012-kaopeilian/backend/app/api/v1/endpoints/speech.py

"""
语音识别 API
"""
from typing import Optional
from fastapi import APIRouter, Depends, File, UploadFile, Form, HTTPException
from pydantic import BaseModel

from app.core.database import get_db
from app.api.deps import get_current_user
from app.models.user import User
from app.services.speech_recognition import (
    get_speech_recognition_service,
    SpeechRecognitionError,
)

router = APIRouter()


class SpeechRecognitionRequest(BaseModel):
    """语音识别请求（文本形式）"""
    text: str
    session_id: Optional[int] = None


class SpeechRecognitionResponse(BaseModel):
    """语音识别响应"""
    code: int = 200
    message: str = "识别成功"
    data: dict


@router.post("/recognize/text", response_model=SpeechRecognitionResponse)
async def recognize_text(
    request: SpeechRecognitionRequest,
    current_user: User = Depends(get_current_user),
):
    """
    处理前端已识别的语音文本
    用于 Web Speech API 识别后的文本传输
    """
    service = get_speech_recognition_service("simple")
    
    try:
        text = await service.recognize_text(request.text)
        return SpeechRecognitionResponse(
            code=200,
            message="识别成功",
            data={
                "text": text,
                "session_id": request.session_id,
            }
        )
    except SpeechRecognitionError as e:
        raise HTTPException(status_code=400, detail=str(e))


@router.post("/recognize/audio", response_model=SpeechRecognitionResponse)
async def recognize_audio(
    audio: UploadFile = File(...),
    format: str = Form(default="wav"),
    sample_rate: int = Form(default=16000),
    engine: str = Form(default="aliyun"),
    current_user: User = Depends(get_current_user),
):
    """
    识别音频文件
    
    支持的音频格式：wav, pcm, mp3, ogg, opus
    支持的识别引擎：aliyun, xunfei
    """
    # 读取音频数据
    audio_data = await audio.read()
    
    if len(audio_data) == 0:
        raise HTTPException(status_code=400, detail="音频文件为空")
    
    if len(audio_data) > 10 * 1024 * 1024:  # 10MB 限制
        raise HTTPException(status_code=400, detail="音频文件过大，最大支持 10MB")
    
    service = get_speech_recognition_service(engine)
    
    try:
        text = await service.recognize_audio(audio_data, format, sample_rate)
        return SpeechRecognitionResponse(
            code=200,
            message="识别成功",
            data={
                "text": text,
                "format": format,
                "sample_rate": sample_rate,
                "engine": engine,
            }
        )
    except SpeechRecognitionError as e:
        raise HTTPException(status_code=400, detail=str(e))
    except NotImplementedError as e:
        raise HTTPException(status_code=501, detail=str(e))


@router.get("/engines")
async def get_available_engines(
    current_user: User = Depends(get_current_user),
):
    """
    获取可用的语音识别引擎列表
    """
    import os
    
    engines = [
        {
            "id": "simple",
            "name": "浏览器语音识别",
            "description": "使用浏览器内置的 Web Speech API 进行语音识别",
            "available": True,
        },
        {
            "id": "aliyun",
            "name": "阿里云智能语音",
            "description": "使用阿里云 NLS 服务进行高精度语音识别",
            "available": all([
                os.getenv("ALIYUN_ACCESS_KEY_ID"),
                os.getenv("ALIYUN_ACCESS_KEY_SECRET"),
                os.getenv("ALIYUN_NLS_APP_KEY"),
            ]),
        },
        {
            "id": "xunfei",
            "name": "讯飞语音识别",
            "description": "使用讯飞 IAT 服务进行语音识别",
            "available": all([
                os.getenv("XUNFEI_APP_ID"),
                os.getenv("XUNFEI_API_KEY"),
                os.getenv("XUNFEI_API_SECRET"),
            ]),
        },
    ]
    
    return {
        "code": 200,
        "message": "获取成功",
        "data": {
            "engines": engines,
            "default": "simple",
        }
    }