from worker_app.celery_app import celery_app
from packages.domain import AssetClassification, ClassificationJob, ClassificationJobStatus
from packages.adapters.in_memory import InMemoryClassificationJobRepository


@celery_app.task(name="worker.classify_asset")
def classify_asset(job_id: str) -> dict:
    """
    Classify asset task.
    
    Steps:
    1. Fetch ClassificationJob from repository
    2. Fetch Asset from repository
    3. Run classification model (placeholder: mock classification)
    4. Update ClassificationJob with result
    5. Return result
    """
    # TODO: Replace with real repository injection
    job_repo = InMemoryClassificationJobRepository()
    
    job = job_repo.get(job_id)
    if job is None:
        return {"status": "failed", "error": "job not found"}
    
    try:
        # Update job status to PROCESSING
        job.status = ClassificationJobStatus.PROCESSING
        job_repo.update(job)
        
        # Mock classification (in real implementation: use ML model, vision API, etc.)
        # For now, randomly classify based on asset_id hash
        asset_id_hash = sum(ord(c) for c in job.asset_id)
        classifications = list(AssetClassification)
        classification = classifications[asset_id_hash % len(classifications)]
        confidence = 0.85
        
        # Update job status to COMPLETED
        job.status = ClassificationJobStatus.COMPLETED
        job.classification = classification.value
        job.confidence = confidence
        job_repo.update(job)
        
        return {
            "status": "completed",
            "job_id": job.id,
            "classification": classification.value,
            "confidence": confidence,
        }
    except Exception as e:
        # Update job status to FAILED
        job.status = ClassificationJobStatus.FAILED
        job.error_message = str(e)
        job_repo.update(job)
        
        return {
            "status": "failed",
            "job_id": job.id,
            "error": str(e),
        }