add edge tts

tts/edge_tts_engine.py

@@ -0,0 +1,150 @@
+"""
+Edge-TTS 引擎实现
+"""
+import edge_tts
+from typing import Optional
+from io import BytesIO
+from .base import TTSEngine
+from utils.logger import logger
+
+
+class EdgeTTSEngine(TTSEngine):
+    """
+    Microsoft Edge TTS 引擎实现
+    
+    支持多种语言和声音，免费使用。
+    """
+
+    def __init__(self):
+        """初始化 Edge TTS 引擎"""
+        self.engine_name = "edge-tts"
+        self.engine_version = "1.0.0"
+        logger.info(f"Initialized {self.engine_name} engine")
+
+    async def synthesize(
+        self,
+        text: str,
+        language: str = "zh-CN",
+        voice: Optional[str] = None,
+        rate: float = 1.0,
+        pitch: float = 1.0,
+    ) -> BytesIO:
+        """
+        使用 Edge TTS 将文本合成为语音
+
+        Args:
+            text: 要合成的文本
+            language: 语言代码，默认 zh-CN (中文)
+            voice: 声音 ID，如果为 None 则使用语言默认声音
+            rate: 语速，1.0 为正常速度
+            pitch: 音调，1.0 为正常音调
+
+        Returns:
+            BytesIO 对象，包含合成后的 MP3 音频数据
+        """
+        try:
+            # 如果没有指定声音，使用语言默认声音
+            if voice is None:
+                voice = self._get_default_voice(language)
+
+            logger.debug(
+                f"Synthesizing text with Edge TTS - "
+                f"language={language}, voice={voice}, rate={rate}, pitch={pitch}"
+            )
+
+            # 构建速率和音调字符串（+/-值的百分比形式）
+            rate_str = f"{(rate - 1) * 100:+.0f}%"
+            pitch_str = f"{(pitch - 1) * 100:+.0f}Hz"
+
+            # 创建 Edge TTS 客户端并合成
+            communicate = edge_tts.Communicate(
+                text=text,
+                voice=voice,
+                rate=rate_str,
+                pitch=pitch_str,
+            )
+
+            # 收集所有音频数据块
+            audio_data = BytesIO()
+            async for chunk in communicate.stream():
+                if chunk["type"] == "audio":
+                    audio_data.write(chunk["data"])
+
+            audio_data.seek(0)
+            logger.debug(
+                f"Successfully synthesized text. Audio size: {audio_data.getbuffer().nbytes} bytes"
+            )
+            return audio_data
+
+        except Exception as e:
+            logger.error(f"Error synthesizing text with Edge TTS: {str(e)}")
+            raise
+
+    async def get_supported_voices(self, language: str = "zh-CN") -> list[dict]:
+        """
+        获取指定语言支持的声音列表
+
+        Args:
+            language: 语言代码，例如 'zh-CN'、'en-US' 等
+
+        Returns:
+            声音列表，包含 name、voice_id、locale 等信息
+        """
+        try:
+            logger.debug(f"Fetching supported voices for language: {language}")
+            voices = await edge_tts.list_voices()
+
+            # 筛选指定语言的声音
+            filtered_voices = [
+                {
+                    "name": voice.get("ShortName", ""),
+                    "voice_id": voice.get("ShortName", ""),
+                    "locale": voice.get("Locale", ""),
+                    "display_name": voice.get("DisplayName", ""),
+                    "gender": voice.get("Gender", ""),
+                }
+                for voice in voices
+                if voice.get("Locale", "").startswith(language.split("-")[0])
+            ]
+
+            logger.debug(f"Found {len(filtered_voices)} voices for language {language}")
+            return filtered_voices
+
+        except Exception as e:
+            logger.error(f"Error fetching supported voices: {str(e)}")
+            raise
+
+    def get_engine_name(self) -> str:
+        """获取引擎名称"""
+        return self.engine_name
+
+    def get_engine_version(self) -> str:
+        """获取引擎版本"""
+        return self.engine_version
+
+    @staticmethod
+    def _get_default_voice(language: str) -> str:
+        """
+        获取指定语言的默认声音
+
+        Args:
+            language: 语言代码
+
+        Returns:
+            默认声音 ID
+        """
+        # 预定义的语言默认声音映射表
+        default_voices = {
+            "zh-CN": "zh-CN-XiaoxiaoNeural",  # 中文（简体）- 晓晓
+            "zh-TW": "zh-TW-HsiaoChen",  # 中文（繁体）
+            "en-US": "en-US-AriaNeural",  # 英文（美国）
+            "en-GB": "en-GB-SoniaNeural",  # 英文（英国）
+            "ja-JP": "ja-JP-NanamiNeural",  # 日语
+            "ko-KR": "ko-KR-SunHiNeural",  # 韩语
+            "fr-FR": "fr-FR-CelesteNeural",  # 法语
+            "de-DE": "de-DE-ConraadNeural",  # 德语
+            "es-ES": "es-ES-AlvaroNeural",  # 西班牙语
+            "ru-RU": "ru-RU-DmitryNeural",  # 俄语
+        }
+
+        return default_voices.get(language, "zh-CN-XiaoxiaoNeural")