Merge pull request #3176 from xinnan-tech/perf-tool-add-online-search

Perf：新增联网搜索工具类、修复新闻描述配置获取源问题

main/manager-api/src/main/resources/db/changelog/202605121509.sql

@@ -0,0 +1,6 @@
+-- 新增联网搜索工具配置字段
+SET @data_exists = (SELECT COUNT(*) FROM ai_model_provider WHERE id = 'SYSTEM_PLUGIN_WEB_SEARCH');
+SET @sql = IF(@data_exists = 0,
+    'INSERT INTO `ai_model_provider` (`id`, `model_type`, `provider_code`, `name`, `fields`, `sort`, `creator`, `create_date`, `updater`, `update_date`) VALUES (''SYSTEM_PLUGIN_WEB_SEARCH'', ''Plugin'', ''web_search'', ''联网搜索'', ''[{\"key\": \"provider\", \"type\": \"string\", \"label\": \"搜索源：metaso / tavily\", \"default\": \"metaso\", \"editing\": false, \"selected\": false}, {\"key\": \"description\", \"type\": \"string\", \"label\": \"工具描述\", \"default\": \"联网搜索工具。当用户明确需要联网搜索问题时使用此工具。\", \"editing\": false, \"selected\": false}, {\"key\": \"max_results\", \"type\": \"string\", \"label\": \"返回数量\", \"default\": \"5\", \"editing\": false, \"selected\": false}, {\"key\": \"api_key\", \"type\": \"string\", \"label\": \"apiKey\", \"default\": \"mk-XXXX\", \"editing\": false, \"selected\": false}]'', 80, 1988490863118454785, ''2026-05-11 17:15:35'', 1988490863118454785, ''2026-05-12 10:55:49'')',
+    'SELECT ''data already exists, skip'' AS msg');
+PREPARE stmt FROM @sql; EXECUTE stmt; DEALLOCATE PREPARE stmt;

main/manager-api/src/main/resources/db/changelog/db.changelog-master.yaml

@@ -662,3 +662,10 @@ databaseChangeLog:
         - sqlFile:
             encoding: utf8
             path: classpath:db/changelog/202605081008.sql
+  - changeSet:
+      id: 202605121509
+      author: DaGou12138
+      changes:
+        - sqlFile:
+            encoding: utf8
+            path: classpath:db/changelog/202605121509.sql

main/xiaozhi-server/app.py

@@ -47,13 +47,6 @@ async def main():
     check_ffmpeg_installed()
     config = load_config()
 
-    # 初始化插件描述（需要在load_config之后、服务启动之前）
-    try:
-        from plugins_func.functions.get_news_from_newsnow import init_news_description
-        init_news_description(config)
-    except Exception:
-        pass
-
     # auth_key优先级：配置文件server.auth_key > manager-api.secret > 自动生成
     # auth_key用于jwt认证，比如视觉分析接口的jwt认证、ota接口的token生成与websocket认证
     # 获取配置文件中的auth_key

main/xiaozhi-server/config.yaml

@@ -171,6 +171,19 @@ plugins:
     api_key: "ragflow-xxx"
     # ragflow知识库id
     dataset_ids: ["123456789"]
+  # 联网搜索插件配置
+  # 支持秘塔(metaso)和Tavily两个搜索源
+  # 秘塔API Key申请：https://metaso.cn/search-api/api-keys
+  # Tavily API Key申请：https://app.tavily.com/home
+  web_search:
+    # 搜索源：metaso / tavily
+    provider: "metaso"
+    # 工具描述，用于告知LLM何时调用此工具，可根据需要自定义
+    description: "联网搜索工具。当用户明确需要联网搜索问题时使用此工具。"
+    # 返回结果数量（默认5）
+    max_results: 5
+    # 秘塔API Key（mk-前缀） / Tavily API Key（tvly-前缀）
+    api_key: "mk-xxx"
 # 声纹识别配置
 voiceprint:
   # 声纹接口地址
@@ -253,6 +266,7 @@ Intent:
     # 系统默认已经记载"handle_exit_intent(退出识别)"、"play_music(音乐播放)"插件，请勿重复加载
     # 下面是加载查天气、角色切换、加载查新闻的插件示例
     functions:
+      - web_search
       - get_weather
       - get_news_from_newsnow
       - play_music
@@ -264,6 +278,7 @@ Intent:
     # 下面是加载查天气、角色切换、加载查新闻的插件示例
     functions:
       - change_role
+      - web_search                  # 联网搜索
       - get_weather
       # - search_from_ragflow
       # - get_news_from_chinanews

main/xiaozhi-server/core/providers/tools/server_plugins/plugin_executor.py

@@ -87,6 +87,11 @@ class ServerPluginExecutor(ToolExecutor):
                         func_item.description["function"][
                             "description"
                         ] = fun_description
+
+                # 新闻插件：根据配置更新新闻源参数描述
+                if func_name == "get_news_from_newsnow":
+                    self._init_news_source_description(func_item, func_name)
+
                 tools[func_name] = ToolDefinition(
                     name=func_name,
                     description=func_item.description,
@@ -98,3 +103,20 @@ class ServerPluginExecutor(ToolExecutor):
     def has_tool(self, tool_name: str) -> bool:
         """检查是否有指定的服务端插件工具"""
         return tool_name in all_function_registry
+
+    def _init_news_source_description(self, func_item, func_name):
+        """根据连接配置初始化新闻工具的参数描述"""
+        news_sources = (
+            self.config.get("plugins", {})
+            .get(func_name, {})
+            .get("news_sources", "")
+        )
+        if not news_sources:
+            news_sources = "澎湃新闻;百度热搜;财联社"
+        sources_str = news_sources.replace(";", "、")
+        try:
+            func_item.description["function"]["parameters"]["properties"]["source"][
+                "description"
+            ] = f"新闻源的标准中文名称，例如{sources_str}等。可选参数，如果不提供则使用默认新闻源"
+        except (KeyError, TypeError):
+            pass

main/xiaozhi-server/plugins_func/functions/get_news_from_newsnow.py

@@ -55,15 +55,7 @@ CHANNEL_MAP = {
 DEFAULT_NEWS_SOURCES = "澎湃新闻;百度热搜;财联社"
 
 def _get_newsnow_config(conn):
-    """从连接配置中获取newsnow插件配置，优先用conn.common_config，兜底用conn.config"""
-    # 优先从公共配置获取（保留本地config.yaml的配置）
-    common_plugins = getattr(conn, "common_config", {}).get("plugins", {})
-    common_newsnow = common_plugins.get("get_news_from_newsnow", {})
-    common_sources = common_newsnow.get("news_sources", "")
-    if isinstance(common_sources, str) and common_sources.strip():
-        return common_sources
-
-    # 兜底从连接配置获取
+    # 从连接配置获取
     plugins = conn.config.get("plugins", {})
     newsnow = plugins.get("get_news_from_newsnow", {})
     sources = newsnow.get("news_sources", "")
@@ -91,24 +83,6 @@ def get_news_sources_from_config(conn):
 # 从默认配置获取可用的新闻源名称（运行时由get_news_sources_from_config动态获取）
 example_sources_str = DEFAULT_NEWS_SOURCES.replace(";","、")
 
-def init_news_description(config: dict):
-    """项目启动时调用一次，根据配置更新工具描述中的新闻源示例"""
-    from types import SimpleNamespace
-    from plugins_func.register import all_function_registry
-
-    # 复用get_news_sources_from_config，用SimpleNamespace模拟conn
-    conn_wrapper = SimpleNamespace(config=config)
-    news_sources = get_news_sources_from_config(conn_wrapper)
-
-    sources_str = news_sources.replace(";","、")
-
-    func_item = all_function_registry.get("get_news_from_newsnow")
-    if func_item:
-        func_item.description["function"]["parameters"]["properties"]["source"][
-            "description"
-        ] = f"新闻源的标准中文名称，例如{sources_str}等。可选参数，如果不提供则使用默认新闻源"
-        logger.bind(tag=TAG).info(f"新闻工具描述已初始化，可用新闻源: {sources_str}")
-
 GET_NEWS_FROM_NEWSNOW_FUNCTION_DESC = {
     "type": "function",
     "function": {

main/xiaozhi-server/plugins_func/functions/web_search.py

@@ -0,0 +1,158 @@
+import requests
+from config.logger import setup_logging
+from plugins_func.register import (
+    register_function,
+    ToolType,
+    ActionResponse,
+    Action,
+)
+from typing import TYPE_CHECKING
+
+if TYPE_CHECKING:
+    from core.connection import ConnectionHandler
+
+TAG = __name__
+logger = setup_logging()
+
+_DEFAULT_DESCRIPTION = (
+    "联网搜索工具。当用户明确需要联网搜索问题时使用此工具。"
+)
+
+WEB_SEARCH_FUNCTION_DESC = {
+    "type": "function",
+    "function": {
+        "name": "web_search",
+        "description": _DEFAULT_DESCRIPTION,
+        "parameters": {
+            "type": "object",
+            "properties": {
+                "query": {
+                    "type": "string",
+                    "description": "搜索关键词或问题",
+                }
+            },
+            "required": ["query"],
+        },
+    },
+}
+
+
+def _search_metaso(api_key: str, query: str, max_results: int) -> str:
+    """调用秘塔搜索API"""
+    url = "https://metaso.cn/api/v1/search"
+    headers = {
+        "Authorization": f"Bearer {api_key}",
+        "Content-Type": "application/json",
+    }
+    payload = {
+        "q": query,
+        "size": max_results,
+        "stream": False,
+        "scope": "webpage",
+        "includeSummary": True,
+        "includeRawContent": False,
+        "conciseSnippet": False,
+    }
+    logger.bind(tag=TAG).debug(f"秘塔搜索请求 | URL: {url} | payload: {payload}")
+    response = requests.post(url, json=payload, headers=headers, timeout=15)
+    response.raise_for_status()
+    data = response.json()
+    logger.bind(tag=TAG).debug(f"秘塔搜索响应 | status: {response.status_code}")
+
+    webpages = data.get("webpages", [])
+    if not webpages:
+        return "未找到相关搜索结果。"
+
+    lines = ["【联网搜索结果】"]
+    for i, item in enumerate(webpages, 1):
+        title = item.get("title", "无标题")
+        snippet = item.get("summary", "")
+        date = item.get("date", "")
+        lines.append(f"{i}. 标题：{title}")
+        if date:
+            lines.append(f"   日期：{date}")
+        if snippet:
+            lines.append(f"   摘要：{snippet}")
+
+    return "\n".join(lines)
+
+
+def _search_tavily(api_key: str, query: str, max_results: int) -> str:
+    """调用Tavily搜索API"""
+    url = "https://api.tavily.com/search"
+    headers = {
+        "Authorization": f"Bearer {api_key}",
+        "Content-Type": "application/json",
+    }
+    payload = {
+        "query": query,
+        "max_results": max_results,
+        "search_depth": "advanced",
+        "include_answer": "advanced",
+    }
+    logger.bind(tag=TAG).debug(f"Tavily搜索请求 | URL: {url} | payload: {payload}")
+    response = requests.post(url, json=payload, headers=headers, timeout=15)
+    response.raise_for_status()
+    data = response.json()
+    logger.bind(tag=TAG).debug(f"Tavily搜索响应 | status: {response.status_code} | data: {data}")
+
+    results = data.get("results", [])
+    if not results:
+        return "未找到相关搜索结果。"
+
+    answer = data.get("answer", "")
+    lines = [f"【联网搜索结果】\n总结：{answer}"]
+    # for i, item in enumerate(results, 1):
+    #     title = item.get("title", "无标题")
+    #     summary = item.get("content", "")
+    #     lines.append(f"{i}. 标题：{title}")
+    #     if summary:
+    #         lines.append(f"   摘要：{summary}")
+
+    return "\n".join(lines)
+
+
+@register_function("web_search", WEB_SEARCH_FUNCTION_DESC, ToolType.SYSTEM_CTL)
+def web_search(conn: "ConnectionHandler", query: str = None):
+    logger.bind(tag=TAG).info(f"web_search 被调用 | query={query}")
+    if not query:
+        return ActionResponse(Action.REQLLM, "请提供搜索关键词。", None)
+
+    web_search_config = conn.config.get("plugins", {}).get("web_search", {})
+    provider = web_search_config.get("provider", "").lower()
+    max_results = int(web_search_config.get("max_results", 3))
+    logger.bind(tag=TAG).info(f"web_search 配置 | provider={provider} | max_results={max_results} | config_keys={list(web_search_config.keys())}")
+
+    api_key = web_search_config.get("api_key", "")
+    if not api_key:
+        return ActionResponse(
+            Action.REQLLM,
+            "联网搜索功能未配置API Key，请在配置文件中填写。",
+            None,
+        )
+
+    if provider == "metaso":
+        search_fn = lambda: _search_metaso(api_key, query, max_results)
+    elif provider == "tavily":
+        search_fn = lambda: _search_tavily(api_key, query, max_results)
+    else:
+        return ActionResponse(
+            Action.REQLLM,
+            f"联网搜索功能未配置或配置的搜索源无效（当前：{provider}），请检查配置。",
+            None,
+        )
+
+    try:
+        result_text = search_fn()
+        logger.bind(tag=TAG).info(f"搜索结果组装完成:\n{result_text}")
+    except requests.exceptions.Timeout:
+        logger.bind(tag=TAG).error("联网搜索请求超时")
+        result_text = "联网搜索请求超时，请稍后重试。"
+    except requests.exceptions.RequestException as e:
+        logger.bind(tag=TAG).error(f"联网搜索请求失败: {e}")
+        result_text = "联网搜索请求失败，请稍后重试。"
+    except Exception as e:
+        logger.bind(tag=TAG).error(f"联网搜索异常: {e}")
+        result_text = "联网搜索出现异常，请稍后重试。"
+
+    return ActionResponse(Action.REQLLM, result_text, None)