feat: 接入 edge-tts 解决文字转语音的问题 close #642

xiaomusic/config.py

@@ -228,6 +228,12 @@ class Config:
     web_music_proxy: bool = (
         os.getenv("XIAOMUSIC_WEB_MUSIC_PROXY", "true").lower() == "true"
     )
+    # edge-tts 语音角色
+    edge_tts_voice: str = os.getenv("XIAOMUSIC_EDGE_TTS_VOICE", "")
+    # 是否启用定时清理临时文件
+    enable_auto_clean_temp: bool = (
+        os.getenv("XIAOMUSIC_ENABLE_AUTO_CLEAN_TEMP", "true").lower() == "true"
+    )
 
     def append_keyword(self, keys, action):
         for key in keys.split(","):

xiaomusic/crontab.py

@@ -1,4 +1,5 @@
 import json
+import os
 
 from apscheduler.schedulers.asyncio import AsyncIOScheduler
 from apscheduler.triggers.base import BaseTrigger
@@ -218,3 +219,34 @@ class Crontab:
             self.log.info("crontab reload_config ok")
         except Exception as e:
             self.log.exception(f"Execption {e}")
+
+        # 添加定时清理临时文件任务
+        if xiaomusic.config.enable_auto_clean_temp:
+
+            async def clean_temp_job():
+                try:
+                    temp_dir = xiaomusic.config.temp_dir
+                    if not os.path.exists(temp_dir):
+                        self.log.info(f"临时目录不存在: {temp_dir}")
+                        return
+
+                    files = os.listdir(temp_dir)
+                    deleted_count = 0
+                    for filename in files:
+                        try:
+                            file_path = os.path.join(temp_dir, filename)
+                            if os.path.isfile(file_path):
+                                os.remove(file_path)
+                                deleted_count += 1
+                                self.log.debug(f"已删除临时文件: {file_path}")
+                        except Exception as e:
+                            self.log.warning(f"删除文件失败 {file_path}: {e}")
+
+                    self.log.info(
+                        f"定时清理临时文件完成，共删除 {deleted_count} 个文件"
+                    )
+                except Exception as e:
+                    self.log.exception(f"清理临时文件异常: {e}")
+
+            self.add_job("0 3 * * *", clean_temp_job)
+            self.log.info("已添加每日凌晨3点定时清理临时文件任务")

xiaomusic/device_player.py

@@ -77,6 +77,8 @@ class XiaoMusicDevice:
 
         # 添加歌曲定时器
         self._add_song_timer = None
+        # TTS 播放定时器
+        self._tts_timer = None
 
     @property
     def did(self):
@@ -599,24 +601,82 @@ class XiaoMusicDevice:
     async def text_to_speech(self, value):
         """文字转语音"""
         try:
-            # 有 tts command 优先使用 tts command 说话
-            if self.hardware in TTS_COMMAND:
-                tts_cmd = TTS_COMMAND[self.hardware]
-                self.log.info("Call MiIOService tts.")
-                value = value.replace(" ", ",")  # 不能有空格
-                await miio_command(
-                    self.auth_manager.miio_service,
-                    self.did,
-                    f"{tts_cmd} {value}",
-                )
+            # 检查是否配置了 edge-tts 语音角色
+            if self.config.edge_tts_voice:
+                await self._text_to_speech_edge_tts(value)
             else:
-                self.log.debug("Call MiNAService tts.")
-                await self.auth_manager.mina_service.text_to_speech(
-                    self.device_id, value
-                )
+                # 使用原有的 TTS 逻辑
+                # 有 tts command 优先使用 tts command 说话
+                if self.hardware in TTS_COMMAND:
+                    tts_cmd = TTS_COMMAND[self.hardware]
+                    self.log.info("Call MiIOService tts.")
+                    value = value.replace(" ", ",")  # 不能有空格
+                    await miio_command(
+                        self.auth_manager.miio_service,
+                        self.did,
+                        f"{tts_cmd} {value}",
+                    )
+                else:
+                    self.log.debug("Call MiNAService tts.")
+                    await self.auth_manager.mina_service.text_to_speech(
+                        self.device_id, value
+                    )
         except Exception as e:
             self.log.exception(f"Execption {e}")
 
+    async def _text_to_speech_edge_tts(self, value):
+        """使用 edge-tts 进行文字转语音"""
+        from xiaomusic.utils.music_utils import get_local_music_duration
+        from xiaomusic.utils.network_utils import text_to_mp3
+
+        try:
+            # 取消之前的 TTS 定时器
+            if self._tts_timer:
+                self._tts_timer.cancel()
+                self._tts_timer = None
+                self.log.info("已取消之前的 TTS 定时器")
+
+            # 使用 edge-tts 生成 MP3 文件
+            self.log.info(
+                f"使用 edge-tts 生成语音: {value}, voice: {self.config.edge_tts_voice}"
+            )
+            mp3_path = await text_to_mp3(
+                text=value,
+                save_dir=self.config.temp_dir,
+                voice=self.config.edge_tts_voice,
+            )
+            self.log.info(f"edge-tts 生成的文件路径: {mp3_path}")
+
+            # 生成播放 URL
+            url = self.xiaomusic.music_url_handler._get_file_url(mp3_path)
+            self.log.info(f"TTS 播放 URL: {url}")
+
+            # 播放 TTS 音频
+            await self.group_player_play(url)
+
+            # 获取 MP3 时长
+            duration = await get_local_music_duration(mp3_path, self.config)
+            self.log.info(f"TTS 音频时长: {duration} 秒")
+
+            # 创建定时器，时长到后停止
+            if duration > 0:
+
+                async def _tts_timeout():
+                    await asyncio.sleep(duration)
+                    try:
+                        self.log.info("TTS 播放定时器时间到")
+                        if self._tts_timer:
+                            self._tts_timer = None
+                            await self.stop(arg1="notts")
+                    except Exception as e:
+                        self.log.error(f"TTS 定时器异常: {e}")
+
+                self._tts_timer = asyncio.create_task(_tts_timeout())
+                self.log.info(f"已设置 TTS 定时器，{duration} 秒后停止")
+
+        except Exception as e:
+            self.log.exception(f"edge-tts 播放失败: {e}")
+
     async def group_player_play(self, url, name=""):
         """同一组设备播放"""
         device_id_list = self.xiaomusic.get_group_device_id_list(self.group_name)
@@ -841,6 +901,11 @@ class XiaoMusicDevice:
             self._stop_timer = None
             self.log.info("cancel_all_timer _stop_timer.cancel")
 
+        if self._tts_timer:
+            self._tts_timer.cancel()
+            self._tts_timer = None
+            self.log.info("cancel_all_timer _tts_timer.cancel")
+
     @classmethod
     def dict_clear(cls, d):
         """清空设备字典并取消所有定时器"""

xiaomusic/music_url.py

@@ -138,6 +138,21 @@ class MusicUrlHandler:
             str: 本地音乐播放URL
         """
         filename = self.music_library.get_filename(name)
+        self.log.info(
+            f"_get_local_music_url local music. name:{name}, filename:{filename}"
+        )
+        return self._get_file_url(filename)
+
+    def _get_file_url(self, filepath):
+        """根据文件路径生成可访问的URL
+
+        Args:
+            filepath: 文件的完整路径
+
+        Returns:
+            str: 文件访问URL
+        """
+        filename = filepath
 
         # 处理文件路径
         if filename.startswith(self.config.music_path):
@@ -146,9 +161,7 @@ class MusicUrlHandler:
         if filename.startswith("/"):
             filename = filename[1:]
 
-        self.log.info(
-            f"_get_local_music_url local music. name:{name}, filename:{filename}"
-        )
+        self.log.info(f"_get_file_url filepath:{filepath}, filename:{filename}")
 
         # 构造URL
         encoded_name = urllib.parse.quote(filename)

xiaomusic/static/default/setting.html

@@ -101,6 +101,31 @@ var vConsole = new window.VConsole();
         <label for="temp_path">临时文件目录:</label>
         <input id="temp_path" type="text" value="music/tmp" />
 
+        <label for="enable_auto_clean_temp">启用定时清理临时文件(每日凌晨3点):</label>
+        <select id="enable_auto_clean_temp">
+          <option value="true" selected>开启</option>
+          <option value="false">关闭</option>
+        </select>
+
+        <label for="edge_tts_voice">Edge-TTS 语音角色:</label>
+        <select id="edge_tts_voice">
+          <option value="" selected>不使用(默认)</option>
+          <option value="zh-CN-XiaoxiaoNeural">晓晓 (女声,温柔)</option>
+          <option value="zh-CN-XiaoyiNeural">晓伊 (女声,活泼)</option>
+          <option value="zh-CN-YunjianNeural">云健 (男声,成熟)</option>
+          <option value="zh-CN-YunxiNeural">云希 (男声,阳光)</option>
+          <option value="zh-CN-YunxiaNeural">云夏 (男声,少年)</option>
+          <option value="zh-CN-YunyangNeural">云扬 (男声,新闻)</option>
+          <option value="zh-CN-liaoning-XiaobeiNeural">晓北 (女声,东北)</option>
+          <option value="zh-CN-shaanxi-XiaoniNeural">晓妮 (女声,陕西)</option>
+          <option value="zh-HK-HiuGaaiNeural">曉佳 (女声,粤语)</option>
+          <option value="zh-HK-HiuMaanNeural">曉曼 (女声,粤语)</option>
+          <option value="zh-HK-WanLungNeural">雲龍 (男声,粤语)</option>
+          <option value="zh-TW-HsiaoChenNeural">曉臻 (女声,台湾)</option>
+          <option value="zh-TW-YunJheNeural">雲哲 (男声,台湾)</option>
+          <option value="zh-TW-HsiaoYuNeural">曉雨 (女声,台湾)</option>
+        </select>
+
         <label for="ffmpeg_location">ffmpeg路径:</label>
         <input id="ffmpeg_location" type="text" value="./ffmpeg/bin" />