from fastapi import FastAPI, Query
from fastapi.staticfiles import StaticFiles
from kokoro import KPipeline
import soundfile as sf
from pydub import AudioSegment
import os
from fastapi.responses import JSONResponse
from pydantic import BaseModel
import uuid
import time
app = FastAPI()

# 挂载 static 目录
os.makedirs("static", exist_ok=True)
app.mount("/static", StaticFiles(directory="static"), name="static")

class AudioResponse(BaseModel):
    audio_url: str
    time: float
# 初始化KPipeline
pipeline = KPipeline(lang_code='z')
@app.get("/generate_audio/", response_model=AudioResponse)
async def generate_audio(text: str = Query(..., description="输入要转换为语音的文本"),
                         voice: str = Query("zf_xiaoxiao", description="选择语音类型")):
    per=time.time()
    # 生成音频
    generator = pipeline(text, voice=voice, speed=1, split_pattern=r'\n+')
    
    # 保存每个音频片段
    audio_files = []
    for i, (gs, ps, audio) in enumerate(generator):
        audio_filename = f"static/temp_{i}.wav"
        sf.write(audio_filename, audio, 24000)
        audio_files.append(audio_filename)
    
    # 合并音频片段
    combined_audio = AudioSegment.empty()
    for audio_file in audio_files:
        audio_segment = AudioSegment.from_wav(audio_file)
        combined_audio += audio_segment
    
    # 生成唯一的音频文件名
    unique_filename = f"audio_{uuid.uuid4().hex}.wav"
    combined_filepath = os.path.join("static", unique_filename)
    
    # 保存合并后的音频文件
    combined_audio.export(combined_filepath, format="wav")
    
    # 清理临时音频文件
    for audio_file in audio_files:
        os.remove(audio_file)
    end=time.time()
    # 返回合并后的音频文件URL
    audio_url = f"/static/{unique_filename}"
    return {"audio_url": audio_url,"time": end-per}

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=8000)