Files
orchard/backend/app/purge_seed_data.py
Mondo Diaz c4c9c20763 Remove tag system, use versions only for artifact references
Tags were mutable aliases that caused confusion alongside the immutable
version system. This removes tags entirely, keeping only PackageVersion
for artifact references.

Changes:
- Remove tags and tag_history tables (migration 012)
- Remove Tag model, TagRepository, and 6 tag API endpoints
- Update cache system to create versions instead of tags
- Update frontend to display versions instead of tags
- Remove tag-related schemas and types
- Update artifact cleanup service for version-based ref_count
2026-02-03 12:18:19 -06:00

203 lines
6.6 KiB
Python

"""
Purge seed/demo data from the database.
This is used when transitioning an environment from dev/test to production-like.
Triggered by setting ORCHARD_PURGE_SEED_DATA=true environment variable.
"""
import logging
import os
from sqlalchemy.orm import Session
from .models import (
Project,
Package,
Artifact,
Upload,
PackageVersion,
ArtifactDependency,
Team,
TeamMembership,
User,
AccessPermission,
)
from .storage import get_storage
logger = logging.getLogger(__name__)
# Seed data identifiers (from seed.py)
SEED_PROJECT_NAMES = [
"frontend-libs",
"backend-services",
"mobile-apps",
"internal-tools",
]
SEED_TEAM_SLUG = "demo-team"
SEED_USERNAMES = [
"alice",
"bob",
"charlie",
"diana",
"eve",
"frank",
]
def should_purge_seed_data() -> bool:
"""Check if seed data should be purged based on environment variable."""
return os.environ.get("ORCHARD_PURGE_SEED_DATA", "").lower() == "true"
def purge_seed_data(db: Session) -> dict:
"""
Purge all seed/demo data from the database.
Returns a dict with counts of deleted items.
"""
logger.warning("PURGING SEED DATA - This will delete demo projects, users, and teams")
results = {
"dependencies_deleted": 0,
"versions_deleted": 0,
"uploads_deleted": 0,
"artifacts_deleted": 0,
"packages_deleted": 0,
"projects_deleted": 0,
"permissions_deleted": 0,
"team_memberships_deleted": 0,
"users_deleted": 0,
"teams_deleted": 0,
"s3_objects_deleted": 0,
}
storage = get_storage()
# Find seed projects
seed_projects = db.query(Project).filter(Project.name.in_(SEED_PROJECT_NAMES)).all()
seed_project_ids = [p.id for p in seed_projects]
if not seed_projects:
logger.info("No seed projects found, nothing to purge")
return results
logger.info(f"Found {len(seed_projects)} seed projects to purge")
# Find packages in seed projects
seed_packages = db.query(Package).filter(Package.project_id.in_(seed_project_ids)).all()
seed_package_ids = [p.id for p in seed_packages]
# Find artifacts in seed packages (via uploads)
seed_uploads = db.query(Upload).filter(Upload.package_id.in_(seed_package_ids)).all()
seed_artifact_ids = list(set(u.artifact_id for u in seed_uploads))
# Delete in order (respecting foreign keys)
# 1. Delete artifact dependencies
if seed_artifact_ids:
count = db.query(ArtifactDependency).filter(
ArtifactDependency.artifact_id.in_(seed_artifact_ids)
).delete(synchronize_session=False)
results["dependencies_deleted"] = count
logger.info(f"Deleted {count} artifact dependencies")
# 2. Delete package versions
if seed_package_ids:
count = db.query(PackageVersion).filter(
PackageVersion.package_id.in_(seed_package_ids)
).delete(synchronize_session=False)
results["versions_deleted"] = count
logger.info(f"Deleted {count} package versions")
# 3. Delete uploads
if seed_package_ids:
count = db.query(Upload).filter(Upload.package_id.in_(seed_package_ids)).delete(
synchronize_session=False
)
results["uploads_deleted"] = count
logger.info(f"Deleted {count} uploads")
# 4. Delete S3 objects for seed artifacts
if seed_artifact_ids:
seed_artifacts = db.query(Artifact).filter(Artifact.id.in_(seed_artifact_ids)).all()
for artifact in seed_artifacts:
if artifact.s3_key:
try:
storage.client.delete_object(Bucket=storage.bucket, Key=artifact.s3_key)
results["s3_objects_deleted"] += 1
except Exception as e:
logger.warning(f"Failed to delete S3 object {artifact.s3_key}: {e}")
logger.info(f"Deleted {results['s3_objects_deleted']} S3 objects")
# 5. Delete artifacts (only those with ref_count that would be 0 after our deletions)
# Since we deleted all versions pointing to these artifacts, we can delete them
if seed_artifact_ids:
count = db.query(Artifact).filter(Artifact.id.in_(seed_artifact_ids)).delete(
synchronize_session=False
)
results["artifacts_deleted"] = count
logger.info(f"Deleted {count} artifacts")
# 6. Delete packages
if seed_package_ids:
count = db.query(Package).filter(Package.id.in_(seed_package_ids)).delete(
synchronize_session=False
)
results["packages_deleted"] = count
logger.info(f"Deleted {count} packages")
# 7. Delete access permissions for seed projects
if seed_project_ids:
count = db.query(AccessPermission).filter(
AccessPermission.project_id.in_(seed_project_ids)
).delete(synchronize_session=False)
results["permissions_deleted"] = count
logger.info(f"Deleted {count} access permissions")
# 8. Delete seed projects
count = db.query(Project).filter(Project.name.in_(SEED_PROJECT_NAMES)).delete(
synchronize_session=False
)
results["projects_deleted"] = count
logger.info(f"Deleted {count} projects")
# 9. Find and delete seed team
seed_team = db.query(Team).filter(Team.slug == SEED_TEAM_SLUG).first()
if seed_team:
# Delete team memberships first
count = db.query(TeamMembership).filter(
TeamMembership.team_id == seed_team.id
).delete(synchronize_session=False)
results["team_memberships_deleted"] = count
logger.info(f"Deleted {count} team memberships")
# Delete the team
db.delete(seed_team)
results["teams_deleted"] = 1
logger.info(f"Deleted team: {SEED_TEAM_SLUG}")
# 10. Delete seed users (but NOT admin)
seed_users = db.query(User).filter(User.username.in_(SEED_USERNAMES)).all()
for user in seed_users:
# Delete any remaining team memberships for this user
db.query(TeamMembership).filter(TeamMembership.user_id == user.id).delete(
synchronize_session=False
)
# Delete any access permissions for this user
# Note: AccessPermission.user_id is VARCHAR (username), not UUID
db.query(AccessPermission).filter(AccessPermission.user_id == user.username).delete(
synchronize_session=False
)
db.delete(user)
results["users_deleted"] += 1
if results["users_deleted"] > 0:
logger.info(f"Deleted {results['users_deleted']} seed users")
db.commit()
logger.warning("SEED DATA PURGE COMPLETE")
logger.info(f"Purge results: {results}")
return results