From f8ece2d984307d14620c48b1a4c30c70db702135 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E9=A6=99=E8=8D=89=E5=91=B3=E7=9A=84=E7=BA=B3=E8=A5=BF?= =?UTF-8?q?=E5=A6=B2?= <151599587+NyaOH-Nahida@users.noreply.github.com> Date: Tue, 11 Mar 2025 21:04:04 +0800 Subject: [PATCH] =?UTF-8?q?add:=20GSV=E6=9C=AC=E5=9C=B0=E8=AF=AD=E9=9F=B3?= =?UTF-8?q?=E5=90=88=E6=88=90=E7=9B=B8=E5=85=B3=20(#266)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- main/xiaozhi-server/config.yaml | 30 +++++ main/xiaozhi-server/core/providers/tts/GSV.py | 123 ++++++++++++++++++ 2 files changed, 153 insertions(+) create mode 100644 main/xiaozhi-server/core/providers/tts/GSV.py diff --git a/main/xiaozhi-server/config.yaml b/main/xiaozhi-server/config.yaml index a441930f..c8fbdb9f 100644 --- a/main/xiaozhi-server/config.yaml +++ b/main/xiaozhi-server/config.yaml @@ -311,6 +311,36 @@ TTS: media_type: "wav" streaming_mode: false threshold: 30 + GSV: + # GSV原神本地语音合成 + # 整合包下载：https://pan.ai-hobbyist.com/Infer%20Packs/TTS + # 国内镜像：https://alist.hancat.work/d/cdn/GPT-Sovits%E8%AF%AD%E9%9F%B3%E5%90%88%E6%88%90%E6%95%B4%E5%90%88%E5%8C%85%EF%BC%8C%E6%94%AF%E6%8C%81%E5%A4%9A%E4%BA%BA%E5%AF%B9%E8%AF%9D%E3%80%82%E6%B0%B8%E4%B9%85%E5%85%8D%E8%B4%B9%EF%BC%8C%E4%B8%A5%E7%A6%81%E5%80%92%E5%8D%96%EF%BC%81.7z + # 本TTS需要配合模型来使用，模型可以去这里下载： + # 二次元 & 游戏角色 GPT-Sovits V2 语音合成模型：https://www.modelscope.cn/models/aihobbyist/Anime_GPT-Sovits_Models + # 下载后将其解压到model文件夹 + # URL需要修改批处理文件才可生效，默认WebUI端口号为8080，API端口号为WebUI端口号+1 + url: "http://127.0.0.1:9881/infer_single" + app_key: "" + audio_dl_url: "http://127.0.0.1:9881" + model_name: "【原神】须弥" + speaker_name: "纳西妲" + prompt_text_lang: "中文" + emocton: "中立_neutral" + text_lang: "中文" + top_k: 10 + top_p: 1 + temperature: 1 + text_split_method: "按标点符号切" + batch_size: 10 + batch_size: 0.75 + split_bucket: true + speed_facter: 1 + fragment_interval: 0.3 + media_type: "wav" + parallel_infer: true + repetition_penalty: 1.35 + seed: -1 + output_file: tmp/ MinimaxTTS: # Minimax语音合成服务，需要先在minimax平台创建账户充值，并获取登录信息 # 平台地址：https://platform.minimaxi.com/ diff --git a/main/xiaozhi-server/core/providers/tts/GSV.py b/main/xiaozhi-server/core/providers/tts/GSV.py new file mode 100644 index 00000000..7a156c29 --- /dev/null +++ b/main/xiaozhi-server/core/providers/tts/GSV.py @@ -0,0 +1,123 @@ +import os +import uuid +import requests +from config.logger import setup_logging +from datetime import datetime +from core.providers.tts.base import TTSProviderBase + +TAG = __name__ +logger = setup_logging() + +class TTSProvider(TTSProviderBase): + def __init__(self, config, delete_audio_file): + super().__init__(config, delete_audio_file) + # POST请求地址 + self.url = config.get("url") + # 从配置中加载所有参数 + self.app_key = config.get("app_key") + self.audio_dl_url = config.get("audio_dl_url") + self.model_name = config.get("model_name") + self.speaker_name = config.get("speaker_name") + self.prompt_text_lang = config.get("prompt_text_lang") + self.emotion = config.get("emotion") + self.text_lang = config.get("text_lang") + self.top_k = config.get("top_k") + self.top_p = config.get("top_p") + self.temperature = config.get("temperature") + self.text_split_method = config.get("text_split_method") + self.batch_size = config.get("batch_size") + self.batch_threshold = config.get("batch_size") + self.split_bucket = config.get("split_bucket") + self.speed_facter = config.get("speed_facter") + self.fragment_interval = config.get("fragment_interval") + self.media_type = config.get("media_type") + self.parallel_infer = config.get("parallel_infer") + self.repetition_penalty = config.get("repetition_penalty") + self.seed = config.get("seed") + + def generate_filename(self): + """根据媒体类型生成文件名""" + extension = f".{self.media_type}" + return os.path.join( + self.output_file, + f"tts-{datetime.now().date()}@{uuid.uuid4().hex}{extension}" + ) + + async def text_to_speak(self, text, output_file): + """文本转语音并保存音频文件""" + # 构造请求体 + request_body = { + "app_key": self.app_key, + "audio_dl_url": self.audio_dl_url, + "model_name": self.model_name, + "speaker_name": self.speaker_name, + "prompt_text_lang": self.prompt_text_lang, + "emotion": self.emotion, + "text": text, + "text_lang": self.text_lang, + "top_k": self.top_k, + "top_p": self.top_p, + "temperature": self.temperature, + "text_split_method": self.text_split_method, + "batch_size": self.batch_size, + "batch_threshold": self.batch_threshold, + "split_bucket": self.split_bucket, + "speed_facter": self.speed_facter, + "fragment_interval": self.fragment_interval, + "media_type": self.media_type, + "parallel_infer": self.parallel_infer, + "repetition_penalty": self.repetition_penalty, + "seed": self.seed + } + + body = { + "app_key": "", + "audio_dl_url": "http://127.0.0.1:9881", + "model_name": "【原神】须弥", + "speaker_name": "纳西妲", + "prompt_text_lang": "中文", + "emotion": "中立_neutral", + "text": text, + "text_lang": "中文", + "top_k": 10, + "top_p": 1, + "temperature": 1, + "text_split_method": "按标点符号切", + "batch_size": 10, + "batch_threshold": 0.75, + "split_bucket": True, + "speed_facter": 1, + "fragment_interval": 0.3, + "media_type": "wav", + "parallel_infer": True, + "repetition_penalty": 1.35, + "seed": -1 + } + + try: + # 发送POST请求 + resp = requests.post(self.url, json=body) + resp.raise_for_status() + + # 解析响应 + result = resp.json() + if "audio_url" not in result: + logger.bind(tag=TAG).error("响应中缺少音频地址") + return + + # 下载音频文件 + audio_resp = requests.get(result["audio_url"]) + audio_resp.raise_for_status() + + # 保存文件 + with open(output_file, "wb") as f: + f.write(audio_resp.content) + + logger.bind(tag=TAG).info(f"音频文件已保存至: {output_file}") + + except requests.exceptions.HTTPError as e: + logger.bind(tag=TAG).error(f"HTTP请求失败: {str(e)}") + except requests.exceptions.JSONDecodeError: + logger.bind(tag=TAG).error("响应解析失败，无效的JSON格式") + except Exception as e: + logger.bind(tag=TAG).error(f"语音合成失败: {str(e)}") \ No newline at end of file