feat: the membership sweep, and the state that makes its failures readable (387 C3)

A daily sweep that walks each platform's roster into `platform_membership`.
Daily because memberships change on a BILLING cycle, not a download cadence.

**`membership_sync` is the part that earns its keep.** Without it three very
different situations are one indistinguishable state — the account subscribes
to nothing, the sweep never ran, the sweep failed — and all three leave zero
rows in `platform_membership`. "You are tracking 12 sources you no longer
subscribe to" is correct in the first case and an invitation to cancel things
the operator is actively paying for in the other two. So C4 gates its
CONCLUSIONS on `last_success_at`, not merely its display, and `roster_is_fresh`
is computed server-side so no caller can forget to.

Two timestamps rather than one: `last_attempt_at` moves every run,
`last_success_at` only on a clean walk. The gap between them is the signal —
a sweep hammering a broken credential every day must not look healthy because
it ran recently, and there is a test for exactly that.

Rejected shortcuts, both tempting: `MAX(platform_membership.last_seen_at)`
cannot tell "synced fine, found nothing" from "never synced"; `task_run` is
worse, since its retention prunes ok rows after 24h and a sweep that last
succeeded three days ago would leave no trace at all.

**The fetch completes before anything is written.** That ordering is the safety
property: a walk that dies mid-pagination writes nothing, so a failure can
never leave a roster half this week's and half last week's. `touch_membership`
never deletes, so a failure cannot empty the roster either — but "intact"
should mean intact, not merely non-empty.

Rule 89's four, each where it actually lives: recovery is "run it again"
(upsert, no deletes); retention is C1's age-out-never-delete, because
disappearing IS the signal; the wall-clock deadline is per-platform and
distinct from the per-REQUEST timeout the client already has (rule 156 — a
paginated roster answering every page slowly-but-within-timeout would never
trip that one and would sit on a worker indefinitely); duration comes from the
existing TaskRun signal plumbing.

**A bug caught in review, not production:** the broad `except Exception` would
have swallowed Celery's SoftTimeLimitExceeded — which is an ORDINARY Exception
subclass, not a BaseException — letting the sweep run past the soft limit into
the hard one, where it is SIGKILLed mid-transaction. A sweep that cannot be
stopped is worse than one that fails. Now re-raised explicitly, with a test
that also asserts SoftTimeLimitExceeded is still an Exception, so the re-raise
cannot quietly become dead code.

Rule 164 is why this ships with UI rather than backend-only: a roster that
never synced must be VISIBLE as such. The card says "never synced" in words and
states no count at all — rendering it as 0 is the precise conflation the whole
step exists to prevent — while a real zero behind a real sync is reported as
zero, because that one IS an answer. Pinned in both directions.

Three independent gates decide whether a platform is swept — registered here,
client exposes `iter_memberships`, credential exists — each silent, so adding
SubscribeStar (D1) is one line and nothing else. A missing credential is not an
error: recording a failure would light up the UI for a feature never enabled.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LNXXULQDjVZmbuNa2G9mD9

backend/app/models/__init__.py

@@ -21,6 +21,7 @@ from .import_batch import ImportBatch
 from .import_settings import ImportSettings
 from .import_task import ImportTask
 from .library_audit_run import LibraryAuditRun
+from .membership_sync import MembershipSync
 from .ml_settings import MLSettings
 from .patreon_failed_media import PatreonFailedMedia
 from .patreon_seen_media import PatreonSeenMedia
@@ -82,6 +83,7 @@ __all__ = [
     "ImportTask",
     "ImportSettings",
     "LibraryAuditRun",
+    "MembershipSync",
     "MLSettings",
     "HeadAutoApplyRun",
     "HeadMetric",

backend/app/models/membership_sync.py

@@ -0,0 +1,77 @@
+"""membership_sync — did the roster actually sync, and when.
+
+Milestone 387, step C3.
+
+`platform_membership` records what was SEEN. This records whether looking
+happened at all, and that is a different fact — the one that makes an empty
+roster readable.
+
+## Why this table has to exist
+
+Without it, three very different situations are one indistinguishable state:
+
+* the account genuinely subscribes to nothing,
+* the sweep has never run,
+* the sweep ran and failed.
+
+All three produce zero rows in `platform_membership`. Telling the operator
+"you are tracking 12 sources you do not subscribe to" is correct in the first
+case and catastrophic in the other two — it is an invitation to cancel things
+they are actively paying for. C4 must therefore gate its CONCLUSIONS on
+`last_success_at`, not merely display it.
+
+`MAX(platform_membership.last_seen_at)` was the tempting shortcut and does not
+work: it cannot distinguish "synced fine, found nothing" from "never synced".
+`task_run` was the other candidate and is worse — its retention prunes ok rows
+after 24h, so a sweep that last succeeded three days ago would leave no trace
+at all.
+
+## Separate attempt and success timestamps, deliberately
+
+`last_attempt_at` moves every run; `last_success_at` moves only on a clean
+walk. The GAP between them is the staleness signal, and keeping them apart is
+what lets the UI say "last synced 3 days ago, last tried 20 minutes ago,
+failing" — which is a different message from either half alone.
+"""
+
+from datetime import datetime
+
+from sqlalchemy import DateTime, Integer, String, Text, UniqueConstraint, func
+from sqlalchemy.orm import Mapped, mapped_column
+
+from .base import Base
+
+
+class MembershipSync(Base):
+    __tablename__ = "membership_sync"
+    __table_args__ = (
+        UniqueConstraint("platform", name="uq_membership_sync_platform"),
+    )
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True)
+    platform: Mapped[str] = mapped_column(String(64), nullable=False)
+
+    # Moves on EVERY run, success or not — so "we are trying" is visible even
+    # while "we are succeeding" is not.
+    last_attempt_at: Mapped[datetime | None] = mapped_column(
+        DateTime(timezone=True), nullable=True
+    )
+    # Moves only on a COMPLETE walk. This is the freshness signal C4 gates its
+    # conclusions on; NULL means never — which must never be rendered as zero.
+    last_success_at: Mapped[datetime | None] = mapped_column(
+        DateTime(timezone=True), nullable=True
+    )
+    # How many memberships the last SUCCESSFUL walk saw. Paired with
+    # last_success_at so "0" is only ever readable as a real zero.
+    last_count: Mapped[int | None] = mapped_column(Integer, nullable=True)
+
+    # Cleared on success. Plain String, no CHECK — this carries an exception
+    # class name (PatreonAuthError, PatreonDriftError, ...) and the vocabulary
+    # is whatever the client raises, exactly as source.error_type works.
+    last_error_type: Mapped[str | None] = mapped_column(String(64), nullable=True)
+    last_error_message: Mapped[str | None] = mapped_column(Text, nullable=True)
+
+    updated_at: Mapped[datetime] = mapped_column(
+        DateTime(timezone=True), nullable=False,
+        server_default=func.now(), onupdate=func.now(),
+    )