From 4d50eeb8303fb4d9514404a303363319c4c72174 Mon Sep 17 00:00:00 2001 From: Chingfeng Li Date: Wed, 6 Aug 2025 14:53:59 +0800 Subject: [PATCH] =?UTF-8?q?=E4=BF=AE=E6=AD=A3self=E9=94=99=E8=AF=AF?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../core/providers/tts/aliyun_stream.py | 4 +++- .../core/providers/tts/huoshan_double_stream.py | 4 +++- .../core/providers/tts/linkerai.py | 17 ++++++----------- 3 files changed, 12 insertions(+), 13 deletions(-) diff --git a/main/xiaozhi-server/core/providers/tts/aliyun_stream.py b/main/xiaozhi-server/core/providers/tts/aliyun_stream.py index 52451cba..4929491c 100644 --- a/main/xiaozhi-server/core/providers/tts/aliyun_stream.py +++ b/main/xiaozhi-server/core/providers/tts/aliyun_stream.py @@ -482,8 +482,10 @@ class TTSProvider(TTSProviderBase): finally: self._monitor_task = None - def to_tts_stream(self, text: str, opus_handler=self.handle_opus) -> None: + def to_tts_stream(self, text: str, opus_handler=None) -> None: """非流式TTS处理，用于测试及保存音频文件的场景""" + if opus_handler is None: + opus_handler = self.handle_opus try: # 创建新的事件循环 loop = asyncio.new_event_loop() 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 f45b9a96..118cb4fe 100644 --- a/main/xiaozhi-server/core/providers/tts/huoshan_double_stream.py +++ b/main/xiaozhi-server/core/providers/tts/huoshan_double_stream.py @@ -629,7 +629,7 @@ class TTSProvider(TTSProviderBase): def wav_to_opus_data_audio_raw_stream(self, raw_data_var, is_end=False, callback: Callable[[Any], Any]=None): return self.opus_encoder.encode_pcm_to_opus_stream(raw_data_var, is_end, callback=callback) - def to_tts_stream(self, text: str, opus_handler=self.handle_opus) -> None: + def to_tts_stream(self, text: str, opus_handler=None) -> None: """非流式生成音频数据，用于生成音频及测试场景 Args: @@ -639,6 +639,8 @@ class TTSProvider(TTSProviderBase): Returns: list: 音频数据列表 """ + if opus_handler is None: + opus_handler = self.handle_opus try: # 创建事件循环 loop = asyncio.new_event_loop() diff --git a/main/xiaozhi-server/core/providers/tts/linkerai.py b/main/xiaozhi-server/core/providers/tts/linkerai.py index a9877d43..09d3d82a 100644 --- a/main/xiaozhi-server/core/providers/tts/linkerai.py +++ b/main/xiaozhi-server/core/providers/tts/linkerai.py @@ -233,15 +233,15 @@ class TTSProvider(TTSProviderBase): logger.bind(tag=TAG).error(f"TTS请求异常: {e}") self.tts_audio_queue.put((SentenceType.LAST, [], None)) - def to_tts_stream(self, text: str) -> list: + def to_tts_stream(self, text: str, opus_handler=None) -> list: """非流式TTS处理，用于测试及保存音频文件的场景 Args: text: 要转换的文本 - - Returns: - list: 返回opus编码后的音频数据列表 + opus_handler: opus数据处理方法 """ + if opus_handler is None: + opus_handler = self.handle_opus start_time = time.time() text = MarkdownCleaner.clean_markdown(text) @@ -291,14 +291,9 @@ class TTSProvider(TTSProviderBase): # 最后一帧可能不足，用0填充 frame = frame + b"\x00" * (frame_bytes - len(frame)) - opus = self.opus_encoder.encode_pcm_to_opus( - frame, end_of_stream=(i + frame_bytes >= len(pcm_data)) + self.opus_encoder.encode_pcm_to_opus_stream( + frame, end_of_stream=(i + frame_bytes >= len(pcm_data)), callback=opus_handler ) - if opus: - opus_datas.extend(opus) - - return opus_datas except Exception as e: logger.bind(tag=TAG).error(f"TTS请求异常: {e}") - return []