声音克隆合成V3接口 python代码示例

import os
import time
import base64
import requests

def _ensure_dir(path: str) -> None:
   directory = os.path.dirname(path)
   if directory and not os.path.exists(directory):
       os.makedirs(directory, exist_ok=True)

def _save_bytes_to_file(data: bytes, save_path: str) -> str:
   _ensure_dir(save_path)
   with open(save_path, "wb") as f:
       f.write(data)
   return os.path.abspath(save_path)

def call_text_to_clone_voice_v3(
   base_url: str,
   key: str,
   source_audio_url: str,
   target_text: str,
   save_dir: str = "testdemo/outputs",
   speed: str = "1.0",
   pitch: str = "1.0",
   ref_text: str = "我的声音将用于平台克隆，并合法使用，为自己的行为负责",
   req_type: int = 2,
   notify_url: str  = None,
   speakerid: str = None,
   audio_format: str = "mp3",
) -> str:
   """调用 /api/texttoclonevoicev3/index 并将结果保存到本地或返回任务ID。

   req_type:
     - 2: 同步返回音频的base64，直接落盘
     - 1: 异步模式，必须提供 notify_url，本函数返回任务ID
   """
   audio_format = (audio_format or "mp3").lower()
   if audio_format not in ("mp3", "wav"):
       audio_format = "mp3"

   endpoint = f"{base_url.rstrip('/')}/api/texttoclonevoicev3/index"
   payload = {
       "key": key,
       "source_audio_url": source_audio_url,
       "target_text": target_text,
       "speed": str(speed),
       "pitch": str(pitch),
       "ref_text": ref_text,
       "type": str(req_type),
   }

   if speakerid:
       payload["speakerid"] = speakerid

   if req_type == 1:
       if not notify_url:
           raise ValueError("notify_url is required when type=1 (async mode)")
       payload["notify_url"] = notify_url

   resp = requests.post(endpoint, data=payload, timeout=120)
   resp.raise_for_status()
   result = resp.json()

   if result.get("code") != 200:
       raise RuntimeError(f"API error: {result}")

   data = result.get("data") or {}
   timestamp = int(time.time())

   if req_type == 2:
       audio_b64 = data.get("audio_base64")
       if not audio_b64:
           raise RuntimeError("Missing audio_base64 in response for type=2")
       audio_bytes = base64.b64decode(audio_b64)
       save_path = os.path.join(save_dir, f"texttoclonevoicev3_{timestamp}.{audio_format}")
       return _save_bytes_to_file(audio_bytes, save_path)

   # type == 1
   taskid = data.get("taskid")
   if not taskid:
       raise RuntimeError("Missing taskid in response for type=1")
   return str(taskid)


if __name__ == "__main__":
   # 仅保留：调用 texttoclonevoicev3 接口（同步返回base64并落盘）
 BASE_URL_V3 = "https://api.hihookeji.com"
   API_KEY_V3 = "平台申请"
   REF_TEXT_V3 = "我的声音将用于平台克隆，并合法使用，为自己的行为负责"
   EXAMPLE_SOURCE_AUDIO_URL = "https://example.mp3"
   TARGET_TEXT_3 = "这是声音克隆3同步合成示例。"
   try:
       out3 = call_text_to_clone_voice_v3(
           base_url=BASE_URL_V3,
           key=API_KEY_V3,
           source_audio_url=EXAMPLE_SOURCE_AUDIO_URL,
           target_text=TARGET_TEXT_3,
           speed="1.0",
           pitch="1.0",
           ref_text=REF_TEXT_V3,
           req_type=2
       )
       print(f"/api/texttoclonevoicev3/index 已保存到: {out3}")
   except Exception as e:
       print("/api/texttoclonevoicev3/index 调用失败:", e)