update: 增加唤醒时声纹处理 ，1秒内发送至大模型 【需优化唤醒锁机制，中途会遭受打断(偶发)，考虑忽略检测】

main/xiaozhi-server/core/utils/util.py

@@ -1,7 +1,6 @@
 import re
 import os
 import json
-import wave
 import copy
 import socket
 import requests
@@ -12,6 +11,7 @@ from io import BytesIO
 from core.utils import p3
 from pydub import AudioSegment
 from typing import Callable, Any
+from core.providers.tts.dto.dto import SentenceType
 
 TAG = __name__
 emoji_map = {
@@ -274,33 +274,23 @@ def pcm_to_data_stream(raw_data, is_opus=True, callback: Callable[[Any], Any] =
             frame_data = chunk if isinstance(chunk, bytes) else bytes(chunk)
             callback(frame_data)
 
+def play_audio_response(conn, response):
+    """音频响应处理"""
+    conn.tts.tts_audio_queue.put((SentenceType.FIRST, [], response.get("text")))
+    play_audio_frames(conn, response.get("file_path"))
+    conn.tts.tts_audio_queue.put((SentenceType.LAST, [], None))
 
-def opus_datas_to_wav_bytes(opus_datas, sample_rate=16000, channels=1):
-    """
-    将opus帧列表解码为wav字节流
-    """
-    decoder = opuslib_next.Decoder(sample_rate, channels)
-    pcm_datas = []
 
-    frame_duration = 60  # ms
-    frame_size = int(sample_rate * frame_duration / 1000)  # 960
-
-    for opus_frame in opus_datas:
-        # 解码为PCM（返回bytes，2字节/采样点）
-        pcm = decoder.decode(opus_frame, frame_size)
-        pcm_datas.append(pcm)
-
-    pcm_bytes = b"".join(pcm_datas)
-
-    # 写入wav字节流
-    wav_buffer = BytesIO()
-    with wave.open(wav_buffer, "wb") as wf:
-        wf.setnchannels(channels)
-        wf.setsampwidth(2)  # 16bit
-        wf.setframerate(sample_rate)
-        wf.writeframes(pcm_bytes)
-    return wav_buffer.getvalue()
+def play_audio_frames(conn, file_path):
+    """播放音频文件并处理发送帧数据"""
+    def handle_audio_frame(frame_data):
+        conn.tts.tts_audio_queue.put((SentenceType.MIDDLE, frame_data, None))
 
+    audio_to_data_stream(
+        file_path,
+        is_opus=True,
+        callback=handle_audio_frame
+    )
 
 def check_vad_update(before_config, new_config):
     if (

main/xiaozhi-server/core/utils/wakeup_word.py

@@ -1,140 +0,0 @@
-import os
-import re
-import yaml
-import time
-import hashlib
-import portalocker
-from typing import Dict
-
-
-class FileLock:
-    def __init__(self, file, timeout=5):
-        self.file = file
-        self.timeout = timeout
-        self.start_time = None
-
-    def __enter__(self):
-        self.start_time = time.time()
-        while True:
-            try:
-                portalocker.lock(self.file, portalocker.LOCK_EX | portalocker.LOCK_NB)
-                return self.file
-            except portalocker.LockException:
-                if time.time() - self.start_time > self.timeout:
-                    raise TimeoutError("获取文件锁超时")
-                time.sleep(0.1)
-
-    def __exit__(self, exc_type, exc_val, exc_tb):
-        portalocker.unlock(self.file)
-
-
-class WakeupWordsConfig:
-    def __init__(self):
-        self.config_file = "data/.wakeup_words.yaml"
-        self.assets_dir = "config/assets/wakeup_words"
-        self._ensure_directories()
-        self._config_cache = None
-        self._last_load_time = 0
-        self._cache_ttl = 1  # 缓存有效期（秒）
-        self._lock_timeout = 5  # 文件锁超时时间（秒）
-
-    def _ensure_directories(self):
-        """确保必要的目录存在"""
-        os.makedirs(os.path.dirname(self.config_file), exist_ok=True)
-        os.makedirs(self.assets_dir, exist_ok=True)
-
-    def _load_config(self) -> Dict:
-        """加载配置文件，使用缓存机制"""
-        current_time = time.time()
-
-        # 如果缓存有效，直接返回缓存
-        if (
-            self._config_cache is not None
-            and current_time - self._last_load_time < self._cache_ttl
-        ):
-            return self._config_cache
-
-        try:
-            with open(self.config_file, "a+") as f:
-                with FileLock(f, timeout=self._lock_timeout):
-                    f.seek(0)
-                    content = f.read()
-                    config = yaml.safe_load(content) if content else {}
-                    self._config_cache = config
-                    self._last_load_time = current_time
-                    return config
-        except (TimeoutError, IOError) as e:
-            print(f"加载配置文件失败: {e}")
-            return {}
-        except Exception as e:
-            print(f"加载配置文件时发生未知错误: {e}")
-            return {}
-
-    def _save_config(self, config: Dict):
-        """保存配置到文件，使用文件锁保护"""
-        try:
-            with open(self.config_file, "w") as f:
-                with FileLock(f, timeout=self._lock_timeout):
-                    yaml.dump(config, f, allow_unicode=True)
-                    self._config_cache = config
-                    self._last_load_time = time.time()
-        except (TimeoutError, IOError) as e:
-            print(f"保存配置文件失败: {e}")
-            raise
-        except Exception as e:
-            print(f"保存配置文件时发生未知错误: {e}")
-            raise
-
-    def get_wakeup_response(self, voice: str) -> Dict:
-        voice = hashlib.md5(voice.encode()).hexdigest()
-        """获取唤醒词回复配置"""
-        config = self._load_config()
-
-        if not config or voice not in config:
-            return None
-
-        # 检查文件大小
-        file_path = config[voice]["file_path"]
-        if not os.path.exists(file_path) or os.stat(file_path).st_size < (15 * 1024):
-            return None
-
-        return config[voice]
-
-    def update_wakeup_response(self, voice: str, file_path: str, text: str):
-        """更新唤醒词回复配置"""
-        try:
-            # 过滤表情符号
-            filtered_text = re.sub(r'[\U0001F600-\U0001F64F\U0001F900-\U0001F9FF]', '', text)
-
-            config = self._load_config()
-            voice_hash = hashlib.md5(voice.encode()).hexdigest()
-            config[voice_hash] = {
-                "voice": voice,
-                "file_path": file_path,
-                "time": time.time(),
-                "text": filtered_text,
-            }
-            self._save_config(config)
-        except Exception as e:
-            print(f"更新唤醒词回复配置失败: {e}")
-            raise
-
-    def generate_file_path(self, voice: str) -> str:
-        """生成音频文件路径，使用voice的哈希值作为文件名"""
-        try:
-            # 生成voice的哈希值
-            voice_hash = hashlib.md5(voice.encode()).hexdigest()
-            file_path = os.path.join(self.assets_dir, f"{voice_hash}.wav")
-
-            # 如果文件已存在，先删除
-            if os.path.exists(file_path):
-                try:
-                    os.remove(file_path)
-                except Exception as e:
-                    print(f"删除已存在的音频文件失败: {e}")
-                    raise
-
-            return file_path
-        except Exception as e:
-            print(f"生成音频文件路径失败: {e}")
-            raise