add:实时打断对话

README.md

@@ -27,13 +27,13 @@
 ## 已实现
 
 - `xiaozhi-esp32` 通信 WebSocket 协议
+- 支持实时打断对话，试一下在聊天的时候说一句`你好小智`
 - 支持国语、粤语、英语、日语、韩语 5 种语言识别（FunASR（默认））
 - 自由更换 LLM（支持ChatGLM（默认）、Dify、DeepSeek）
 - 自由更换 TTS（支持EdgeTTS（默认）、火山引擎豆包TTS）
 
 ## 正在实现
 
-- 打断对话
 - 按键手动对话
 - 长时间不聊天进入休眠状态
 - 对话记忆
@@ -213,7 +213,11 @@ https://espressif.github.io/esp-launchpad/
 
 建议：大模型和TTS都是依赖接口，如果网络环境不佳，可以考虑换成本地模型。或多尝试切换不同的接口模型。
 
-## 3、更多问题，可联系我们反馈
+## 3、为啥我的ChatGLMLLM回复有点问题，明明它才是小智，却把我当小智。
+
+建议：第一步可以先`配置文件`里的调整提示词。第二步，`配置文件`中用的模型是免费的模型：`glm-4-flash`，可以考虑换成收费版本的。
+
+## 4、更多问题，可联系我们反馈
 
 ![图片](docs/images/wechat.jpg)
 

README_en.md

@@ -32,13 +32,13 @@ To fully experience this project, follow these steps:
 ## Implemented
 
 - `xiaozhi-esp32` WebSocket communication protocol
+- Support interrupt dialogue in real time, try to say a sentence when chatting `你好小智`
 - Support for 5 languages: Mandarin, Cantonese, English, Japanese, Korean (FunASR - default)
 - Flexible LLM switching (ChatGLM - default, Dify, DeepSeek)
 - Flexible TTS switching (EdgeTTS - default, ByteDance Doubao TTS)
 
 ## In Progress
 
-- Conversation interruption
 - Manual button-triggered dialogue
 - Sleep mode after inactivity
 - Dialogue memory
@@ -223,7 +223,12 @@ both are slow, the network environment may need to be optimized.
 Suggestions: Both big models and TTS are dependent interfaces. If the network environment is not good, you can consider
 changing the local model. Or try to switch different interface models.
 
-## 3、For more questions, contact us to feedback
+## 3、Why is my ChatGLMLLM replying to a bit? Obviously it is Xiaozhi, but treats me as Xiaozhi.
+
+Suggestion: The first step can be to adjust the prompt words in the configuration file. The second step, the model used
+in the configuration file is the free model: `glm-4-flash`. You might consider switching to the paid version.
+
+## 4、For more questions, contact us to feedback
 
 ![图片](docs/images/wechat.jpg)
 

core/connection.py

@@ -12,6 +12,7 @@ from collections import deque
 from core.utils.util import is_segment
 from core.utils.dialogue import Message, Dialogue
 from core.handle.textHandle import handleTextMessage
+from core.handle.abortHandle import handleAbortMessage
 from core.handle.helloHandle import handleHelloMessage
 from core.utils.util import get_string_no_punctuation_or_emoji
 from concurrent.futures import ThreadPoolExecutor, TimeoutError
@@ -28,6 +29,7 @@ class ConnectionHandler:
         self.session_id = None
         self.prompt = None
         self.welcome_msg = None
+        self.client_abort = False
 
         # 线程任务相关
         self.loop = asyncio.get_event_loop()
@@ -100,6 +102,8 @@ class ConnectionHandler:
             msg_json = json.loads(message)
             if msg_json["type"] == "hello":
                 await handleHelloMessage(self, "你好")
+            if msg_json["type"] == "abort":
+                await handleAbortMessage(self)
         except json.JSONDecodeError:
             await handleTextMessage(self, message)
 
@@ -126,6 +130,11 @@ class ConnectionHandler:
         self.llm_finish_task = False
         for content in llm_responses:
             response_message.append(content)
+            # 如果中途被打断，就停止生成
+            if self.client_abort:
+                start = len(response_message)
+                break
+
             end_time = time.time()  # 记录结束时间
             self.logger.debug(f"大模型返回时间时间: {end_time - start_time} 秒, 生成token={content}")
             if is_segment(response_message):
@@ -169,9 +178,11 @@ class ConnectionHandler:
                 except Exception as e:
                     self.logger.error(f"TTS 任务出错: {e}")
                     continue
-                asyncio.run_coroutine_threadsafe(
+                if not self.client_abort:
-                    sendAudioMessage(self, opus_datas, duration, text), self.loop
+                    # 如果没有中途打断就发送语音
-                )
+                    asyncio.run_coroutine_threadsafe(
+                        sendAudioMessage(self, opus_datas, duration, text), self.loop
+                    )
                 if self.tts.delete_audio_file and os.path.exists(tts_file):
                     os.remove(tts_file)
             except Exception as e:

core/handle/abortHandle.py

@@ -0,0 +1,16 @@
+import json
+import logging
+
+logger = logging.getLogger(__name__)
+
+
+async def handleAbortMessage(conn):
+    logger.info("Abort message received")
+    # 设置成打断状态，会自动打断llm、tts任务
+    conn.client_abort = True
+    # 打断屏显任务
+    conn.stop_all_tasks()
+    # 打断客户端说话状态
+    await conn.websocket.send(json.dumps({"type": "tts", "state": "stop", "session_id": conn.session_id}))
+    conn.clearSpeakStatus()
+    logger.info("Abort message received-end")

core/handle/audioHandle.py

@@ -20,6 +20,7 @@ async def handleAudioMessage(conn, audio):
     conn.asr_audio.append(audio)
     # 如果本段有声音，且已经停止了
     if conn.client_voice_stop:
+        conn.client_abort = False
         conn.asr_server_receive = False
         text, file_path = conn.asr.speech_to_text(conn.asr_audio, conn.session_id)
         logger.info(f"识别文本: {text}")