Phase B (Insights & Alerts) PR4-11 in one drop. Builds on the schema +
service skeletons committed in PRs 1-3.
PR4 Analytics dashboard — 4 chart types (funnel/timeline/breakdown/source),
date-range picker (today/7d/30d/90d), CSV+PNG export per card.
PR5 Alert rail UI + /alerts page — topbar bell w/ live count, dashboard
right-rail, three-tab page (active/dismissed/resolved), socket-driven
invalidation. Bell lazy-loads list on popover open to keep cold pages
fast in non-dashboard routes.
PR6 EOI queue tab on documents hub — filters to in-flight EOIs, count
surfaces in tab label.
PR7 Interests-by-berth tab on berth detail — replaces the stub.
PR8 Expense duplicate detection — BullMQ job runs scan on create, yellow
banner on detail w/ Merge / Not-a-duplicate, transactional merge
consolidates receipts and archives the source.
PR9 Receipt scanner PWA + multi-provider AI — port-scoped /scan route in
its own (scanner) group with no dashboard chrome, dynamic per-port
manifest, OpenAI + Claude provider abstraction, admin OCR settings
page (port-level + super-admin global default w/ opt-in fallback),
test-connection endpoint, manual-entry fallback when no key is
configured. Verify form always shown before save — no ghost rows.
PR10 Audit log read view — swap to tsvector full-text search on the
existing GIN index, cursor pagination, filters for entity/action/user
/date range, batched actor-email resolution.
PR11 Real-API tests — opt-in receipt-ocr.spec (admin save+test, optional
real-receipt parse via REALAPI_RECEIPT_FIXTURE) and alert-engine
socket-fanout spec gated behind RUN_ALERT_ENGINE_REALAPI. Both skip
cleanly without their gate envs so CI stays green.
Test totals: vitest 690 -> 713, smoke 130 -> 138, realapi +2 opt-in.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The dashboard and residential interest smoke tests were intermittently
failing with the page rendering empty/skeleton state. Root causes:
1. ui-store persisted currentPortId/Slug, but those are URL-derived state.
After login lands on /<first-port-by-name>/dashboard, localStorage holds
that port. Hard-navigating to /port-nimara/... rehydrated the store with
the stale id, and useQuery fired with the wrong port before
PortProvider's URL-sync useEffect could correct it. Drop both fields
from partialize — PortProvider re-derives them from the route every
navigation.
2. apiFetch's slug-to-port fallback fired N parallel /api/v1/admin/ports
calls when N components mounted simultaneously with an empty store.
Dedupe in-flight lookups so a stampede collapses into one round-trip.
Also tightened four flaky smoke tests that depended on a fixed 3s wait or
non-waiting isVisible({timeout}) — replaced with expect(...).toBeVisible
or expect.poll so they handle dev-mode JIT cold-start delays cleanly.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Failures were mostly stale selectors, not product regressions:
- .or() traps matching the topbar "+ New" button → use specific names
(Add Webhook, New Field, New Template)
- broad /create|add|new/ patterns → same fix
- [role="dialog"] overlay matched before content → getByRole('dialog').last()
- locator('input') picked hidden Radix Select inputs → getByPlaceholder /
getByRole('combobox', { name })
- 11-global-search rewritten for the inline topbar search (the cmdk
CommandDialog the old tests targeted was replaced)
- missing .first() causing strict-mode failures on notifications heading,
version history text, nav links
- dashboard landing test: no h1 exists, target KPI text instead
- activity-feed: items aren't anchors; match action badge text
- monitoring data-leak check scoped to <main> (sidebar has Email/Documents)
- admin API without port context returns 400 (not 403) for non-admins —
accept 400 as a valid "blocked" status in the sales-agent test
Also dropped dead imports and unused locals surfaced by lint-staged.
Full suite: 124 passed (11.2m).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
