feat(ml): schedule presentation auto-hide sweep + retention (#141 step 6)

scheduled_presentation_auto_apply (daily beat) runs presentation_auto_apply_sweep
— idempotent, so an interrupted run just re-runs next cycle (that's the recovery),
wall-clock bounded by soft/hard task time limits. prune_presentation_reviews
(daily beat) drops RESOLVED review flags older than 30 days (rule 89 retention).
Tests run both tasks via a monkeypatched session factory. Milestone 141 complete:
the presentation-chrome auto-hide + conflict-flagged review is now live end-to-end.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CDgx8bQS5YrGRK76v8HUnM

backend/app/celery_app.py

@@ -152,6 +152,15 @@ def make_celery() -> Celery:
                 "schedule": 86400.0,  # soft auto-apply: drop auto-tags now below
                 # their threshold (m139); no-op unless the auto-apply switch is on
             },
+            "presentation-auto-apply-daily": {
+                "task": "backend.app.tasks.ml.scheduled_presentation_auto_apply",
+                "schedule": 86400.0,  # auto-hide banner/editor chrome (#141);
+                # no-op unless presentation_auto_apply_enabled
+            },
+            "prune-presentation-reviews-daily": {
+                "task": "backend.app.tasks.ml.prune_presentation_reviews",
+                "schedule": 86400.0,  # retention: drop resolved review flags >30d
+            },
             "snapshot-head-metrics-daily": {
                 "task": "backend.app.tasks.maintenance.snapshot_head_metrics",
                 "schedule": 86400.0,

backend/app/tasks/ml.py

@@ -594,6 +594,46 @@ def scheduled_ccip_auto_apply() -> str:
         return f"applied={applied}"
 
 
+@celery.task(
+    name="backend.app.tasks.ml.scheduled_presentation_auto_apply",
+    soft_time_limit=1800, time_limit=2100,
+)
+def scheduled_presentation_auto_apply() -> str:
+    """Auto-hide presentation chrome (banner / editor screenshot) on a daily
+    passive sweep (#141). No-op unless presentation_auto_apply_enabled. Idempotent
+    — already-hidden images are skipped — so an interrupted run simply re-runs next
+    cycle (that IS the recovery). Wall-clock bounded by the task time limits."""
+    from ..services.ml.heads import presentation_auto_apply_sweep
+
+    SessionLocal = _sync_session_factory()
+    with SessionLocal() as session:
+        result = presentation_auto_apply_sweep(session)
+    return f"applied={result['n_applied']} flagged={result['n_flagged']}"
+
+
+@celery.task(name="backend.app.tasks.ml.prune_presentation_reviews")
+def prune_presentation_reviews() -> str:
+    """Retention (rule 89): drop RESOLVED presentation-review flags older than 30
+    days — the operator has acted on them, so they're just history (#141)."""
+    from datetime import UTC, datetime, timedelta
+
+    from sqlalchemy import delete as sa_delete
+
+    from ..models import PresentationReview
+
+    cutoff = datetime.now(UTC) - timedelta(days=30)
+    SessionLocal = _sync_session_factory()
+    with SessionLocal() as session:
+        res = session.execute(
+            sa_delete(PresentationReview).where(
+                PresentationReview.resolved_at.is_not(None),
+                PresentationReview.resolved_at < cutoff,
+            )
+        )
+        session.commit()
+    return f"pruned={res.rowcount}"
+
+
 @celery.task(
     name="backend.app.tasks.ml.scheduled_retract_auto_tags",
     soft_time_limit=1800, time_limit=2100,

tests/test_presentation_scheduled.py

@@ -0,0 +1,98 @@
+"""Scheduled presentation tasks (#141): the daily auto-hide sweep wrapper +
+resolved-flag retention. The task opens its own session via
+_sync_session_factory, monkeypatched here to run against the test's db_sync."""
+from datetime import UTC, datetime, timedelta
+
+import pytest
+from sqlalchemy import select
+
+from backend.app.models import (
+    ImageRecord,
+    MLSettings,
+    PresentationReview,
+    Tag,
+    TagHead,
+)
+from backend.app.tasks.ml import (
+    prune_presentation_reviews,
+    scheduled_presentation_auto_apply,
+)
+
+pytestmark = pytest.mark.integration
+
+
+class _Ctx:
+    def __init__(self, s):
+        self.s = s
+
+    def __enter__(self):
+        return self.s
+
+    def __exit__(self, *a):
+        return False
+
+
+def _sf(db_sync):
+    class _SM:
+        def __call__(self):
+            return _Ctx(db_sync)
+
+    return _SM()
+
+
+def _img(db, sha, emb=None):
+    img = ImageRecord(
+        path=f"/images/{sha}.jpg", sha256=sha * 32, size_bytes=1, mime="image/jpeg",
+        width=1, height=1, origin="imported_filesystem",
+        integrity_status="unknown", siglip_embedding=emb,
+    )
+    db.add(img)
+    db.flush()
+    return img
+
+
+def test_scheduled_presentation_sweep_hides_chrome(db_sync, monkeypatch):
+    monkeypatch.setattr(
+        "backend.app.tasks.ml._sync_session_factory", lambda: _sf(db_sync)
+    )
+    banner = db_sync.execute(
+        select(Tag).where(Tag.is_system.is_(True), Tag.name == "banner")
+    ).scalar_one()
+    s = db_sync.execute(select(MLSettings).where(MLSettings.id == 1)).scalar_one()
+    w = [0.0] * 1152
+    w[0] = 3.0
+    db_sync.add(TagHead(
+        tag_id=banner.id, embedding_version=s.embedder_model_version,
+        weights=w, bias=0.0, suggest_threshold=0.5, auto_apply_threshold=0.5,
+        n_pos=60, n_neg=90, ap=0.9, precision_cv=0.98, recall=0.7,
+    ))
+    emb = [0.0] * 1152
+    emb[0] = 3.0
+    _img(db_sync, "sc", emb)
+    db_sync.commit()
+    assert "applied=1" in scheduled_presentation_auto_apply()
+
+
+def test_prune_presentation_reviews_drops_old_resolved(db_sync, monkeypatch):
+    monkeypatch.setattr(
+        "backend.app.tasks.ml._sync_session_factory", lambda: _sf(db_sync)
+    )
+    banner = db_sync.execute(
+        select(Tag).where(Tag.is_system.is_(True), Tag.name == "banner")
+    ).scalar_one()
+    a, b, c = _img(db_sync, "p1"), _img(db_sync, "p2"), _img(db_sync, "p3")
+    old = datetime.now(UTC) - timedelta(days=40)
+    recent = datetime.now(UTC) - timedelta(days=1)
+    # old resolved → pruned; recent resolved → kept; unresolved → kept.
+    db_sync.add(PresentationReview(
+        image_record_id=a.id, tag_id=banner.id, conflict_score=0.7, resolved_at=old))
+    db_sync.add(PresentationReview(
+        image_record_id=b.id, tag_id=banner.id, conflict_score=0.7, resolved_at=recent))
+    db_sync.add(PresentationReview(
+        image_record_id=c.id, tag_id=banner.id, conflict_score=0.7))
+    db_sync.commit()
+    assert prune_presentation_reviews() == "pruned=1"
+    remaining = db_sync.execute(
+        select(PresentationReview.image_record_id)
+    ).scalars().all()
+    assert set(remaining) == {b.id, c.id}