refactor(cache): 重构缓存系统为分层语义缓存

将原有的基于文件的 `ToolCache` 替换为全新的 `CacheManager`，引入了更复杂和高效的分层语义缓存机制。

新系统特性：
- **分层缓存**:
  - L1 缓存: 内存字典 (KV) + FAISS (向量)，用于极速访问。
  - L2 缓存: SQLite (KV) + ChromaDB (向量)，用于持久化存储。
- **语义缓存**: 利用嵌入模型 (Embedding) 对查询进行向量化，实现基于语义相似度的缓存命中，显著提高了缓存命中率。
- **自动失效**: 缓存键包含工具源代码的哈希值，当工具代码更新时，相关缓存会自动失效，避免了脏数据问题。
- **异步支持**: 缓存的 `get` 和 `set` 方法现在是异步的，以适应项目中异步化的工具调用流程。

`web_search_tool` 已更新以使用新的 `CacheManager`，在调用缓存时传递 `tool_class` 和 `semantic_query` 以充分利用新功能。

Co-Authored-By: tt-P607 <68868379+tt-P607@users.noreply.github.com>

src/common/cache_manager.py

@@ -1,344 +1,225 @@
+import time
 import json
+import sqlite3
+import chromadb
 import hashlib
-import re
+import inspect
+import numpy as np
+import faiss
 from typing import Any, Dict, Optional
-from datetime import datetime, timedelta
-from pathlib import Path
-from difflib import SequenceMatcher
-
 from src.common.logger import get_logger
+from src.llm_models.utils_model import LLMRequest
+from src.config.api_ada_configs import ModelTaskConfig
+from src.config.config import global_config, model_config
 
 logger = get_logger("cache_manager")
 
+class CacheManager:
+    """
+    一个支持分层和语义缓存的通用工具缓存管理器。
+    采用单例模式，确保在整个应用中只有一个缓存实例。
+    L1缓存: 内存字典 (KV) + FAISS (Vector)。
+    L2缓存: SQLite (KV) + ChromaDB (Vector)。
+    """
+    _instance = None
 
-class ToolCache:
-    """工具缓存管理器，用于缓存工具调用结果，支持近似匹配"""
+    def __new__(cls, *args, **kwargs):
+        if not cls._instance:
+            cls._instance = super(CacheManager, cls).__new__(cls)
+        return cls._instance
 
-    def __init__(
-        self,
-        cache_dir: str = "data/tool_cache",
-        max_age_hours: int = 24,
-        similarity_threshold: float = 0.65,
-    ):
+    def __init__(self, default_ttl: int = 3600, db_path: str = "data/cache.db", chroma_path: str = "data/chroma_db"):
         """
-        初始化缓存管理器
-
-        Args:
-            cache_dir: 缓存目录路径
-            max_age_hours: 缓存最大存活时间（小时）
-            similarity_threshold: 近似匹配的相似度阈值 (0-1)
+        初始化缓存管理器。
         """
-        self.cache_dir = Path(cache_dir)
-        self.max_age = timedelta(hours=max_age_hours)
-        self.max_age_seconds = max_age_hours * 3600
-        self.similarity_threshold = similarity_threshold
-        self.cache_dir.mkdir(parents=True, exist_ok=True)
+        if not hasattr(self, '_initialized'):
+            self.default_ttl = default_ttl
+            
+            # L1 缓存 (内存)
+            self.l1_kv_cache: Dict[str, Dict[str, Any]] = {}
+            embedding_dim = global_config.lpmm_knowledge.embedding_dimension
+            self.l1_vector_index = faiss.IndexFlatIP(embedding_dim)
+            self.l1_vector_id_to_key: Dict[int, str] = {}
+            
+            # L2 缓存 (持久化)
+            self.db_path = db_path
+            self._init_sqlite()
+            self.chroma_client = chromadb.PersistentClient(path=chroma_path)
+            self.chroma_collection = self.chroma_client.get_or_create_collection(name="semantic_cache")
+            
+            # 嵌入模型
+            self.embedding_model = LLMRequest(model_config.model_task_config.embedding)
 
-    @staticmethod
-    def _normalize_query(query: str) -> str:
+            self._initialized = True
+            logger.info("缓存管理器已初始化: L1 (内存+FAISS), L2 (SQLite+ChromaDB)")
+
+    def _init_sqlite(self):
+        """初始化SQLite数据库和表结构。"""
+        with sqlite3.connect(self.db_path) as conn:
+            cursor = conn.cursor()
+            cursor.execute("""
+                CREATE TABLE IF NOT EXISTS cache (
+                    key TEXT PRIMARY KEY,
+                    value TEXT,
+                    expires_at REAL
+                )
+            """)
+            conn.commit()
+
+    def _generate_key(self, tool_name: str, function_args: Dict[str, Any], tool_class: Any) -> str:
+        """生成确定性的缓存键，包含代码哈希以实现自动失效。"""
+        try:
+            source_code = inspect.getsource(tool_class)
+            code_hash = hashlib.md5(source_code.encode()).hexdigest()
+        except (TypeError, OSError) as e:
+            code_hash = "unknown"
+            logger.warning(f"无法获取 {tool_class.__name__} 的源代码，代码哈希将为 'unknown'。错误: {e}")
+        try:
+            sorted_args = json.dumps(function_args, sort_keys=True)
+        except TypeError:
+            sorted_args = repr(sorted(function_args.items()))
+        return f"{tool_name}::{sorted_args}::{code_hash}"
+
+    async def get(self, tool_name: str, function_args: Dict[str, Any], tool_class: Any, semantic_query: Optional[str] = None) -> Optional[Any]:
         """
-        标准化查询文本，用于相似度比较
-
-        Args:
-            query: 原始查询文本
-
-        Returns:
-            标准化后的查询文本
+        从缓存获取结果，查询顺序: L1-KV -> L1-Vector -> L2-KV -> L2-Vector。
         """
-        if not query:
-            return ""
+        # 步骤 1: L1 精确缓存查询
+        key = self._generate_key(tool_name, function_args, tool_class)
+        logger.debug(f"生成的缓存键: {key}")
+        if semantic_query:
+            logger.debug(f"使用的语义查询: '{semantic_query}'")
 
-        # 纯 Python 实现
-        normalized = query.lower()
-        normalized = re.sub(r"[^\w\s]", " ", normalized)
-        normalized = " ".join(normalized.split())
-        return normalized
+        if key in self.l1_kv_cache:
+            entry = self.l1_kv_cache[key]
+            if time.time() < entry["expires_at"]:
+                logger.info(f"命中L1键值缓存: {key}")
+                return entry["data"]
+            else:
+                del self.l1_kv_cache[key]
 
-    def _calculate_similarity(self, text1: str, text2: str) -> float:
-        """
-        计算两个文本的相似度
+        # 步骤 2: L1/L2 语义和L2精确缓存查询
+        query_embedding = None
+        if semantic_query and self.embedding_model:
+            embedding_result = await self.embedding_model.get_embedding(semantic_query)
+            if embedding_result:
+                query_embedding = np.array([embedding_result], dtype='float32')
 
-        Args:
-            text1: 文本1
-            text2: 文本2
+        # 步骤 2a: L1 语义缓存 (FAISS)
+        if query_embedding is not None and self.l1_vector_index.ntotal > 0:
+            faiss.normalize_L2(query_embedding)
+            distances, indices = self.l1_vector_index.search(query_embedding, 1)
+            if indices.size > 0 and distances[0][0] > 0.75: # IP 越大越相似
+                hit_index = indices[0][0]
+                l1_hit_key = self.l1_vector_id_to_key.get(hit_index)
+                if l1_hit_key and l1_hit_key in self.l1_kv_cache:
+                    logger.info(f"命中L1语义缓存: {l1_hit_key}")
+                    return self.l1_kv_cache[l1_hit_key]["data"]
 
-        Returns:
-            相似度分数 (0-1)
-        """
-        if not text1 or not text2:
-            return 0.0
+        # 步骤 2b: L2 精确缓存 (SQLite)
+        with sqlite3.connect(self.db_path) as conn:
+            cursor = conn.cursor()
+            cursor.execute("SELECT value, expires_at FROM cache WHERE key = ?", (key,))
+            row = cursor.fetchone()
+            if row:
+                value, expires_at = row
+                if time.time() < expires_at:
+                    logger.info(f"命中L2键值缓存: {key}")
+                    data = json.loads(value)
+                    # 回填 L1
+                    self.l1_kv_cache[key] = {"data": data, "expires_at": expires_at}
+                    return data
+                else:
+                    cursor.execute("DELETE FROM cache WHERE key = ?", (key,))
+                    conn.commit()
 
-        # 纯 Python 实现
-        norm_text1 = self._normalize_query(text1)
-        norm_text2 = self._normalize_query(text2)
+        # 步骤 2c: L2 语义缓存 (ChromaDB)
+        if query_embedding is not None:
+            results = self.chroma_collection.query(query_embeddings=query_embedding.tolist(), n_results=1)
+            if results and results['ids'] and results['ids'][0]:
+                distance = results['distances'][0][0] if results['distances'] and results['distances'][0] else 'N/A'
+                logger.debug(f"L2语义搜索找到最相似的结果: id={results['ids'][0]}, 距离={distance}")
+                if distance != 'N/A' and distance < 0.75:
+                    l2_hit_key = results['ids'][0]
+                    logger.info(f"命中L2语义缓存: key='{l2_hit_key}', 距离={distance:.4f}")
+                    with sqlite3.connect(self.db_path) as conn:
+                        cursor = conn.cursor()
+                    cursor.execute("SELECT value, expires_at FROM cache WHERE key = ?", (l2_hit_key if isinstance(l2_hit_key, str) else l2_hit_key[0],))
+                    row = cursor.fetchone()
+                    if row:
+                        value, expires_at = row
+                        if time.time() < expires_at:
+                            data = json.loads(value)
+                            logger.debug(f"L2语义缓存返回的数据: {data}")
+                            # 回填 L1
+                            self.l1_kv_cache[key] = {"data": data, "expires_at": expires_at}
+                            if query_embedding is not None:
+                                new_id = self.l1_vector_index.ntotal
+                                faiss.normalize_L2(query_embedding)
+                                self.l1_vector_index.add(x=query_embedding)
+                                self.l1_vector_id_to_key[new_id] = key
+                            return data
 
-        if norm_text1 == norm_text2:
-            return 1.0
-
-        return SequenceMatcher(None, norm_text1, norm_text2).ratio()
-
-    @staticmethod
-    def _generate_cache_key(tool_name: str, function_args: Dict[str, Any]) -> str:
-        """
-        生成缓存键
-
-        Args:
-            tool_name: 工具名称
-            function_args: 函数参数
-
-        Returns:
-            缓存键字符串
-        """
-        # 将参数排序后序列化，确保相同参数产生相同的键
-        sorted_args = json.dumps(function_args, sort_keys=True, ensure_ascii=False)
-
-        # 纯 Python 实现
-        cache_string = f"{tool_name}:{sorted_args}"
-        return hashlib.md5(cache_string.encode("utf-8")).hexdigest()
-
-    def _get_cache_file_path(self, cache_key: str) -> Path:
-        """获取缓存文件路径"""
-        return self.cache_dir / f"{cache_key}.json"
-
-    def _is_cache_expired(self, cached_time: datetime) -> bool:
-        """检查缓存是否过期"""
-        return datetime.now() - cached_time > self.max_age
-
-    def _find_similar_cache(
-        self, tool_name: str, function_args: Dict[str, Any]
-    ) -> Optional[Dict[str, Any]]:
-        """
-        查找相似的缓存条目
-
-        Args:
-            tool_name: 工具名称
-            function_args: 函数参数
-
-        Returns:
-            相似的缓存结果，如果不存在则返回None
-        """
-        query = function_args.get("query", "")
-        if not query:
-            return None
-
-        candidates = []
-        cache_data_list = []
-
-        # 遍历所有缓存文件，收集候选项
-        for cache_file in self.cache_dir.glob("*.json"):
-            try:
-                with open(cache_file, "r", encoding="utf-8") as f:
-                    cache_data = json.load(f)
-
-                # 检查是否是同一个工具
-                if cache_data.get("tool_name") != tool_name:
-                    continue
-
-                # 检查缓存是否过期
-                cached_time = datetime.fromisoformat(cache_data["timestamp"])
-                if self._is_cache_expired(cached_time):
-                    continue
-
-                # 检查其他参数是否匹配（除了query）
-                cached_args = cache_data.get("function_args", {})
-                args_match = True
-                for key, value in function_args.items():
-                    if key != "query" and cached_args.get(key) != value:
-                        args_match = False
-                        break
-
-                if not args_match:
-                    continue
-
-                # 收集候选项
-                cached_query = cached_args.get("query", "")
-                candidates.append((cached_query, len(cache_data_list)))
-                cache_data_list.append(cache_data)
-
-            except Exception as e:
-                logger.warning(f"检查缓存文件时出错: {cache_file}, 错误: {e}")
-                continue
-
-        if not candidates:
-            logger.debug(
-                f"未找到相似缓存: {tool_name}, 查询: '{query}'，相似度阈值: {self.similarity_threshold}"
-            )
-            return None
-
-        # 纯 Python 实现
-        best_match = None
-        best_similarity = 0.0
-
-        for cached_query, index in candidates:
-            similarity = self._calculate_similarity(query, cached_query)
-            if similarity > best_similarity and similarity >= self.similarity_threshold:
-                best_similarity = similarity
-                best_match = cache_data_list[index]
-
-        if best_match is not None:
-            cached_query = best_match["function_args"].get("query", "")
-            logger.info(
-                f"相似缓存命中，相似度: {best_similarity:.2f}, 原查询: '{cached_query}', 当前查询: '{query}'"
-            )
-            return best_match["result"]
-
-        logger.debug(
-            f"未找到相似缓存: {tool_name}, 查询: '{query}'，相似度阈值: {self.similarity_threshold}"
-        )
+        logger.debug(f"缓存未命中: {key}")
         return None
 
-    def get(
-        self, tool_name: str, function_args: Dict[str, Any]
-    ) -> Optional[Dict[str, Any]]:
-        """
-        从缓存获取结果，支持精确匹配和近似匹配
+    async def set(self, tool_name: str, function_args: Dict[str, Any], tool_class: Any, data: Any, ttl: Optional[int] = None, semantic_query: Optional[str] = None):
+        """将结果存入所有缓存层。"""
+        if ttl is None:
+            ttl = self.default_ttl
+        if ttl <= 0:
+            return
 
-        Args:
-            tool_name: 工具名称
-            function_args: 函数参数
+        key = self._generate_key(tool_name, function_args, tool_class)
+        expires_at = time.time() + ttl
+        
+        # 写入 L1
+        self.l1_kv_cache[key] = {"data": data, "expires_at": expires_at}
 
-        Returns:
-            缓存的结果，如果不存在或已过期则返回None
-        """
-        # 首先尝试精确匹配
-        cache_key = self._generate_cache_key(tool_name, function_args)
-        cache_file = self._get_cache_file_path(cache_key)
+        # 写入 L2
+        value = json.dumps(data)
+        with sqlite3.connect(self.db_path) as conn:
+            cursor = conn.cursor()
+            cursor.execute("REPLACE INTO cache (key, value, expires_at) VALUES (?, ?, ?)", (key, value, expires_at))
+            conn.commit()
 
-        if cache_file.exists():
-            try:
-                with open(cache_file, "r", encoding="utf-8") as f:
-                    cache_data = json.load(f)
+        # 写入语义缓存
+        if semantic_query and self.embedding_model:
+            embedding_result = await self.embedding_model.get_embedding(semantic_query)
+            if embedding_result:
+                embedding = np.array([embedding_result], dtype='float32')
+                # 写入 L1 Vector
+                new_id = self.l1_vector_index.ntotal
+                faiss.normalize_L2(embedding)
+                self.l1_vector_index.add(x=embedding)
+                self.l1_vector_id_to_key[new_id] = key
+                # 写入 L2 Vector
+                self.chroma_collection.add(embeddings=embedding.tolist(), ids=[key])
 
-                # 检查缓存是否过期
-                cached_time = datetime.fromisoformat(cache_data["timestamp"])
-                if self._is_cache_expired(cached_time):
-                    logger.debug(f"缓存已过期: {cache_key}")
-                    cache_file.unlink()  # 删除过期缓存
-                else:
-                    logger.debug(f"精确匹配缓存: {tool_name}")
-                    return cache_data["result"]
+        logger.info(f"已缓存条目: {key}, TTL: {ttl}s")
 
-            except (json.JSONDecodeError, KeyError, ValueError) as e:
-                logger.warning(f"读取缓存文件失败: {cache_file}, 错误: {e}")
-                # 删除损坏的缓存文件
-                if cache_file.exists():
-                    cache_file.unlink()
+    def clear_l1(self):
+        """清空L1缓存。"""
+        self.l1_kv_cache.clear()
+        self.l1_vector_index.reset()
+        self.l1_vector_id_to_key.clear()
+        logger.info("L1 (内存+FAISS) 缓存已清空。")
 
-        # 如果精确匹配失败，尝试近似匹配
-        return self._find_similar_cache(tool_name, function_args)
+    def clear_l2(self):
+        """清空L2缓存。"""
+        with sqlite3.connect(self.db_path) as conn:
+            cursor = conn.cursor()
+            cursor.execute("DELETE FROM cache")
+            conn.commit()
+        self.chroma_client.delete_collection(name="semantic_cache")
+        self.chroma_collection = self.chroma_client.get_or_create_collection(name="semantic_cache")
+        logger.info("L2 (SQLite & ChromaDB) 缓存已清空。")
 
-    def set(
-        self, tool_name: str, function_args: Dict[str, Any], result: Dict[str, Any]
-    ) -> None:
-        """
-        将结果保存到缓存
+    def clear_all(self):
+        """清空所有缓存。"""
+        self.clear_l1()
+        self.clear_l2()
+        logger.info("所有缓存层级已清空。")
 
-        Args:
-            tool_name: 工具名称
-            function_args: 函数参数
-            result: 缓存结果
-        """
-        cache_key = self._generate_cache_key(tool_name, function_args)
-        cache_file = self._get_cache_file_path(cache_key)
-
-        cache_data = {
-            "tool_name": tool_name,
-            "function_args": function_args,
-            "result": result,
-            "timestamp": datetime.now().isoformat(),
-        }
-
-        try:
-            with open(cache_file, "w", encoding="utf-8") as f:
-                json.dump(cache_data, f, ensure_ascii=False, indent=2)
-            logger.debug(f"缓存已保存: {tool_name} -> {cache_key}")
-        except Exception as e:
-            logger.error(f"保存缓存失败: {cache_file}, 错误: {e}")
-
-    def clear_expired(self) -> int:
-        """
-        清理过期缓存
-
-        Returns:
-            删除的文件数量
-        """
-        removed_count = 0
-
-        for cache_file in self.cache_dir.glob("*.json"):
-            try:
-                with open(cache_file, "r", encoding="utf-8") as f:
-                    cache_data = json.load(f)
-
-                cached_time = datetime.fromisoformat(cache_data["timestamp"])
-                if self._is_cache_expired(cached_time):
-                    cache_file.unlink()
-                    removed_count += 1
-                    logger.debug(f"删除过期缓存: {cache_file}")
-
-            except Exception as e:
-                logger.warning(f"清理缓存文件时出错: {cache_file}, 错误: {e}")
-                # 删除损坏的文件
-                try:
-                    cache_file.unlink()
-                    removed_count += 1
-                except (OSError, json.JSONDecodeError, KeyError, ValueError):
-                    logger.warning(f"删除损坏的缓存文件失败: {cache_file}, 错误: {e}")
-
-        logger.info(f"清理完成，删除了 {removed_count} 个过期缓存文件")
-        return removed_count
-
-    def clear_all(self) -> int:
-        """
-        清空所有缓存
-
-        Returns:
-            删除的文件数量
-        """
-        removed_count = 0
-
-        for cache_file in self.cache_dir.glob("*.json"):
-            try:
-                cache_file.unlink()
-                removed_count += 1
-            except Exception as e:
-                logger.warning(f"删除缓存文件失败: {cache_file}, 错误: {e}")
-
-        logger.info(f"清空缓存完成，删除了 {removed_count} 个文件")
-        return removed_count
-
-    def get_stats(self) -> Dict[str, Any]:
-        """
-        获取缓存统计信息
-
-        Returns:
-            缓存统计信息字典
-        """
-        total_files = 0
-        expired_files = 0
-        total_size = 0
-
-        for cache_file in self.cache_dir.glob("*.json"):
-            try:
-                total_files += 1
-                total_size += cache_file.stat().st_size
-
-                with open(cache_file, "r", encoding="utf-8") as f:
-                    cache_data = json.load(f)
-
-                cached_time = datetime.fromisoformat(cache_data["timestamp"])
-                if self._is_cache_expired(cached_time):
-                    expired_files += 1
-
-            except (OSError, json.JSONDecodeError, KeyError, ValueError):
-                expired_files += 1  # 损坏的文件也算作过期
-
-        return {
-            "total_files": total_files,
-            "expired_files": expired_files,
-            "total_size_bytes": total_size,
-            "cache_dir": str(self.cache_dir),
-            "max_age_hours": self.max_age.total_seconds() / 3600,
-            "similarity_threshold": self.similarity_threshold,
-        }
-
-tool_cache = ToolCache()
+# 全局实例
+tool_cache = CacheManager()

src/common/cache_manager_backup.py

@@ -0,0 +1,344 @@
+import json
+import hashlib
+import re
+from typing import Any, Dict, Optional
+from datetime import datetime, timedelta
+from pathlib import Path
+from difflib import SequenceMatcher
+
+from src.common.logger import get_logger
+
+logger = get_logger("cache_manager")
+
+
+class ToolCache:
+    """工具缓存管理器，用于缓存工具调用结果，支持近似匹配"""
+
+    def __init__(
+        self,
+        cache_dir: str = "data/tool_cache",
+        max_age_hours: int = 24,
+        similarity_threshold: float = 0.65,
+    ):
+        """
+        初始化缓存管理器
+
+        Args:
+            cache_dir: 缓存目录路径
+            max_age_hours: 缓存最大存活时间（小时）
+            similarity_threshold: 近似匹配的相似度阈值 (0-1)
+        """
+        self.cache_dir = Path(cache_dir)
+        self.max_age = timedelta(hours=max_age_hours)
+        self.max_age_seconds = max_age_hours * 3600
+        self.similarity_threshold = similarity_threshold
+        self.cache_dir.mkdir(parents=True, exist_ok=True)
+
+    @staticmethod
+    def _normalize_query(query: str) -> str:
+        """
+        标准化查询文本，用于相似度比较
+
+        Args:
+            query: 原始查询文本
+
+        Returns:
+            标准化后的查询文本
+        """
+        if not query:
+            return ""
+
+        # 纯 Python 实现
+        normalized = query.lower()
+        normalized = re.sub(r"[^\w\s]", " ", normalized)
+        normalized = " ".join(normalized.split())
+        return normalized
+
+    def _calculate_similarity(self, text1: str, text2: str) -> float:
+        """
+        计算两个文本的相似度
+
+        Args:
+            text1: 文本1
+            text2: 文本2
+
+        Returns:
+            相似度分数 (0-1)
+        """
+        if not text1 or not text2:
+            return 0.0
+
+        # 纯 Python 实现
+        norm_text1 = self._normalize_query(text1)
+        norm_text2 = self._normalize_query(text2)
+
+        if norm_text1 == norm_text2:
+            return 1.0
+
+        return SequenceMatcher(None, norm_text1, norm_text2).ratio()
+
+    @staticmethod
+    def _generate_cache_key(tool_name: str, function_args: Dict[str, Any]) -> str:
+        """
+        生成缓存键
+
+        Args:
+            tool_name: 工具名称
+            function_args: 函数参数
+
+        Returns:
+            缓存键字符串
+        """
+        # 将参数排序后序列化，确保相同参数产生相同的键
+        sorted_args = json.dumps(function_args, sort_keys=True, ensure_ascii=False)
+
+        # 纯 Python 实现
+        cache_string = f"{tool_name}:{sorted_args}"
+        return hashlib.md5(cache_string.encode("utf-8")).hexdigest()
+
+    def _get_cache_file_path(self, cache_key: str) -> Path:
+        """获取缓存文件路径"""
+        return self.cache_dir / f"{cache_key}.json"
+
+    def _is_cache_expired(self, cached_time: datetime) -> bool:
+        """检查缓存是否过期"""
+        return datetime.now() - cached_time > self.max_age
+
+    def _find_similar_cache(
+        self, tool_name: str, function_args: Dict[str, Any]
+    ) -> Optional[Dict[str, Any]]:
+        """
+        查找相似的缓存条目
+
+        Args:
+            tool_name: 工具名称
+            function_args: 函数参数
+
+        Returns:
+            相似的缓存结果，如果不存在则返回None
+        """
+        query = function_args.get("query", "")
+        if not query:
+            return None
+
+        candidates = []
+        cache_data_list = []
+
+        # 遍历所有缓存文件，收集候选项
+        for cache_file in self.cache_dir.glob("*.json"):
+            try:
+                with open(cache_file, "r", encoding="utf-8") as f:
+                    cache_data = json.load(f)
+
+                # 检查是否是同一个工具
+                if cache_data.get("tool_name") != tool_name:
+                    continue
+
+                # 检查缓存是否过期
+                cached_time = datetime.fromisoformat(cache_data["timestamp"])
+                if self._is_cache_expired(cached_time):
+                    continue
+
+                # 检查其他参数是否匹配（除了query）
+                cached_args = cache_data.get("function_args", {})
+                args_match = True
+                for key, value in function_args.items():
+                    if key != "query" and cached_args.get(key) != value:
+                        args_match = False
+                        break
+
+                if not args_match:
+                    continue
+
+                # 收集候选项
+                cached_query = cached_args.get("query", "")
+                candidates.append((cached_query, len(cache_data_list)))
+                cache_data_list.append(cache_data)
+
+            except Exception as e:
+                logger.warning(f"检查缓存文件时出错: {cache_file}, 错误: {e}")
+                continue
+
+        if not candidates:
+            logger.debug(
+                f"未找到相似缓存: {tool_name}, 查询: '{query}'，相似度阈值: {self.similarity_threshold}"
+            )
+            return None
+
+        # 纯 Python 实现
+        best_match = None
+        best_similarity = 0.0
+
+        for cached_query, index in candidates:
+            similarity = self._calculate_similarity(query, cached_query)
+            if similarity > best_similarity and similarity >= self.similarity_threshold:
+                best_similarity = similarity
+                best_match = cache_data_list[index]
+
+        if best_match is not None:
+            cached_query = best_match["function_args"].get("query", "")
+            logger.info(
+                f"相似缓存命中，相似度: {best_similarity:.2f}, 原查询: '{cached_query}', 当前查询: '{query}'"
+            )
+            return best_match["result"]
+
+        logger.debug(
+            f"未找到相似缓存: {tool_name}, 查询: '{query}'，相似度阈值: {self.similarity_threshold}"
+        )
+        return None
+
+    def get(
+        self, tool_name: str, function_args: Dict[str, Any]
+    ) -> Optional[Dict[str, Any]]:
+        """
+        从缓存获取结果，支持精确匹配和近似匹配
+
+        Args:
+            tool_name: 工具名称
+            function_args: 函数参数
+
+        Returns:
+            缓存的结果，如果不存在或已过期则返回None
+        """
+        # 首先尝试精确匹配
+        cache_key = self._generate_cache_key(tool_name, function_args)
+        cache_file = self._get_cache_file_path(cache_key)
+
+        if cache_file.exists():
+            try:
+                with open(cache_file, "r", encoding="utf-8") as f:
+                    cache_data = json.load(f)
+
+                # 检查缓存是否过期
+                cached_time = datetime.fromisoformat(cache_data["timestamp"])
+                if self._is_cache_expired(cached_time):
+                    logger.debug(f"缓存已过期: {cache_key}")
+                    cache_file.unlink()  # 删除过期缓存
+                else:
+                    logger.debug(f"精确匹配缓存: {tool_name}")
+                    return cache_data["result"]
+
+            except (json.JSONDecodeError, KeyError, ValueError) as e:
+                logger.warning(f"读取缓存文件失败: {cache_file}, 错误: {e}")
+                # 删除损坏的缓存文件
+                if cache_file.exists():
+                    cache_file.unlink()
+
+        # 如果精确匹配失败，尝试近似匹配
+        return self._find_similar_cache(tool_name, function_args)
+
+    def set(
+        self, tool_name: str, function_args: Dict[str, Any], result: Dict[str, Any]
+    ) -> None:
+        """
+        将结果保存到缓存
+
+        Args:
+            tool_name: 工具名称
+            function_args: 函数参数
+            result: 缓存结果
+        """
+        cache_key = self._generate_cache_key(tool_name, function_args)
+        cache_file = self._get_cache_file_path(cache_key)
+
+        cache_data = {
+            "tool_name": tool_name,
+            "function_args": function_args,
+            "result": result,
+            "timestamp": datetime.now().isoformat(),
+        }
+
+        try:
+            with open(cache_file, "w", encoding="utf-8") as f:
+                json.dump(cache_data, f, ensure_ascii=False, indent=2)
+            logger.debug(f"缓存已保存: {tool_name} -> {cache_key}")
+        except Exception as e:
+            logger.error(f"保存缓存失败: {cache_file}, 错误: {e}")
+
+    def clear_expired(self) -> int:
+        """
+        清理过期缓存
+
+        Returns:
+            删除的文件数量
+        """
+        removed_count = 0
+
+        for cache_file in self.cache_dir.glob("*.json"):
+            try:
+                with open(cache_file, "r", encoding="utf-8") as f:
+                    cache_data = json.load(f)
+
+                cached_time = datetime.fromisoformat(cache_data["timestamp"])
+                if self._is_cache_expired(cached_time):
+                    cache_file.unlink()
+                    removed_count += 1
+                    logger.debug(f"删除过期缓存: {cache_file}")
+
+            except Exception as e:
+                logger.warning(f"清理缓存文件时出错: {cache_file}, 错误: {e}")
+                # 删除损坏的文件
+                try:
+                    cache_file.unlink()
+                    removed_count += 1
+                except (OSError, json.JSONDecodeError, KeyError, ValueError):
+                    logger.warning(f"删除损坏的缓存文件失败: {cache_file}, 错误: {e}")
+
+        logger.info(f"清理完成，删除了 {removed_count} 个过期缓存文件")
+        return removed_count
+
+    def clear_all(self) -> int:
+        """
+        清空所有缓存
+
+        Returns:
+            删除的文件数量
+        """
+        removed_count = 0
+
+        for cache_file in self.cache_dir.glob("*.json"):
+            try:
+                cache_file.unlink()
+                removed_count += 1
+            except Exception as e:
+                logger.warning(f"删除缓存文件失败: {cache_file}, 错误: {e}")
+
+        logger.info(f"清空缓存完成，删除了 {removed_count} 个文件")
+        return removed_count
+
+    def get_stats(self) -> Dict[str, Any]:
+        """
+        获取缓存统计信息
+
+        Returns:
+            缓存统计信息字典
+        """
+        total_files = 0
+        expired_files = 0
+        total_size = 0
+
+        for cache_file in self.cache_dir.glob("*.json"):
+            try:
+                total_files += 1
+                total_size += cache_file.stat().st_size
+
+                with open(cache_file, "r", encoding="utf-8") as f:
+                    cache_data = json.load(f)
+
+                cached_time = datetime.fromisoformat(cache_data["timestamp"])
+                if self._is_cache_expired(cached_time):
+                    expired_files += 1
+
+            except (OSError, json.JSONDecodeError, KeyError, ValueError):
+                expired_files += 1  # 损坏的文件也算作过期
+
+        return {
+            "total_files": total_files,
+            "expired_files": expired_files,
+            "total_size_bytes": total_size,
+            "cache_dir": str(self.cache_dir),
+            "max_age_hours": self.max_age.total_seconds() / 3600,
+            "similarity_threshold": self.similarity_threshold,
+        }
+
+tool_cache = ToolCache()

src/main.py

@@ -17,6 +17,7 @@ from src.common.server import get_global_server, Server
 from src.mood.mood_manager import mood_manager
 from rich.traceback import install
 from src.manager.schedule_manager import schedule_manager
+from src.common.cache_manager import tool_cache
 # from src.api.main import start_api_server
 
 # 导入新的插件管理器和热重载管理器

src/plugins/built_in/web_search_tool/plugin.py

@@ -88,7 +88,8 @@ class WebSurfingTool(BaseTool):
             return {"error": "搜索查询不能为空。"}
 
         # 检查缓存
-        cached_result = tool_cache.get(self.name, function_args)
+        query = function_args.get("query")
+        cached_result = await tool_cache.get(self.name, function_args, tool_class=self.__class__, semantic_query=query)
         if cached_result:
             logger.info(f"缓存命中: {self.name} -> {function_args}")
             return cached_result
@@ -109,7 +110,8 @@ class WebSurfingTool(BaseTool):
         
         # 保存到缓存
         if "error" not in result:
-            tool_cache.set(self.name, function_args, result)
+            query = function_args.get("query")
+            await tool_cache.set(self.name, function_args, self.__class__, result, semantic_query=query)
             
         return result
 
@@ -463,7 +465,7 @@ class URLParserTool(BaseTool):
         执行URL内容提取和总结。优先使用Exa，失败后尝试本地解析。
         """
         # 检查缓存
-        cached_result = tool_cache.get(self.name, function_args)
+        cached_result = await tool_cache.get(self.name, function_args, tool_class=self.__class__)
         if cached_result:
             logger.info(f"缓存命中: {self.name} -> {function_args}")
             return cached_result
@@ -577,7 +579,7 @@ class URLParserTool(BaseTool):
         
         # 保存到缓存
         if "error" not in result:
-            tool_cache.set(self.name, function_args, result)
+            await tool_cache.set(self.name, function_args, self.__class__, result)
 
         return result