan_executor.py +++ b/src/chat/planner_actions/plan_executor.py @@ -96,7 +96,7 @@ class PlanExecutor: self.execution_stats["failed_executions"] += len(execution_results) - successful_count logger.info( - f"动作执行完成: 总数={len(plan.decided_actions)}, 成功={successful_count}, 失败={len(execution_results) - successful_count}" + f"规划执行完成: 总数={len(plan.decided_actions)}, 成功={successful_count}, 失败={len(execution_results) - successful_count}" ) return { @@ -124,7 +124,7 @@ class PlanExecutor: reply_content = "" try: - logger.info(f"执行回复动作: {action_info.action_type}, 原因: {action_info.reasoning}") + logger.info(f"执行回复动作: {action_info.action_type} (原因: {action_info.reasoning})") # 获取用户ID - 兼容对象和字典 if hasattr(action_info.action_message, "user_info"): @@ -156,7 +156,7 @@ class PlanExecutor: ) success = True - logger.info(f"回复动作执行成功: {action_info.action_type}") + logger.info(f"回复动作 '{action_info.action_type}' 执行成功。") except Exception as e: error_message = str(e) @@ -214,7 +214,7 @@ class PlanExecutor: error_message = "" try: - logger.info(f"执行其他动作: {action_info.action_type}, 原因: {action_info.reasoning}") + logger.info(f"执行其他动作: {action_info.action_type} (原因: {action_info.reasoning})") # 构建动作参数 action_params = { @@ -228,7 +228,7 @@ class PlanExecutor: await self.action_manager.execute_action(action_name=action_info.action_type, **action_params) success = True - logger.info(f"其他动作执行成功: {action_info.action_type}") + logger.info(f"其他动作 '{action_info.action_type}' 执行成功。") except Exception as e: error_message = str(e) diff --git a/src/chat/planner_actions/planner.py b/src/chat/planner_actions/planner.py index d648c2292..a423b76ed 100644 --- a/src/chat/planner_actions/planner.py +++ b/src/chat/planner_actions/planner.py @@ -128,7 +128,7 @@ class ActionPlanner: reply_not_available = False if not should_reply and "reply" in initial_plan.available_actions: - logger.info(f"消息兴趣度不足({latest_score.total_score:.2f})，移除reply动作") + logger.info(f"兴趣度不足 ({latest_score.total_score:.2f})，移除'回复'动作。") reply_not_available = True # base_threshold = self.interest_scoring.reply_threshold @@ -136,9 +136,8 @@ class ActionPlanner: non_reply_action_interest_threshold = global_config.affinity_flow.non_reply_action_interest_threshold if score < non_reply_action_interest_threshold: logger.info( - f"❌ 兴趣度不足非回复动作阈值: {score:.3f} < {non_reply_action_interest_threshold:.3f}，直接返回no_action" + f"兴趣度 {score:.3f} 低于非回复动作阈值 {non_reply_action_interest_threshold:.3f}，不执行任何动作。" ) - logger.info(f"📊 最低要求: {non_reply_action_interest_threshold:.3f}") # 直接返回 no_action from src.common.data_models.info_data_model import ActionPlannerInfo