roa2web-service-auto/backend/modules/data_entry/services/cleanup_service.py

"""
Cleanup service for auto-deleting expired failed receipts.

US-008: Backend - Auto-Cleanup Erori După 7 Zile
- Finds receipts with processing_status='failed' and processing_completed_at < now() - 7 days
- Deletes the receipts and their attached files from storage
- Runs at startup and then daily as a background task
"""

import asyncio
import logging
from datetime import datetime, timedelta
from pathlib import Path
from typing import Optional

from sqlalchemy import select, and_
from sqlalchemy.ext.asyncio import AsyncSession
from sqlalchemy.orm import selectinload

from backend.modules.data_entry.db.models.receipt import Receipt, ReceiptAttachment
from backend.modules.data_entry.config import settings

logger = logging.getLogger(__name__)

# Cleanup configuration
CLEANUP_RETENTION_DAYS = 7
CLEANUP_INTERVAL_HOURS = 24

# In-memory storage for last cleanup stats (optional - for login notification)
_last_cleanup_stats: dict = {
    "count": 0,
    "timestamp": None
}


def get_last_cleanup_stats() -> dict:
    """Get stats from the last cleanup run for notification purposes."""
    return _last_cleanup_stats.copy()


async def cleanup_expired_failed_receipts(session: AsyncSession) -> int:
    """
    Find and delete receipts with processing_status='failed' older than 7 days.

    This function:
    1. Queries for failed receipts where processing_completed_at < now() - 7 days
    2. Deletes attachment files from disk
    3. Deletes the receipt records (cascade deletes attachment records)

    Args:
        session: AsyncSession for database operations

    Returns:
        Number of receipts deleted
    """
    global _last_cleanup_stats

    cutoff_date = datetime.utcnow() - timedelta(days=CLEANUP_RETENTION_DAYS)

    # Find expired failed receipts with their attachments
    query = select(Receipt).options(
        selectinload(Receipt.attachments)
    ).where(
        and_(
            Receipt.processing_status == "failed",
            Receipt.processing_completed_at.isnot(None),
            Receipt.processing_completed_at < cutoff_date
        )
    )

    result = await session.execute(query)
    expired_receipts = result.scalars().all()

    if not expired_receipts:
        logger.debug("[Cleanup] No expired failed receipts found")
        return 0

    deleted_count = 0
    deleted_files = 0

    upload_base_path = settings.upload_path_resolved

    for receipt in expired_receipts:
        try:
            # Delete attachment files from disk
            for attachment in receipt.attachments:
                file_path = upload_base_path / attachment.file_path
                if file_path.exists():
                    try:
                        file_path.unlink()
                        deleted_files += 1
                        logger.debug(f"[Cleanup] Deleted file: {file_path}")
                    except OSError as e:
                        logger.warning(f"[Cleanup] Failed to delete file {file_path}: {e}")

                # Also try to clean up empty parent directories
                parent_dir = file_path.parent
                if parent_dir.exists() and parent_dir != upload_base_path:
                    try:
                        # Only remove if directory is empty
                        if not any(parent_dir.iterdir()):
                            parent_dir.rmdir()
                            logger.debug(f"[Cleanup] Removed empty directory: {parent_dir}")
                    except OSError:
                        pass  # Directory not empty or permission issue, skip

            # Delete receipt (cascade deletes attachment records in DB)
            await session.delete(receipt)
            deleted_count += 1

        except Exception as e:
            logger.error(f"[Cleanup] Error deleting receipt {receipt.id}: {e}")
            continue

    # Commit all deletions
    if deleted_count > 0:
        await session.commit()

    # Update stats for notification
    _last_cleanup_stats = {
        "count": deleted_count,
        "files_deleted": deleted_files,
        "timestamp": datetime.utcnow().isoformat()
    }

    logger.info(f"[Cleanup] Cleaned up {deleted_count} expired failed receipts ({deleted_files} files)")

    return deleted_count


async def run_cleanup_task(get_session_func) -> None:
    """
    Background task that runs cleanup at startup and then every 24 hours.

    Args:
        get_session_func: Async generator function that yields database sessions
    """
    logger.info("[Cleanup] Starting cleanup background task")

    # Run immediately at startup
    try:
        async for session in get_session_func():
            count = await cleanup_expired_failed_receipts(session)
            if count > 0:
                logger.info(f"[Cleanup] Initial cleanup: {count} receipts removed")
            break
    except Exception as e:
        logger.error(f"[Cleanup] Initial cleanup failed: {e}")

    # Then run every 24 hours
    while True:
        try:
            await asyncio.sleep(CLEANUP_INTERVAL_HOURS * 3600)

            async for session in get_session_func():
                count = await cleanup_expired_failed_receipts(session)
                if count > 0:
                    logger.info(f"[Cleanup] Daily cleanup: {count} receipts removed")
                break

        except asyncio.CancelledError:
            logger.info("[Cleanup] Cleanup task cancelled")
            raise
        except Exception as e:
            logger.error(f"[Cleanup] Daily cleanup failed: {e}")
            # Continue running even if one cleanup fails


# Global reference to cleanup task for graceful shutdown
_cleanup_task: Optional[asyncio.Task] = None


async def start_cleanup_task(get_session_func) -> bool:
    """
    Start the cleanup background task.

    Args:
        get_session_func: Async generator function that yields database sessions

    Returns:
        True if task started successfully, False otherwise
    """
    global _cleanup_task

    if _cleanup_task is not None and not _cleanup_task.done():
        logger.warning("[Cleanup] Cleanup task already running")
        return False

    try:
        _cleanup_task = asyncio.create_task(run_cleanup_task(get_session_func))
        logger.info("[Cleanup] ✅ Cleanup background task started")
        return True
    except Exception as e:
        logger.error(f"[Cleanup] Failed to start cleanup task: {e}")
        return False


async def stop_cleanup_task() -> None:
    """Stop the cleanup background task gracefully."""
    global _cleanup_task

    if _cleanup_task is not None and not _cleanup_task.done():
        _cleanup_task.cancel()
        try:
            await _cleanup_task
        except asyncio.CancelledError:
            pass
        logger.info("[Cleanup] Cleanup task stopped")

    _cleanup_task = None


def is_cleanup_task_running() -> bool:
    """Check if the cleanup task is currently running."""
    return _cleanup_task is not None and not _cleanup_task.done()