- scanner: gate the consent backstop with a per-(user,doc_type) one-shot
marker so a standing admin-disable doesn't re-enqueue idempotent deletes
every scan tick; the marker clears when the type is re-enabled. Derive
_TEXT_BACKSTOP_DOC_TYPES from INDEXED_DOC_TYPES so new indexed types are
covered automatically
- semantic: extract _consent_narrowed_doc_types so the search-side narrowing
is unit-testable; add tests for restrict/intersect/disjoint/empty
- purge route: cap doc_types length (abuse guard) -> 400
- tests: one-shot + re-enable backstop, too-many-doc_types 400
Deferred (noted on PR): per-document allowed_doc_types call is cache-hot;
purge "last error wins" — both logged. SonarCloud broad-except hotspots are
deliberate (noqa BLE001), reviewable in the UI.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- purge route: 400 (not 500) on a valid-JSON non-object body
- scanner: backstop-purge admin-disabled note/news_item/deck_card points
(their deletion-tracking lives inside the skipped scan_* fns), mirroring the
files path; gated on a concrete allow-set so fail-open never deletes
- processor: record_ingest_dropped("admin_disabled") so consent-skipped index
tasks are observable/alertable
- app.py: list /api/v1/vector-sync/purge in the endpoints log line
- capabilities: drop empty-string doc types; return frozenset throughout
- purge: document the count-before-delete approximation
- tests: non-object body -> 400, ProvisioningRequiredError -> 428, cache TTL
expiry refetch, and the scanner consent backstop
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Consume the astrolabe.semantic_search capability as the source of truth for
which content sources an admin has approved for semantic search, and enforce
it independently of Astrolabe (this server queries Qdrant directly).
- capabilities.py: cached per-user reader for enabled_doc_types (TTL+LRU,
fail-open so older Astrolabe / transient OCS errors don't break search)
- semantic search: intersect requested doc_types with the allowed set;
restrict to the allowed set when none requested; short-circuit when empty
- scanner: skip disabled sources during discovery (files discovery yields
nothing when disabled, so the existing grace-period reconcile purges them)
- processor: drop near-real-time index tasks for disabled doc_types
(webhook events bypass the scanner gate); deletes always proceed
- vector/purge.py + POST /api/v1/vector-sync/purge: admin-only global
delete-by-doc_type, called by Astrolabe when a source is disabled so
consent is binding on data-at-rest
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
OCR is the paid, opt-in tier (DOCUMENT_OCR_ENABLED, default off). The per-tier
escalation gate already declines to hop to OCR when it's disabled (the pre-OCR
tier is terminal — no surprise cost), but that left operators blind to how much
OCR demand exists.
evaluate_escalation now returns a structured EscalationDecision:
- "hop" — a higher tier can run; the caller raises EscalateError (queue-hop).
- "suppressed" — the ideal next tier (e.g. ocr) exists but is DISABLED; the caller
indexes the current tier's output as terminal and records the
would-be hop on the new astrolabe_document_escalation_suppressed_total
{from_tier,to_tier,reason} counter instead of hopping.
- None — index as-is (good text, or no such tier at all).
So with OCR off, escalation_suppressed_total{to_tier="ocr"} is the latent OCR
demand an operator weighs before enabling OCR; enabling it converts these into
real document_escalation_total{to_tier="ocr"} hops. next_available_tier gains an
ignore_enabled flag to compute the *ideal* (enabled-gate-ignored) target.
Tests: registry suppressed vs hop vs terminal (incl. structured-hop-not-suppressed
when OCR off but structured available); _parse_pdf_tier records suppressed +
indexes without raising.
Deck #324 (parent #323).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A stall is often systemic (a Qdrant/embedding outage stalls every in-flight
job), so reclaiming the whole batch at now() every */5min tick would
thundering-herd a recovering dependency, bypassing TieredEscalationStrategy's
per-job backoff. reclaim_stalled_ingest_jobs now offsets retry_at by a fixed
delay (INGEST_RECLAIM_RETRY_DELAY_SECONDS, default 30s; 0 = legacy immediate).
Also document the hot-vs-restart flag asymmetry: INGEST_ESCALATION_ENABLED is
re-read per job; INGEST_TRANSIENT_MAX_ATTEMPTS is snapshotted at worker startup.
Deck #323.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- tests: make the transient-backoff progression assertion load-independent by
bracketing the get_retry_decision call with before/after timestamps instead of
measuring against a second datetime.now() (no freezegun dependency).
- tests: use pytest.approx for the ingest-queue-depth gauge assertions —
SonarCloud python:S1244 (float == ) was a MAJOR reliability finding that
tripped the new_reliability_rating quality gate.
- processor: tighten the EscalateError lazy-bind comment (file processing already
imports the document stack via get_registry; the gating only spares the
delete / text-doc paths and module-load time).
Deck #323.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Register the periodic stalled-job reclaim on a dedicated ingest-maintenance
queue that every worker drains (any --tier), so reclaim still fires when the
fast fleet is scaled to zero and only ocr workers run. procrastinate's
periodic-defer dedup keeps it single-run across drainers.
- escalation: mark `unsupported`/`forced` reason labels as reserved (not raised).
- processor: note that options/progress_callback are intentionally not threaded
through _parse_pdf_tier yet (symmetric with the inline path).
- tests: assert TieredEscalationStrategy backoff progression (4/8/16/…/300s);
cover get_ingest_pending per-queue aggregation + the legacy job_counts
fallback; add an external-path zero-page no-escalation case; use the canonical
INGEST_QUEUE_FAST instead of the back-compat alias.
Deck #323.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Split external (procrastinate) document processing into per-tier queues so a
document is attempted at most once per tier and requeued to the next tier's
queue on a low-quality parse, using procrastinate's native retry.
- escalation.py: TIER_LADDER (fast->structured->ocr) + EscalateError signal
- registry: process_tier (one tier) + evaluate_escalation post-parse gate
(reuses classify_from_text) + next_available_tier; shared _classify_result
and _oversize_result with the inline pipeline
- processor: process_document(tier=...) runs one tier and raises EscalateError
before embed (junk text never indexed); inline memory path unchanged
- queue/procrastinate: ingest-fast|structured|ocr queues; TieredEscalationStrategy
(queue-hop on EscalateError, bounded same-tier transient retry); queue-aware
task; producer defers to ingest-fast; per-queue counts + all-queue reclaim
- cli: worker --tier {fast,structured,ocr}
- billing: pages_ocr usage event + pipeline_tier metadata (paid OCR billed apart)
- observability: astrolabe_ingest_queue_depth{queue,status} gauge + per-queue
counts in nc_get_vector_sync_status / management status endpoint
- config: INGEST_ESCALATION_ENABLED (default true), INGEST_TRANSIENT_MAX_ATTEMPTS
INGEST_ESCALATION_ENABLED=false and INGEST_QUEUE=memory preserve prior behaviour.
Deck #323.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Quality-gate fixes (new-code conditions on PR #902):
- new_security_hotspots_reviewed: drop the fake "http://nextcloud" host in the
manager tests to https:// (python:S5332 ×2).
- new_security_rating: generate the integration test's fake app password with
secrets.token_urlsafe instead of a hardcoded literal (python:S2068).
- new_reliability_rating: restructure the user_manager sleep so an explicit
await checkpoint lives inside the cancellation scope — await one waiter
directly while watching shutdown via start_soon (python:S7490). Behaviour is
unchanged: timeout, shutdown, or a provisioning ring all end the sleep.
Review nits:
- Move the shutdown test's fail_after(2) to wrap the whole task group so it
actually bounds the task-group exit (was guarding a no-op sleep); drop the
sleep(0) stub (python:S7491).
- Type _wake_on's wait_fn as Callable[[], Awaitable[object]].
- Note in _wire_vector_sync_state why provision_signal is set on the singleton
only, not fanned out to app.state.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- ProvisionSignal.wait() re-arms in a finally so a cancelled wait (shutdown
racing the doorbell) leaves a fresh unset event, not a stale set-but-consumed
one; preserves the no-await-before-swap lost-wakeup guarantee.
- Hoist user_manager_task's _wake_on helper out of the while loop (one object,
not one per iteration).
- Test: assert ProvisionSignal via its public wait() contract instead of the
private _event attribute.
- Add test_provision_app_password_wakes_user_manager covering the
api/passwords.py wake path (previously only LFv2 web + MCP tool were tested).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Background vector sync discovered newly provisioned users only on the
periodic user-manager poll (VECTOR_SYNC_USER_POLL_INTERVAL, default 60s),
delaying first indexing by up to a minute. Add a ProvisionSignal doorbell
that provisioning paths ring after storing a user's app password, waking
user_manager_task to re-poll and spawn the user's scanner at once. The
periodic poll remains the backstop (covers cross-replica provisioning).
- ProvisionSignal (stable reference, wait-and-re-arm) held on
VectorSyncState; closes the lost-wakeup window (no await between observing
the ring and re-arming; anyio.Event stickiness covers a mid-poll ring)
- user_manager_task races its poll timeout against the doorbell + shutdown
- notify_user_provisioned() rung from the three app-password provisioning
sites: Login Flow v2 web, MCP provisioning tool, management/BasicAuth API
Note: the pre-existing scanner_wake_event was never .set() and only wakes
existing scanners; a brand-new user has none, so the manager is what must
be nudged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Round-3 review on PR #891 (no blockers):
- Add test_encode_dav_path_encodes_exactly_once pinning the documented
decoded-input precondition ("already%20encoded.pdf" -> "already%2520...").
- format_exception_group: proper singular/plural ("1 sub-exception" vs
"N sub-exceptions") instead of "(s)".
- oauth_sync: use `if doc_task is not None:` to match processor_task's guard.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Round-1 review on PR #891:
- Guard processor_task's broad except handler against an unbound doc_task
(mirrors multi_user_processor_task): initialise doc_task=None before the loop
and branch the error log. Fixes a latent NameError if receive() raises a
non-TimeoutError/EndOfStream before the first document binds. Regression test
added.
- Drop the unnecessary `from __future__ import annotations` in vector/_errors.py
and express format_exception_group's non-group fast path as an explicit
isinstance check.
- Add a copy_resource Destination-header encoding test (analogue to MOVE);
strengthen the ExceptionGroup test to assert the full leaf repr survives.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two ingest-robustness fixes from card 309 (OHR-Bench smoke-test triage).
WebDAV paths flowed through the client already URL-decoded (unquote on the
PROPFIND/REPORT <d:href>, or raw MCP-tool input), so a '#' reached httpx as a
URL fragment and silently truncated the request -> spurious 404 on otherwise
valid files (e.g. law filenames with '#', commas, double/trailing spaces).
Route every caller-path builder through a new _webdav_path helper that
percent-encodes the path once (preserving separators); the MOVE/COPY
Destination header is encoded too.
Vector-sync runs inside anyio task groups, so a child-task failure surfaced as
a BaseExceptionGroup whose str() is the useless "unhandled errors in a
TaskGroup (N sub-exception)" -- hiding the real ConnectError operators need.
Add format_exception_group to flatten the group to its leaf exceptions and use
it at the broad catch/log sites in processor.py and oauth_sync.py.
Refs: Deck board 12 card 309 (AC #4 filename handling, AC #2 observability).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Simplify the OCS status guard to `if status and status != "ok"` — falsy
(missing/None/"") is tolerated more naturally than the explicit tuple.
- Add `test_value_error_from_ocs_failure_returns_none`, covering the
OCS-failure ValueError flowing through `_get_enabled_apps_or_none` to the
scan-all fallback.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Extract `_app_enabled` to a module-level helper so the gate predicate is
unit-tested directly instead of via an inline copy that could drift.
- Move `import logging` to module scope in test_scanner_app_gating.py.
- Harden `get_enabled_apps` OCS-envelope parsing (`X or {}` / `or []`) so a
present-but-null `ocs`/`data` coerces to empty instead of raising on
`None.get`; add parametrized malformed-envelope tests.
- Use https:// in the test request URL (SonarCloud S5332 hotspot).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The vector-sync scanner polled every indexed app (Notes, Files, News,
Deck) for every provisioned user on each scan cycle. When a user lacks
an app, its REST API returns 404; these were caught (indexing
continued) but flooded tenant logs with repeated 404s, scaling with
users x disabled-apps x scan-frequency and masking real failures.
Add NextcloudClient.get_enabled_apps(), which reads the per-user
/ocs/v2.php/core/navigation/apps endpoint (respects group
restrictions). Chosen over /cloud/capabilities because the News app
advertises no capability and never appears there.
scan_user_documents now resolves the enabled-app set once per cycle and
skips the Notes/News/Deck scans for apps the user lacks. Files stays
unconditional (core Tags API, not a 404 source). Detection failures
fall back to scanning every app (prior behaviour), so a transient
nav-endpoint blip never silently halts indexing; the per-app 404 guards
remain as the safety net.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Tagging an existing file/folder emits only OCP\SystemTag\MapperEvent — never a
Node*Event — so tagged PDFs were previously only picked up by the hourly
scanner. Subscribe to the tag event and reconcile membership so adding/removing
the `vector-index` tag (re)indexes in near-real time.
- webhook_presets: add OCP\SystemTag\MapperEvent to the files_sync preset
(NC 32+, where MapperEvent gained getWebhookSerializable(); harmless on older
servers — it just never fires).
- webhook_parser: parse MapperEvent (objectType=files) into a path-less file
"reconcile" task. The payload carries only a fileid + tagIds (no name/path),
so assign and unassign both collapse to a reconcile.
- processor._reconcile_tag_event: resolve the fileid against the user's current
vector-index PDFs (find_files_by_tag). Present -> index with the resolved
path/etag; absent -> flip to delete. Naturally handles "an unrelated tag
changed" and a tagged folder's own fileid (no-op; the scanner still expands
folders to descendants).
- Unit tests for the parser branch and the reconcile.
The matching admin-UI preset change ships separately in the astrolabe app repo.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Round-1 review follow-ups (PR #857):
- scanner.py: skip the rename-reconcile when the existing metadata point is a
placeholder. reconcile_document_path only touches real chunks, so a not-yet-
indexed file would just incur a 0-point set_payload; the real index writes the
current path anyway.
- test_sharing_state.py: add a dedup-hit case where the file was renamed AND the
user is new to the ACL, asserting both set_payload writes fire (file_path/title
and acl_principals).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The vector-sync pipeline derived an indexed file's display title from the
document's embedded metadata (e.g. a PDF's /Title), falling back to the
filename only when absent. That embedded title frequently disagrees with how
the user named the file in Nextcloud and is confusing in the astrolabe
vector-viz UI (a passive consumer of the `title` payload field).
For files, always derive the title from the Nextcloud filename via a shared
`file_title_from_path` helper. Notes/deck/news keep their metadata titles.
A rename/move in Nextcloud keeps the fileid (doc_id) and content (etag/mtime)
but changes the path, so both the dedup claim and the scanner freshness gate
skip re-embedding and the stored file_path/title go stale. Add
`reconcile_document_path`: a metadata-only set_payload that refreshes
file_path + title on the existing real chunks without re-fetch/re-embed.
Wire it into both skip paths:
- dedup hit (etag unchanged on rename) via claim_existing_index(current_path=...)
- scanner incremental skip (etag changed, mtime stable)
Both reuse already-fetched payloads, so steady-state scans add no extra
round-trip (reconcile is a no-op when the path is unchanged).
Refs: Deck #204
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Clear the module-singleton ingest references (task_producer,
document_send_stream, document_receive_stream) on lifespan shutdown via a new
_clear_vector_sync_state() helper, mirroring the eviction_task_group cleanup.
Defense-in-depth so a late webhook (or a module-singleton integration test)
can't touch a producer/stream backed by an already-closed resource.
- Add IngestTransport.backend_name ("memory"/"postgres") and use it in both
lifespan log lines, removing the last settings.ingest_queue read from the
background-sync setup — the lifespan no longer inspects the backend at all.
- Cover backend_name in the build_transport adapter-selection tests.
Refs: Deck #196
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- LocalTransport.run_consumers increments active_consumer_count per worker
(instead of once after the loop) so the count is accurate if a later
tg.start() raises mid-pool.
- Add LocalTransport.aclose() to explicitly close its owned send/receive stream
ends (belt-and-suspenders against unclosed-resource warnings; anyio aclose is
idempotent, and by shutdown the scanner is already winding down). Reworded the
base IngestTransport.aclose() docstring to point at the overrides.
- Inline ingest_transport.producer at the scanner/user_manager call sites,
dropping the single-use task_producer alias in both lifespan paths.
- Annotate DistributedTransport._producer explicitly as ProcrastinateTaskProducer
so the drain() coupling is visible and ty catches drift.
- Add a unit test for LocalTransport.aclose() (closes the owned streams,
idempotent).
Refs: Deck #196
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Add IngestTransport.active_consumer_count (0 by default; LocalTransport stores
the started count) so app.py logs the worker count without re-checking
INGEST_QUEUE — the last backend-knowledge leak in the lifespan is gone.
- Document that DistributedTransport is postgres/procrastinate-specific by design
(aclose() calls ProcrastinateTaskProducer.drain()); other distributed backends
would be separate IngestTransport subclasses.
- Clarify the _wire_vector_sync_state log line (writes app.state + singleton, not
only the singleton).
- Strengthen the LocalTransport test: assert active_consumer_count transitions
0→N and that each worker receives a distinct cloned receive stream.
Refs: Deck #196
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Note at both metrics-task call sites that receive_stream is None in postgres
mode (get_ingest_pending falls back to procrastinate counts).
- Add test_default_is_exact_true covering the status-endpoint count path.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Finish the hexagonal ports-&-adapters split started in #183. The producer side
already had a TaskProducer port + adapters, but the consumer side was
unabstracted and the INGEST_QUEUE selection leaked into a duplicated
`if use_postgres:` branch across both app.py lifespan paths.
Introduce an IngestTransport ABC (vector/queue/transport.py) that bundles the
producer with running (or not running) the in-process consumer pool, built by a
single build_transport() factory:
- LocalTransport (INGEST_QUEUE=memory): in-process anyio stream drained by an
N-worker pool that run_consumers starts.
- DistributedTransport (INGEST_QUEUE=postgres): wraps ProcrastinateTaskProducer;
run_consumers is a no-op because the consumer is the external `worker` role.
Both lifespan paths now call build_transport + _wire_vector_sync_state (new
helper that centralizes the app.state / module-singleton / browser-app writes) +
transport.run_consumers + transport.aclose(), with no INGEST_QUEUE branching and
no getattr drain probe. Adding a future backend (Redis/NATS/SQS) is one new
adapter + one build_transport arm, with no app.py or scanner change.
Preserves the single-tenant parallelism invariant (one shared multiplexed queue
+ N-worker pool, per-document not per-user dispatch) and documents it in
ADR-028. The worker CLI is unchanged (it is the external consumer).
Refs: Deck #196 (Deck #197 tracks the explicit parallelism regression test)
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Add Validator("VECTOR_SYNC_METRICS_REFRESH_INTERVAL", gte=1) so a 0/negative
value can't turn the publish loop into a busy-spin.
- Annotate count_indexed's qdrant_client param as AsyncQdrantClient.
- Add tests: exact kwarg is forwarded to qdrant count, and the placeholder
filter matches False (excludes placeholders) with chunk_index pinned to 0.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The only queue metric, mcp_vector_sync_queue_size, was updated inline by the
single-user consumer (processor_task) but never by the multi-user consumer
(oauth_processor_task). On multi-user tenants (e.g. blackbox-demo, 5 users) the
gauge read 0 for 24h while the live anyio buffer held ~2214 pending documents
(shown by /api/v1/vector-sync/status). The "indexed" figure was also a chunk
count (16039 points ≈ 480 docs) mislabelled as documents.
Publish a consumer-independent snapshot from a periodic task
(vector/metrics_publisher.vector_sync_metrics_task), spawned in BOTH lifespan
task groups (single-user and multi-user) and every queue backend:
- mcp_vector_sync_pending_documents — outstanding work via
ingest_status.get_ingest_pending() (anyio buffer depth or procrastinate
todo+doing); also keeps the legacy queue_size gauge meaningful on all paths.
- mcp_vector_sync_indexed_documents — distinct documents, counted exactly and
cheaply via the one chunk_index=0 point per document (no facet).
- mcp_vector_sync_indexed_chunks — total non-placeholder points.
The /api/v1/vector-sync/status endpoint now returns indexed_documents (distinct
docs) AND indexed_chunks separately, so documents and chunks are no longer
conflated. The publisher uses approximate Qdrant counts (every-N-seconds gauge);
the on-demand endpoint counts exactly. New knob:
VECTOR_SYNC_METRICS_REFRESH_INTERVAL (default 20s). Fail-safe: a metrics refresh
never disturbs ingest.
BREAKING CHANGE: /api/v1/vector-sync/status field `indexed_documents` now holds
the distinct-document count (was the chunk count); the chunk count moved to the
new `indexed_chunks` field. The Astrolabe UI + the nc_get_vector_sync_status MCP
tool / userinfo page are harmonized in a follow-up (Deck #195).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A file shared across many users — directly, or via a group folder shared
to a group — was parsed and embedded once per user. Chunk point IDs are
user-agnostic (uuid5(tenant_id, doc_id=fileid, chunk_index)), but the
per-user freshness gate filtered Qdrant by user_id, so two readers
ping-ponged: each overwrote the other's points and each kept seeing "not
indexed for me", reprocessing every scan. Production telemetry (note
386945, finding #5) measured identical docs re-processed every few hours
at 7-13s each, with PDF parse ~62% of per-doc cost.
Layer 1 — tenant-wide dedup:
- Thread the scanner's tag-REPORT etag into the file DocumentTask and the
chunk payload; index `etag` as a KEYWORD field.
- vector/sharing_state.find_indexed_content scrolls tenant-wide (no
user_id filter) for a non-placeholder point matching
(doc_id, doc_type, etag), gated on embedding_identity in Python so a
model switch correctly forces a re-embed.
- Scanner skips enqueue and the processor skips fetch/parse/embed when a
match exists (cross-worker race-guard before WebDAV read). Dedup is
fail-safe: a Qdrant error degrades to "process normally".
Layer 2 — observed-access ACL (no admin / GroupFolders API needed):
- Each point carries `acl_principals` = the set of user:<uid> whose
scanner has observed (hence can read) the file. The per-user tag REPORT
is the access oracle; group membership/GroupFolders enumeration is
admin-only and unavailable in multi-user modes.
- build_ownership_filter ORs MatchAny(acl_principals, ["user:<me>"]) so a
deduplicated shared/group-folder point surfaces to every reader;
verify-on-read (_verify_files) remains the precise ACL gate.
- Deletion/eviction become "release one user": drop the principal and
delete the points only when the set empties, so one user untagging a
shared file doesn't evict it for the others. Legacy points without the
field keep the original per-user delete.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
🟡 Document the _doc_queueing_lock ":" delimiter invariant (user_id and the
controlled doc_type enum are colon-free, so the key is collision-safe; a
future doc_type with ":" must not be added).
🟡 API pod no longer opens the procrastinate connector twice on startup: add
ProcrastinateTaskProducer.ensure_schema() (applies the schema on the
already-open pool) and have both lifespan branches build the producer then
ensure_schema — one open/close cycle, matching the worker. build_producer now
returns the concrete producer type.
🟢 Document in ports.py that a long-lived-connection producer may optionally
provide drain() (lifespan probes via getattr).
🟢 Add a unit test that a non-credential pipeline error propagates (for
procrastinate's RetryStrategy) and still closes the client via finally.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
🟡 Document why ProcrastinateTaskProducer.connect() uses `await app.open_async()`
(AwaitableContext: await opens a long-lived pool, closed by drain()) and add a
connect()/drain() lifecycle unit test (InMemoryConnector) asserting the pool is
opened by connect and closed by drain — previously untested.
🟡 get_procrastinate_conninfo: forward connect_timeout from DATABASE_URL or
default 10s so an unreachable DB can't hang worker/API startup indefinitely;
warn only on other dropped query params. + tests.
🟢 INGEST_DELETE_SUCCEEDED_JOBS (default true) makes the worker's succeeded-job
deletion configurable for audit retention.
🟢 Worker startup logs via logger.info (structured/OTel) instead of click.echo.
🟢 INGEST_STALLED_JOB_SECONDS (default 300) makes the crash-reclaim threshold
tunable for slow embedding backends; reclaim reads it per-run.
The broad `except` in _apply_ingest_queue_schema_open is kept deliberately:
procrastinate wraps psycopg errors, so narrowing to psycopg.errors.* would miss
the wrapped DDL-conflict and turn a benign concurrent-apply race into a failure;
the presence re-check re-raises genuine errors.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Stop excluding tests/ from the ty-check pre-commit hook so touched test files
are type-checked. Fix the new ingest tests under the now-active check:
- cast duck-typed JobContext / App test doubles to their declared types;
- narrow the gated Postgres fixture's str | None URL (pytest.skip isn't modelled
as NoReturn by ty).
Pre-existing type issues in untouched test modules are unaffected (the hook
checks only changed files); they'll be cleaned as those files are next touched.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Re-architect document ingest from the shared NATS-glued document-processor to a
per-tenant, in-process model owned by nextcloud-mcp-server (Deck #183). The MCP
server now owns both sides of ingest:
- Producer (api role): the scanner defers one job per changed document into the
app's Postgres via procrastinate (queueing_lock dedup; no execution lock, so a
crashed worker can't deadlock a doc — Qdrant upserts are idempotent).
- Consumer (worker role): `nextcloud-mcp-server worker` drains the queue and runs
the existing process_document pipeline; a periodic task reclaims jobs orphaned
in `doing` by a crash.
INGEST_QUEUE selects the transport (auto: postgres when DATABASE_URL is Postgres,
else the in-process anyio queue for SQLite/dev). procrastinate manages its own
tables (applied on a fresh DB at startup and by `db upgrade`). The vector-sync
status surface reads job counts from Postgres in postgres mode. procrastinate +
psycopg3 ship in the [postgres] extra; the app's own engine still uses asyncpg
(driver unification is a follow-up handled in the rendered Helm chart).
NATS JetStream, the Postgres-queue stub, the bus status subscriber, and nats-py
are removed.
BREAKING CHANGE: the external-NATS-ingest env vars are removed
(INGEST_MODE, STATUS_BACKEND, INGEST_BUS_URL, INGEST_BUS_NUM_REPLICAS,
FACT_EVENT_EMITTER). Use INGEST_QUEUE (memory|postgres) and the `worker`
command instead. TENANT_ID is retained (no longer NATS-subject-charset-validated).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a path_prefix filter to semantic search, honoured on both the MCP tool and
the dense-only visualization/API paths through the shared filter contract.
- build_base_filter_conditions: append FieldCondition(file_path,
MatchText(path_prefix)) when set. file_path is only on doc_type == "file"
points, so a non-empty path_prefix implicitly restricts to files.
- Promote path_prefix to an explicit keyword param on the SearchAlgorithm ABC
and both algorithms; thread it through nc_semantic_search (blank ⇒ no filter),
the /api/v1 search endpoints, and the viz route.
- Add a file_path TEXT payload index to _PAYLOAD_INDEX_FIELDS (no content
re-index; idempotent startup migration). MatchText tokenizes on server Qdrant
and matches by substring on local/embedded qdrant-client — both serve folder
scoping.
- Update ADR-027 (Phase 2 implemented; readiness table; semantics note). Tests.
Refs ADR-027 Phase 2. Deck #177.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a modified_after/modified_before date-range filter to semantic search,
honoured on both the MCP tool path (BM25HybridSearchAlgorithm) and the
dense-only visualization/API path (SemanticSearchAlgorithm) through one shared
contract.
- Promote modified_after/modified_before to explicit keyword params on the
SearchAlgorithm ABC and both concrete algorithms; factor the shared
placeholder+ownership+doc_type+date filter into
access_filter.build_base_filter_conditions so new filters land in one place.
- nc_semantic_search: accept RFC 3339 / ISO 8601 (or Unix seconds) bounds via
utils.validation.parse_modified_timestamp; Annotated/Field constraints on the
numeric args; explicit McpError guard for after > before. Thread the parsed
bounds through the cross-app and per-doc_type dispatch.
- /api/v1 search endpoints + viz route parse the same formats and 400 on bad or
inverted ranges.
- Add a modified_at INTEGER payload index to _PAYLOAD_INDEX_FIELDS; the
idempotent _ensure_payload_indexes() startup path migrates existing
collections with no content re-index.
- Update ADR-027 to resolve the review feedback (validation placement, shared
algorithm contract, deferral of nc_semantic_search_answer, payload index,
RFC-3339-at-the-boundary rationale). Add unit tests.
Refs ADR-027. Deck #177.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Follow-up to PR #814 review.
NatsStatusSubscriber.run() called task_status.started() *after* the fallible
pull_subscribe, so a NATS broker that wasn't ready when the MCP server started
would crash the lifespan instead of retrying. Bus status is a non-critical
observability path, so:
- signal started() before the first subscribe (semantics: "loop is running",
not "subscription succeeded");
- retry a failed subscribe with backoff instead of propagating;
- on a real fetch error (not an idle timeout) drop the subscription and
re-subscribe rather than fetching against a possibly-dead handle.
Also anchor the _content_hash etag-threading TODO to the PR #814 review thread
so it is discoverable outside git blame.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- gateway_client: guard token cache with a lazy anyio.Lock so concurrent
embed calls share one M2M token request instead of racing
- status subscriber: distinguish idle fetch timeouts from real broker
errors (log + 5s backoff) instead of swallowing all and spinning
- nats: warn when the bus URL uses unencrypted transport (non-tls://)
- collection_metadata: accept an optional shared httpx client, make TLS
verify explicit, document the unauthenticated control-plane contract
- replace python -O-stripped asserts with explicit ValueError in the bus
status builder and the api metadata source
- document why the nil-UUID sentinel point can't collide with content ids
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
SonarCloud:
- Resolve 6 S5332 hotspots (http→https in test fixture URLs).
- S6418: hoist the unauthenticated AsyncOpenAI placeholder to a named constant
+ NOSONAR (genuine non-secret; gateway ignores it when unauthenticated).
- Fix two reliability bugs: None-index guard in the gateway token-cache test
(S2259) and float `> 0.0` instead of `!= 0.0` in the sentinel test (S1244).
- status.py idle path sleeps 0.1s instead of sleep(0) (S7491); NOSONAR on the
protocol-required async no-await aclose() stubs (S7503).
Claude review:
- Remove three leftover debug print() calls in app.py (logger.info already
covers them).
- payload_backfill: drop parsed_at from the backfilled-keys docstring (it is
per-document state, not a deployment scalar); add a clean 404 precondition
for BasicAuth deployments without an OAuth token verifier.
- status.py: task_status typed TaskStatus | None (drop type: ignore).
- nats.py: TODO to thread etags for file/deck/news; note etag default → None.
- factory: warn on unknown INGEST_BUS_URL scheme; raise ValueError instead of
assert for the external-mode preconditions.
- docs/configuration.md: document the decomposition hook-point env vars + that
nats-py ships core (lazy-imported) and external+bus uses two NATS connections.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds the seven §10.2 hook-point modules + five env vars so Astrolabe Cloud can
offload document processing to the external document-processor / embedding
gateway. Purely additive: with every setting unset the server behaves exactly
as today, so self-hosters are unaffected (Deck #92).
Hook points (all default to current monolith behavior):
- config: EMBEDDING_PROVIDER, INGEST_MODE, STATUS_BACKEND,
COLLECTION_METADATA_SOURCE, FACT_EVENT_EMITTER (+ supporting settings),
validated in Settings.__post_init__ (fail-fast STATUS_BACKEND=local with
INGEST_MODE=external); shared canonical.py.
- vector/payload_keys.py + acl_hash.py: cross-impl NAMESPACE/point_id (§2.2)
and BLAKE2b-128 ACL hash (§11), pinned by fixtures shared with the
document-processor repo.
- embedding/gateway_client.py: OpenAI-compatible GatewayProvider authenticating
via M2M OIDC client-credentials (separate realm); manual-only registry entry.
- vector/collection_metadata.py: sentinel-point / API metadata source with env
fallback.
- vector/queue/: hexagonal ingest producer ports + memory/NATS adapters
(Postgres seam); INGEST_MODE=external publishes mcp.ingest.requested.{tenant}
instead of the in-memory stream and skips the in-process processor pool. The
lifespan becomes a composition root across both deployment branches.
- vector/queue/status.py: STATUS_BACKEND=bus subscriber feeding a StatusStore
the vector-sync status endpoint reads.
- admin/payload_backfill.py: POST /api/v1/admin/payload-backfill (admin scope);
processor writes the new payload keys; query-side ACL pre-filter gated behind
ACL_PREFILTER_ENABLED (default off).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
When the per-tenant nextcloud-mcp-server Pod OOMKills mid-batch, the
in-memory anyio processor queue is lost but the placeholder Qdrant
points (is_placeholder=true, status=pending) survive. The next Pod's
scanner re-runs, sees the existing placeholders, applies the
5 × VECTOR_SYNC_SCAN_INTERVAL staleness gate (~5h with the deployed
1h scan interval), and skips them. Result: 0 documents indexed for
the duration of the gate after every restart.
Stamps a process-level instance_id (UUID per Pod-process) onto every
placeholder write. A new sweep_orphan_placeholders helper, called
once from starlette_lifespan after the Qdrant client is initialised
and before the scanner / user-manager spawns, scrolls the collection
and deletes any placeholder whose instance_id doesn't match the
current Pod's (including placeholders with no instance_id field —
back-compat for pre-fix Pod versions). The scanner's next cycle
naturally re-creates fresh placeholders and queues work normally;
no DocumentTask reconstruction needed.
Sweep is one-shot at startup, not periodic — the existing staleness
gate still covers same-Pod recovery, and the cross-Pod-restart gap
was the only failure mode. Failure is non-fatal (logged via
vector_sync.orphan_sweep_failed) so a transient Qdrant hiccup at
boot doesn't prevent the scanner from running.
Both lifespan branches (single-user BasicAuth, OAuth / multi-user
BasicAuth) call the sweep via a module-local helper. A new
VECTOR_SYNC_ORPHAN_SWEEP_ENABLED setting (default True) provides
an escape hatch.
Closes Deck #101.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Round 17 reviewer (🟡 Important):
1. docs/configuration.md degraded-migration runbook said `doc_id backfill
failed on …` but the actual log line in qdrant_client.py:415 is
`doc_id backfill scroll failed on …`. Operators grepping the runbook
string would have missed it. Insert the `scroll` qualifier.
2. _create_one_payload_index returned True on the 400 schema-conflict
path, so a wrong-type index discovered at create time skipped the
consolidated `Payload index creation incomplete` summary — but a
wrong-type index discovered via the existing-schema check at line
195-206 did fire it. Tenants whose payload_schema is hidden from
their JWT (Qdrant Cloud collection-scoped tokens) only ever observe
the create-time path, so they never saw the operator-level summary.
Return False so the summary fires in both cases.
3. docs/configuration.md said the upgrade-time delay was `proportional to
point count while writes are issued` — overstating the cost. Writes
are proportional to int-typed points only; the scroll itself is
proportional to total point count. Reword.
Local-mode collection-creation regression (root-cause of failing
single-user / login-flow / multi-user-basic CI jobs):
PR #779 changed the existence probe in get_qdrant_client from
collection_exists() (returned bool in both modes) to get_collection()
+ except UnexpectedResponse(status_code=404). The HTTP-mode client
raises UnexpectedResponse with a 404 body, but the local/in-memory
client raises ValueError(f"Collection {name} not found") — see
qdrant_client/local/async_qdrant_local.py. The narrow except clause
let the ValueError propagate, app.py's lifespan re-raised as
RuntimeError, and the mcp container crashed on first start. Catch
ValueError too, with a `not found` substring guard so genuine
programming bugs (bad collection_name, etc.) still surface.
Tests: extend the existing 400-path test to assert the new
failed_fields contract; add two get_qdrant_client unit tests pinning
the local-mode VE catch (positive case + propagation case).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Detect pre-existing payload indexes with the wrong schema type in
`_ensure_payload_indexes`. The previous "field already in
existing_schema → skip" branch silently survived a collection migrated
from the int-doc_id era where `doc_id` is indexed as INTEGER, letting
`MatchValue(value="123")` searches keep failing with HTTP 400 on Qdrant
Cloud strict mode — exactly the production failure this PR was meant to
fix. New behaviour: compare `existing_schema[field].data_type` against
the declared type; on mismatch log a WARNING and append to
`failed_fields` so the consolidated end-of-function summary picks it up.
No auto-repair (operator intervention only — see docs/configuration.md
recovery procedure). New test exercises the doc_id-INTEGER scenario
end-to-end and asserts both the per-field WARNING and the summary line.
Clarify the `_verify_news_items` malformed-doc_id rationale: the news
API has no per-item endpoint, so a malformed doc_id genuinely cannot be
verified against the source of truth. We err toward false-positive
(keep) over false-negative (drop) — same conservative posture as
`_verify_notes` and `_verify_deck_cards`. The producer-side validation
is the real security boundary; the verifier is defence-in-depth. Both
the inline comment and the WARNING message now spell this out.
Add a TODO in `get_last_indexed_timestamp` flagging the O(N) cost on
every incremental sync tick. The previous single-page `limit=10_000`
silently bounded the scroll; paginating fixed correctness but made the
unbounded cost visible. The follow-up tracker (canonical TODO at
`api/visualization.py`) covers migrating the max-`indexed_at` to a
sentinel point or collection metadata for O(1) lookup.
Consolidate the duplicate non-numeric-doc_type TODOs at
`api/visualization.py:508` and `auth/viz_routes.py:570` into a single
canonical comment in `visualization.py`; `viz_routes.py` is reduced to
a back-reference. Removes the rot risk of "fixed in one place,
forgotten in the other." The canonical comment also references the
O(1) timestamp follow-up in `scanner.py`.
Document the `batch_size = 256` (qdrant_client.py) vs
`_DELETION_TRACKING_PAGE_SIZE = 1024` (scanner.py) split with
cross-referencing comments at each site: the smaller batch is for the
read-write backfill upsert path (Qdrant accepts ~256-point chunks
comfortably); the larger page is for read-only deletion-tracking
scrolls where no per-page write round-trip applies.
Replace `assert qdrant_client is not None` in `scan_user_documents`
with `cast(AsyncQdrantClient, qdrant_client)` plus an explanatory
comment. `assert` is silently elided under `-O`; `cast` is the
conventional zero-cost narrower for branches the type checker can't
infer from the surrounding `if not initial_sync` ternary.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Add chunk_start_offset / chunk_end_offset to _PAYLOAD_INDEX_FIELDS so
the legacy offset-based fallback in search/context.py works on Qdrant
Cloud strict mode (pre-#75 clients have no chunk_index payload).
- Cover chunk_index / chunk_start_offset / chunk_end_offset in the
payload-index summary test; refresh the stale field-list comment.
- Flag the is_valid_nextcloud_doc_id gate at both chunk-context handler
sites with a TODO for future non-numeric doc_types.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- _group_int_doc_ids: use type(value) is not int instead of isinstance,
since bool is an int subclass and would otherwise stringify to
"True"/"False" and corrupt legacy payloads on backfill.
- Replace doc_id.isdigit() guards in 5 boundary sites
(api/visualization, auth/viz_routes, search/context note/news_item/
deck_card branches) with a shared is_valid_nextcloud_doc_id helper
that rejects "0", leading zeros, and Unicode digit classes
(superscripts, Arabic-Indic, Devanagari) which pass isdigit() but
cannot be valid MySQL AUTO_INCREMENT IDs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three coordinated fixes flagged as Important in the round-10 review of
PR #773:
1. Index chunk_index. The chunk-context fast path in
_get_chunk_by_index_from_qdrant and get_chunk_bbox_and_page_from_qdrant
filters on chunk_index, but the field was absent from
_PAYLOAD_INDEX_FIELDS. On Qdrant Cloud strict mode every chunk-context
lookup via chunk_index would 400 and silently fall back to the
document re-fetch path — the exact failure mode the chunk_index
shortcut exists to avoid. Added as INTEGER schema.
2. Catch raw network errors in _ensure_payload_indexes. The
create_payload_index loop only caught UnexpectedResponse, so an
httpx.ConnectError or asyncio.TimeoutError mid-loop would propagate
uncaught — leaving _qdrant_client assigned and silently skipping all
remaining fields. Added a broad Exception catch with the same
per-field containment as the 5xx path: log at ERROR with exc_info,
append to failed_fields, continue. New test covers the path.
3. Lazy-initialise _qdrant_init_lock. Constructing anyio.Lock() at
module import time works for the asyncio backend but anyio's docs
advise instantiating synchronization primitives within an async
context, and pyproject.toml's anyio_mode = "auto" means tests can
run under trio. Moved the construction into get_qdrant_client; safe
under cooperative multitasking because there is no await between the
None-check and the assignment.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add an .isdigit() guard at the top of both chunk-context handlers so a
non-numeric doc_id fails fast with a clear 400 ("doc_id must be numeric,
got 'abc'") rather than silently bottoming out as a 404 from deep inside
get_chunk_with_context. The earlier int(doc_id) coercion was removed when
doc_id became a pure pass-through to Qdrant's keyword payload index, which
also dropped this boundary validation.
Also align test_backfill_emits_progress_log_every_20_batches' scroll stub
with real Qdrant: next_offset is now "next-1" (str) instead of 1 (int),
matching the sibling test_backfill_rewrites_int_doc_ids_to_str. Pure
stub-fidelity fix; production code already treats next_offset as opaque.
Addresses both 🟡 Important items from PR #773 review round 10.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Addresses the four 🟡 important findings from claude-bot review on PR #773:
str (non-Optional) and the guard would silently skip the Qdrant lookup
for an empty string. Removing the guard matches the type signature.
(`all([…, doc_id, …])` rejects None and empty string, plus
`assert doc_id is not None`). No code change needed.
`get_qdrant_client()` with a module-level `anyio.Lock`. Double-checked
locking keeps the steady-state hot path lock-free. Without this,
parallel cold-start callers could all enter the init block and run
`_backfill_doc_id_to_string` + `_ensure_payload_indexes` redundantly
(idempotent, but noisy). Pattern matches `auth/storage.py:2071`.
behavior with three tests covering the float-warning path (the gap
called out in the review), the str/None silent-skip paths, and the
int-grouping happy path.
Verification:
- ruff check / format: clean
- ty check -- nextcloud_mcp_server: clean
- uv run pytest tests/unit/: 969 passed
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two unrelated CI failures on this branch, one fix each:
- tests/integration/test_deck_vector_search.py: pass str(card.id) to
get_chunk_with_context. The function's contract is doc_id: str
(keyword-indexed in Qdrant), and real callers (viz_routes.py URL
path, server/semantic.py via str(result.id)) all stringify. The
test was the only int caller, hitting the .isdigit() guard added
earlier on this branch.
- tests/server/login_flow/test_login_flow_integration.py:
test_check_status_provisioned now accepts scopes=None as valid.
Per ProvisionStatusResponse in models/auth.py, None is the
documented sentinel for "all scopes granted" — and the web
provisioning path (provision_routes.py, used by Astrolabe's
"Enable Semantic Search" flow exercised by the new regression test
added on this branch) stores exactly that. The previous
is-not-None assertion hid behind test order until that flow ran.
- Replace anyio.sleep(0) with anyio.lowlevel.checkpoint()
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Skip and warn instead of stringifying floats / unexpected types in the
backfill helper. A stray doc_id=3.0 would otherwise be rewritten to
"3.0", which producers (str(int)) and the keyword index would never
match, and which int() on the verification side would reject. Also add
a doc_id=0 case to the backfill test to guard against a future
falsy-skip regression.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Reviewer findings (1 blocking + 2 important):
- 🔴 Replace `import asyncio` / `await asyncio.sleep(0)` with
`import anyio` / `await anyio.sleep(0)` in the four async side-effect
helpers (_scroll_raises, _upsert_raises, _get_collection_raises,
_create_index). CLAUDE.md mandates anyio for all async operations;
conftest pins the backend to asyncio so the asyncio.sleep call worked
today, but the inconsistency would surface the moment that pin moves.
- 🟡 Replace the sentinel's zero dense vector with a single non-zero
element (`[1e-9] + [0.0] * (dimension - 1)`). Cosine distance is
mathematically undefined for the zero vector and Qdrant Cloud strict
mode rejects zero-vector upserts. The exact value doesn't matter
(sentinel never participates in a search — no user_id/doc_id/doc_type
payload) but the upsert itself must be valid.
- 🟡 Avoid the duplicate `get_collection` round-trip on every restart.
`_ensure_payload_indexes` now accepts an optional
`existing_schema: dict | None` parameter; when None it fetches
collection_info itself (and the get_collection-failure swallow still
applies), but `get_qdrant_client` already fetches collection_info
for dimension validation in the existing-collection branch — pass
`collection_info.payload_schema or {}` through to skip the second
call. The new-collection branch passes `existing_schema={}`
explicitly since a freshly created collection has no payload schema.
The 🟡 deck_card iteration-fallback finding doesn't apply: the
`isdigit()` guard at context.py:612 returns early before either the
fast-path or the iteration fallback runs, so non-numeric doc_ids
cannot reach the inner `c.id == int(doc_id)` comparison.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
