Paperjet/backend/app/services/trash_sweep.py

50 lines
1.6 KiB
Python

"""Background tasks for sweeping trash and old versions."""
from datetime import UTC, datetime, timedelta
from sqlalchemy.orm import Session
from app.config import settings
from app.models.annotation_state import AnnotationState
from app.models.document import Document
from app.models.version import Version
from app.services.assets import delete_asset_directory
from app.services.storage import delete_pdf_file
from app.services.thumbnails import delete_thumbnail
def sweep_trash(db: Session) -> None:
"""Permanently delete documents that have been in the trash past the retention period."""
cutoff = datetime.now(UTC) - timedelta(days=settings.TRASH_RETENTION_DAYS)
cutoff_iso = cutoff.isoformat()
docs_to_delete = (
db.query(Document)
.filter(Document.deleted_at.is_not(None), Document.deleted_at <= cutoff_iso)
.all()
)
for doc in docs_to_delete:
# Delete files from disk
delete_pdf_file(doc.id)
delete_thumbnail(doc.id)
delete_asset_directory(doc.id)
# Delete DB associations
db.query(AnnotationState).filter(AnnotationState.document_id == doc.id).delete()
db.query(Version).filter(Version.document_id == doc.id).delete()
# Delete document record
db.delete(doc)
db.commit()
def prune_auto_versions(db: Session) -> None:
"""Delete automatic versions older than the retention period."""
cutoff = datetime.now(UTC) - timedelta(days=settings.AUTO_VERSION_RETENTION_DAYS)
cutoff_iso = cutoff.isoformat()
db.query(Version).filter(Version.kind == "auto", Version.created_at <= cutoff_iso).delete()
db.commit()