VOSK的Python服务端部署完整方案

1 main.py

import os
import json
import logging
import tempfile
import numpy as np
from fastapi import FastAPI, File, UploadFile, HTTPException
from fastapi.responses import JSONResponse
from vosk import Model, KaldiRecognizer
from vosk_helper import AudioProcessor

# 初始化FastAPI应用
app = FastAPI(title="VOSK语音识别服务", version="1.1.0")

# 配置日志
logging.basicConfig(
    level=logging.INFO,
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger("ASR-Service")

# 全局模型加载
@app.on_event("startup")
async def load_asr_model():
    try:
        model_path = os.getenv("VOSK_MODEL_PATH", "/app/models/vosk-model-cn")
        if not os.path.exists(model_path):
            raise FileNotFoundError(f"模型路径不存在: {model_path}")
        
        app.state.model = Model(model_path)
        app.state.audio_processor = AudioProcessor()
        logger.info(f"成功加载VOSK模型，版本：{app.state.model.version}")
        
    except Exception as e:
        logger.error(f"模型加载失败: {str(e)}")
        raise RuntimeError("服务初始化失败")

@app.post("/recognize", summary="语音识别接口", response_description="识别结果")
async def recognize_speech(
    file: UploadFile = File(..., description="上传的音频文件（支持wav/mp3/amr格式）")
):
    """
    语音识别处理接口：
    - 支持格式：WAV/MP3/AMR
    - 最大文件尺寸：25MB
    - 返回：识别文本和置信度
    """
    # 文件校验
    if not file.content_type.startswith('audio/'):
        logger.warning(f"非法文件类型: {file.content_type}")
        raise HTTPException(400, detail="仅支持音频文件上传")

    # 创建临时文件
    suffix = os.path.splitext(file.filename)[1]
    with tempfile.NamedTemporaryFile(suffix=suffix, delete=False) as tmp:
        try:
            content = await file.read()
            # 文件大小限制25MB
            if len(content) > 25 * 1024 * 1024:
                raise HTTPException(413, "文件大小超过25MB限制")
            
            tmp.write(content)
            tmp_path = tmp.name

        except Exception as e