meijiaka-zy/python-api/app/api/v1/vidu.py

"""
Vidu API 代理路由
================

提供 Vidu 对口型（lip-sync）任务的提交和查询接口。
前端通过此接口提交任务并轮询状态，无需直接访问 Vidu API。
"""

import logging

from fastapi import APIRouter, Depends, HTTPException
from pydantic import BaseModel, Field

from app.api.deps import get_current_user
from app.models.user import User
from app.schemas.common import ApiResponse, success_response
from app.services.vidu_tts_service import ViduTTSService

logger = logging.getLogger(__name__)

router = APIRouter(prefix="/vidu", tags=["Vidu"])


# ========== 请求/响应模型 ==========


class LipSyncRequest(BaseModel):
    """对口型请求"""

    video_url: str = Field(..., min_length=1, description="原视频 URL")
    audio_url: str | None = Field(None, description="音频 URL（与 text 二选一）")
    text: str | None = Field(None, description="文本内容（与 audio_url 二选一）")
    voice_id: str | None = Field(None, description="音色 ID（文字驱动时生效）")
    speed: float = Field(default=1.0, ge=0.5, le=2.0, description="语速")
    volume: int = Field(default=0, ge=0, le=10, description="音量")
    ref_photo_url: str | None = Field(None, description="人脸参考图 URL")

    @staticmethod
    def validate_at_least_one_audio_source(values: dict) -> dict:
        """验证至少提供 audio_url 或 text 之一"""
        audio_url = values.get("audio_url")
        text = values.get("text")
        if not audio_url and not text:
            raise ValueError("必须提供 audio_url 或 text 之一")
        return values


class LipSyncResponse(BaseModel):
    """对口型任务提交响应"""

    task_id: str = Field(..., description="Vidu 任务 ID")
    message: str = Field(default="任务已提交", description="状态消息")


class LipSyncQueryResponse(BaseModel):
    """对口型任务查询响应"""

    task_id: str = Field(..., description="任务 ID")
    state: str = Field(..., description="任务状态: pending/processing/succeeded/failed")
    video_url: str | None = Field(None, description="生成后的视频 URL（成功时）")
    message: str | None = Field(None, description="状态描述或错误信息")
    creations: list[dict] | None = Field(None, description="Vidu 原始 creations 数据")


# ========== API 路由 ==========


@router.post("/lip-sync", response_model=ApiResponse[LipSyncResponse])
async def create_lip_sync_task(
    request: LipSyncRequest,
    current_user: User = Depends(get_current_user),
):
    """
    提交 Vidu 对口型任务

    需要提供：
    - video_url: 原视频 URL（人物出镜视频）
    - audio_url: 音频 URL（与 text 二选一）
    - text: 文本内容（与 audio_url 二选一，使用 Vidu 内置 TTS）

    返回 Vidu task_id，用于后续轮询查询。
    """
    try:
        # 验证至少提供一种音频来源
        if not request.audio_url and not request.text:
            raise HTTPException(
                status_code=400,
                detail="必须提供 audio_url 或 text 之一",
            )

        service = ViduTTSService()
        task_id = await service.lip_sync_create(
            video_url=request.video_url,
            audio_url=request.audio_url,
            text=request.text,
            voice_id=request.voice_id,
            speed=request.speed,
            volume=request.volume,
            ref_photo_url=request.ref_photo_url,
        )

        logger.info(f"[Vidu] 对口型任务提交成功: task_id={task_id}, user={current_user.id}")

        return success_response(
            data=LipSyncResponse(
                task_id=task_id,
                message="对口型任务已提交，请轮询查询状态",
            )
        )

    except HTTPException:
        raise
    except Exception as e:
        logger.error(f"[Vidu] 提交对口型任务失败: {e}")
        raise HTTPException(status_code=500, detail=f"提交对口型任务失败: {e}")


@router.get("/tasks/{task_id}/creations", response_model=ApiResponse[LipSyncQueryResponse])
async def query_lip_sync_task(
    task_id: str,
    current_user: User = Depends(get_current_user),
):
    """
    查询 Vidu 对口型任务状态

    返回任务状态及生成物信息。
    当 state=succeeded 时，video_url 为生成后的对口型视频 URL。
    """
    try:
        service = ViduTTSService()
        result = await service.lip_sync_query(task_id)

        state = result.get("state", "unknown")
        creations = result.get("creations", [])

        # 提取视频 URL（成功时）
        video_url = None
        if state == "succeeded" and creations:
            # creations 是数组，取第一个
            first_creation = creations[0] if creations else {}
            video_url = first_creation.get("url")

        logger.info(
            f"[Vidu] 查询对口型任务: task_id={task_id}, state={state}, user={current_user.id}"
        )

        return success_response(
            data=LipSyncQueryResponse(
                task_id=task_id,
                state=state,
                video_url=video_url,
                message=result.get("message") if state == "failed" else None,
                creations=creations if creations else None,
            )
        )

    except Exception as e:
        logger.error(f"[Vidu] 查询对口型任务失败: {e}")
        raise HTTPException(status_code=500, detail=f"查询任务失败: {e}")