feat: add production data retention
This commit is contained in:
@@ -0,0 +1,108 @@
|
||||
from __future__ import annotations
|
||||
|
||||
from dataclasses import dataclass
|
||||
from datetime import datetime, timedelta, timezone
|
||||
from typing import Callable
|
||||
|
||||
from sqlalchemy import delete, or_, select
|
||||
from sqlalchemy.orm import Session
|
||||
|
||||
from .models import CatchReport, ExternalObservation, ModerationEvent, ModerationStatus, SourceType, SubmissionAttempt
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class RetentionPolicy:
|
||||
submission_days: int = 1
|
||||
unreviewed_days: int = 30
|
||||
approved_personal_days: int = 180
|
||||
staging_days: int = 90
|
||||
audit_days: int = 365
|
||||
published_payload_days: int = 365
|
||||
|
||||
|
||||
def apply_retention(
|
||||
session: Session, *, policy: RetentionPolicy = RetentionPolicy(),
|
||||
now: datetime | None = None, dry_run: bool = True,
|
||||
delete_object: Callable[[str], None] | None = None,
|
||||
) -> dict[str, int]:
|
||||
current = now or datetime.now(timezone.utc)
|
||||
counts = {
|
||||
"submission_attempts": 0,
|
||||
"user_reports_anonymized": 0,
|
||||
"pending_reports_expired": 0,
|
||||
"screenshots_deleted": 0,
|
||||
"staging_observations_deleted": 0,
|
||||
"published_payloads_cleared": 0,
|
||||
"moderation_events_deleted": 0,
|
||||
}
|
||||
|
||||
attempts = list(session.scalars(select(SubmissionAttempt.id).where(
|
||||
SubmissionAttempt.created_at < current - timedelta(days=policy.submission_days),
|
||||
)))
|
||||
counts["submission_attempts"] = len(attempts)
|
||||
|
||||
candidate_cutoff = current - timedelta(days=min(policy.unreviewed_days, policy.approved_personal_days))
|
||||
candidates = list(session.scalars(select(CatchReport).where(
|
||||
CatchReport.source_type == SourceType.user,
|
||||
CatchReport.reported_at < candidate_cutoff,
|
||||
or_(
|
||||
CatchReport.player_name.is_not(None), CatchReport.source_url.is_not(None),
|
||||
CatchReport.raw_payload.is_not(None), CatchReport.screenshot_key.is_not(None),
|
||||
CatchReport.screenshot_upload_token_hash.is_not(None),
|
||||
),
|
||||
)))
|
||||
reports = [item for item in candidates if item.reported_at.replace(tzinfo=item.reported_at.tzinfo or timezone.utc) < current - timedelta(
|
||||
days=policy.approved_personal_days if item.moderation_status == ModerationStatus.approved else policy.unreviewed_days,
|
||||
)]
|
||||
counts["user_reports_anonymized"] = len(reports)
|
||||
counts["screenshots_deleted"] = sum(bool(item.screenshot_key) for item in reports)
|
||||
counts["pending_reports_expired"] = sum(item.moderation_status == ModerationStatus.pending for item in reports)
|
||||
|
||||
stale = list(session.scalars(select(ExternalObservation).where(
|
||||
ExternalObservation.catch_report_id.is_(None),
|
||||
ExternalObservation.status != "published",
|
||||
ExternalObservation.last_seen_at < current - timedelta(days=policy.staging_days),
|
||||
)))
|
||||
counts["staging_observations_deleted"] = len(stale)
|
||||
published = list(session.scalars(select(ExternalObservation).where(
|
||||
ExternalObservation.status == "published",
|
||||
ExternalObservation.last_seen_at < current - timedelta(days=policy.published_payload_days),
|
||||
)))
|
||||
published = [item for item in published if item.payload]
|
||||
counts["published_payloads_cleared"] = len(published)
|
||||
events = list(session.scalars(select(ModerationEvent.id).where(
|
||||
ModerationEvent.created_at < current - timedelta(days=policy.audit_days),
|
||||
)))
|
||||
counts["moderation_events_deleted"] = len(events)
|
||||
|
||||
if dry_run:
|
||||
return counts
|
||||
|
||||
if counts["screenshots_deleted"] and delete_object is None:
|
||||
raise ValueError("delete_object is required when retained screenshots must be deleted")
|
||||
|
||||
for report in reports:
|
||||
if report.screenshot_key and delete_object:
|
||||
delete_object(report.screenshot_key)
|
||||
if report.moderation_status == ModerationStatus.pending:
|
||||
session.add(ModerationEvent(
|
||||
catch_report=report, created_at=current,
|
||||
previous_status=ModerationStatus.pending, new_status=ModerationStatus.rejected,
|
||||
moderator="retention-policy", reason="pending report retention period expired",
|
||||
))
|
||||
report.moderation_status = ModerationStatus.rejected
|
||||
report.player_name = None
|
||||
report.source_url = None
|
||||
report.raw_payload = None
|
||||
report.screenshot_key = None
|
||||
report.screenshot_upload_token_hash = None
|
||||
for observation in stale:
|
||||
session.delete(observation)
|
||||
for observation in published:
|
||||
observation.payload = {}
|
||||
if attempts:
|
||||
session.execute(delete(SubmissionAttempt).where(SubmissionAttempt.id.in_(attempts)))
|
||||
if events:
|
||||
session.execute(delete(ModerationEvent).where(ModerationEvent.id.in_(events)))
|
||||
session.commit()
|
||||
return counts
|
||||
Reference in New Issue
Block a user