0.1.0

wolai-backend/app/routers/tasks.py

@@ -1,9 +1,11 @@
 from fastapi import APIRouter, HTTPException, status
 
 from app.deps import AuthDep
-from app.schemas.tasks import OcrTaskRequest, TaskStatusResponse
+from app.schemas.tasks import MediaOcrTaskRequest, OcrTaskRequest, TaskStatusResponse
+from app.services.supabase_rest import supabase_rest
 from app.services.task_tracker import task_tracker
-from app.workers.tasks import ocr_pipeline
+from app.workers.tasks import media_ocr_pipeline, ocr_pipeline
+from app.workers.utils import dispatch_task
 
 router = APIRouter(prefix="/tasks")
 
@@ -12,7 +14,8 @@ router = APIRouter(prefix="/tasks")
 async def enqueue_ocr_task(payload: OcrTaskRequest, auth: AuthDep) -> TaskStatusResponse:
     """记录任务并投递 Celery，阶段 0 返回占位任务。"""
     task = task_tracker.create_task(user_id=auth.user_id, document_id=payload.document_id, task_type="ocr")
-    ocr_pipeline.delay(
+    dispatch_task(
+        ocr_pipeline,
         task_id=task.task_id,
         document_id=payload.document_id,
         file_url=str(payload.file_url),
@@ -21,6 +24,31 @@ async def enqueue_ocr_task(payload: OcrTaskRequest, auth: AuthDep) -> TaskStatus
     return task
 
 
+@router.post("/media-ocr")
+async def enqueue_media_ocr(payload: MediaOcrTaskRequest, auth: AuthDep) -> dict:
+    """触发媒体资产 OCR，占位实现先写入模拟文本。"""
+    asset = supabase_rest.select_one("media_assets", {"id": payload.asset_id})
+    if not asset:
+        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Media asset not found")
+
+    document_id = asset.get("document_id")
+    if not document_id:
+        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail="Media asset missing document")
+
+    document = supabase_rest.select_one("documents", {"id": document_id})
+    if not document or document.get("user_id") != auth.user_id:
+        raise HTTPException(status_code=status.HTTP_403_FORBIDDEN, detail="Forbidden")
+
+    supabase_rest.update("media_assets", {"id": payload.asset_id}, {"ocr_status": "processing"})
+    dispatch_task(
+        media_ocr_pipeline,
+        asset_id=payload.asset_id,
+        user_id=auth.user_id,
+        document_id=document_id,
+    )
+    return {"asset_id": payload.asset_id, "status": "queued"}
+
+
 @router.get("/{task_id}", response_model=TaskStatusResponse)
 async def get_task_status(task_id: str, auth: AuthDep) -> TaskStatusResponse:
     task = task_tracker.get_task(task_id=task_id, user_id=auth.user_id)