diff --git a/main/xiaozhi-server/config.yaml b/main/xiaozhi-server/config.yaml index c4662afa..ce6e6780 100644 --- a/main/xiaozhi-server/config.yaml +++ b/main/xiaozhi-server/config.yaml @@ -61,6 +61,8 @@ delete_audio: true close_connection_no_voice_time: 120 # TTS请求超时时间(秒) tts_timeout: 10 +# 工具调用超时时间(秒) +tool_call_timeout: 30 # 开启唤醒词加速 enable_wakeup_words_response_cache: true # 开场是否回复唤醒词 diff --git a/main/xiaozhi-server/core/connection.py b/main/xiaozhi-server/core/connection.py index b707e82a..49b51e75 100644 --- a/main/xiaozhi-server/core/connection.py +++ b/main/xiaozhi-server/core/connection.py @@ -32,7 +32,7 @@ from core.providers.asr.dto.dto import InterfaceType from core.handle.textHandle import handleTextMessage from core.providers.tools.unified_tool_handler import UnifiedToolHandler from plugins_func.loadplugins import auto_import_modules -from plugins_func.register import Action +from plugins_func.register import Action, ActionResponse from core.auth import AuthenticationError from config.config_loader import get_private_config_from_api from core.providers.tts.dto.dto import ContentType, TTSMessageDTO, SentenceType @@ -148,6 +148,8 @@ class ConnectionHandler: self.memory = _memory self.intent = _intent + self.is_exiting = False # 标记是否正在执行退出流程 + # 为每个连接单独管理声纹识别 self.voiceprint_provider = None @@ -327,6 +329,10 @@ class ConnectionHandler: async def _route_message(self, message): """消息路由""" + # 退出状态丢弃所有消息 + if self.is_exiting: + return + # 检查是否已经获取到真实的绑定状态 if not self.bind_completed_event.is_set(): # 还没有获取到真实状态，等待直到获取到真实状态或超时 @@ -1102,11 +1108,24 @@ class ConnectionHandler: ) futures_with_data.append((future, tool_call_data, tool_input)) + # 工具调用超时时间，可配置，默认30秒 + tool_call_timeout = int(self.config.get("tool_call_timeout", 30)) # 等待协程结束（实际等待时长为最慢的那个） tool_results = [] + for future, tool_call_data, tool_input in futures_with_data: - result = future.result() - tool_results.append((result, tool_call_data)) + try: + result = future.result(timeout=tool_call_timeout) + tool_results.append((result, tool_call_data)) + except Exception as e: + self.logger.bind(tag=TAG).error( + f"工具调用超时或异常: {tool_call_data['name']}, 错误: {e}" + ) + # 超时时返回错误响应，避免整个流程卡死 + tool_results.append(( + ActionResponse(action=Action.ERROR, result="哎呀，网络遇到点问题，请稍后再试下！"), + tool_call_data + )) # 使用公共方法上报工具调用结果 enqueue_tool_report(self, tool_call_data['name'], tool_input, str(result.result) if result.result else None, report_tool_call=False) diff --git a/main/xiaozhi-server/core/handle/abortHandle.py b/main/xiaozhi-server/core/handle/abortHandle.py index b7bf1611..b74b2893 100644 --- a/main/xiaozhi-server/core/handle/abortHandle.py +++ b/main/xiaozhi-server/core/handle/abortHandle.py @@ -7,6 +7,10 @@ TAG = __name__ async def handleAbortMessage(conn: "ConnectionHandler"): + if conn.close_after_chat or conn.is_exiting: + conn.logger.bind(tag=TAG).info("退出流程中被打断，直接关闭连接") + return + conn.logger.bind(tag=TAG).info("Abort message received") # 设置成打断状态，会自动打断llm、tts任务 conn.client_abort = True diff --git a/main/xiaozhi-server/core/handle/intentHandler.py b/main/xiaozhi-server/core/handle/intentHandler.py index f092dfb2..0c868758 100644 --- a/main/xiaozhi-server/core/handle/intentHandler.py +++ b/main/xiaozhi-server/core/handle/intentHandler.py @@ -33,6 +33,10 @@ async def handle_user_intent(conn: "ConnectionHandler", text): if await check_direct_exit(conn, filtered_text): return True + # 明确再见不被打断 + if conn.is_exiting: + return True + # 检查是否是唤醒词 if await checkWakeupWords(conn, filtered_text): return True @@ -58,6 +62,7 @@ async def check_direct_exit(conn: "ConnectionHandler", text): if text == cmd: conn.logger.bind(tag=TAG).info(f"识别到明确的退出命令: {text}") await send_stt_message(conn, text) + conn.is_exiting = True await conn.close() return True return False @@ -156,7 +161,9 @@ async def process_intent_result( # 使用executor执行函数调用和结果处理 def process_function_call(): conn.dialogue.put(Message(role="user", content=original_text)) - + + # 工具调用超时时间 + tool_call_timeout = int(conn.config.get("tool_call_timeout", 30)) # 使用统一工具处理器处理所有工具调用 try: result = asyncio.run_coroutine_threadsafe( @@ -164,11 +171,11 @@ async def process_intent_result( conn, function_call_data ), conn.loop, - ).result() + ).result(timeout=tool_call_timeout) except Exception as e: conn.logger.bind(tag=TAG).error(f"工具调用失败: {e}") result = ActionResponse( - action=Action.ERROR, result=str(e), response=str(e) + action=Action.ERROR, result="工具调用超时，请一会再试下哈", response="工具调用超时，请一会再试下哈" ) # 上报工具调用结果 diff --git a/main/xiaozhi-server/core/handle/receiveAudioHandle.py b/main/xiaozhi-server/core/handle/receiveAudioHandle.py index 4f69aa99..280d72cf 100644 --- a/main/xiaozhi-server/core/handle/receiveAudioHandle.py +++ b/main/xiaozhi-server/core/handle/receiveAudioHandle.py @@ -15,6 +15,8 @@ TAG = __name__ async def handleAudioMessage(conn: "ConnectionHandler", audio): + if conn.is_exiting: + return # 当前片段是否有人说话 have_voice = conn.vad.is_vad(conn, audio) # 如果设备刚刚被唤醒，短暂忽略VAD检测 diff --git a/main/xiaozhi-server/core/providers/llm/openai/openai.py b/main/xiaozhi-server/core/providers/llm/openai/openai.py index 58ec725c..3d9882f3 100644 --- a/main/xiaozhi-server/core/providers/llm/openai/openai.py +++ b/main/xiaozhi-server/core/providers/llm/openai/openai.py @@ -17,8 +17,22 @@ class LLMProvider(LLMProviderBase): self.base_url = config.get("base_url") else: self.base_url = config.get("url") - timeout = config.get("timeout", 300) - self.timeout = int(timeout) if timeout else 300 + + timeout_config = config.get("timeout") + if isinstance(timeout_config, dict): + # 细粒度超时配置 + custom_timeout = httpx.Timeout( + pool=timeout_config.get("pool", 2.0), + connect=timeout_config.get("connect", 3.0), + write=timeout_config.get("write", 5.0), + read=timeout_config.get("read", 60.0) + ) + elif isinstance(timeout_config, (int, float)) and timeout_config > 0: + # 兼容旧的单一超时配置（整数或浮点数） + custom_timeout = httpx.Timeout(timeout_config) + else: + # 未配置或配置无效，使用默认值 + custom_timeout = httpx.Timeout(300) param_defaults = { "max_tokens": int, @@ -45,7 +59,7 @@ class LLMProvider(LLMProviderBase): model_key_msg = check_model_key("LLM", self.api_key) if model_key_msg: logger.bind(tag=TAG).error(model_key_msg) - self.client = openai.OpenAI(api_key=self.api_key, base_url=self.base_url, timeout=httpx.Timeout(self.timeout)) + self.client = openai.OpenAI(api_key=self.api_key, base_url=self.base_url, timeout=custom_timeout) @staticmethod def normalize_dialogue(dialogue): diff --git a/main/xiaozhi-server/core/providers/tts/alibl_stream.py b/main/xiaozhi-server/core/providers/tts/alibl_stream.py index 1a639651..5fb860f2 100644 --- a/main/xiaozhi-server/core/providers/tts/alibl_stream.py +++ b/main/xiaozhi-server/core/providers/tts/alibl_stream.py @@ -128,7 +128,7 @@ class TTSProvider(TTSProviderBase): self.start_session(self.conn.sentence_id), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) self.before_stop_play_files.clear() logger.bind(tag=TAG).info("TTS会话启动成功") except Exception as e: @@ -145,7 +145,7 @@ class TTSProvider(TTSProviderBase): self.text_to_speak(message.content_detail, None), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) logger.bind(tag=TAG).debug("TTS文本发送成功") except Exception as e: logger.bind(tag=TAG).error(f"发送TTS文本失败: {str(e)}") @@ -166,7 +166,7 @@ class TTSProvider(TTSProviderBase): self.finish_session(self.conn.sentence_id), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) except Exception as e: logger.bind(tag=TAG).error(f"结束TTS会话失败: {str(e)}") continue diff --git a/main/xiaozhi-server/core/providers/tts/aliyun_stream.py b/main/xiaozhi-server/core/providers/tts/aliyun_stream.py index 9548595e..f773b710 100644 --- a/main/xiaozhi-server/core/providers/tts/aliyun_stream.py +++ b/main/xiaozhi-server/core/providers/tts/aliyun_stream.py @@ -233,7 +233,7 @@ class TTSProvider(TTSProviderBase): self.start_session(self.task_id), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) self.before_stop_play_files.clear() logger.bind(tag=TAG).debug("TTS会话启动成功") @@ -251,7 +251,7 @@ class TTSProvider(TTSProviderBase): self.text_to_speak(message.content_detail, None), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) logger.bind(tag=TAG).debug("TTS文本发送成功") except Exception as e: logger.bind(tag=TAG).error(f"发送TTS文本失败: {str(e)}") @@ -271,7 +271,7 @@ class TTSProvider(TTSProviderBase): self.finish_session(self.task_id), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) except Exception as e: logger.bind(tag=TAG).error(f"结束TTS会话失败: {str(e)}") continue diff --git a/main/xiaozhi-server/core/providers/tts/base.py b/main/xiaozhi-server/core/providers/tts/base.py index be42b14d..d4440f92 100644 --- a/main/xiaozhi-server/core/providers/tts/base.py +++ b/main/xiaozhi-server/core/providers/tts/base.py @@ -36,6 +36,7 @@ class TTSProviderBase(ABC): self.delete_audio_file = delete_audio_file self.audio_file_type = "wav" self.output_file = config.get("output_dir", "tmp/") + self.tts_timeout = int(config.get("tts_timeout", 15)) self.tts_text_queue = queue.Queue() self.tts_audio_queue = queue.Queue() self.tts_audio_first_sentence = True @@ -368,7 +369,7 @@ class TTSProviderBase(ABC): sendAudioMessage(self.conn, sentence_type, audio_datas, text), self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) # 记录输出和报告 if self.conn.max_output_size > 0 and text: diff --git a/main/xiaozhi-server/core/providers/tts/huoshan_double_stream.py b/main/xiaozhi-server/core/providers/tts/huoshan_double_stream.py index a237eedc..75a59f94 100644 --- a/main/xiaozhi-server/core/providers/tts/huoshan_double_stream.py +++ b/main/xiaozhi-server/core/providers/tts/huoshan_double_stream.py @@ -309,7 +309,7 @@ class TTSProvider(TTSProviderBase): self.start_session(self.conn.sentence_id), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) self.before_stop_play_files.clear() logger.bind(tag=TAG).debug("TTS会话启动成功") except Exception as e: @@ -326,7 +326,7 @@ class TTSProvider(TTSProviderBase): self.text_to_speak(message.content_detail, None), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) logger.bind(tag=TAG).debug("TTS文本发送成功") except Exception as e: logger.bind(tag=TAG).error(f"发送TTS文本失败: {str(e)}") @@ -346,7 +346,7 @@ class TTSProvider(TTSProviderBase): self.finish_session(self.conn.sentence_id), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) except Exception as e: logger.bind(tag=TAG).error(f"结束TTS会话失败: {str(e)}") continue diff --git a/main/xiaozhi-server/core/providers/tts/xunfei_stream.py b/main/xiaozhi-server/core/providers/tts/xunfei_stream.py index ed0a58d6..7a806014 100644 --- a/main/xiaozhi-server/core/providers/tts/xunfei_stream.py +++ b/main/xiaozhi-server/core/providers/tts/xunfei_stream.py @@ -179,7 +179,7 @@ class TTSProvider(TTSProviderBase): self.start_session(self.conn.sentence_id), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) self.before_stop_play_files.clear() logger.bind(tag=TAG).info("TTS会话启动成功") @@ -198,7 +198,7 @@ class TTSProvider(TTSProviderBase): self.text_to_speak(message.content_detail, None), loop=self.conn.loop, ) - future.result() + future.result(timeout=self.tts_timeout) logger.bind(tag=TAG).debug("TTS文本发送成功") except Exception as e: logger.bind(tag=TAG).error(f"发送TTS文本失败: {str(e)}") diff --git a/main/xiaozhi-server/plugins_func/functions/handle_exit_intent.py b/main/xiaozhi-server/plugins_func/functions/handle_exit_intent.py index bd24f372..154c4f89 100644 --- a/main/xiaozhi-server/plugins_func/functions/handle_exit_intent.py +++ b/main/xiaozhi-server/plugins_func/functions/handle_exit_intent.py @@ -31,6 +31,7 @@ handle_exit_intent_function_desc = { "handle_exit_intent", handle_exit_intent_function_desc, ToolType.SYSTEM_CTL ) def handle_exit_intent(conn: "ConnectionHandler", say_goodbye: str | None = None): + conn.is_exiting = True # 处理退出意图 try: if say_goodbye is None: