- Added getPatientByInsuranceIdAndDob to storage
- Processor uses insuranceId + DOB as unique key instead of insuranceId alone
- Dependent with same subscriber ID but different DOB gets a new patient record
(bypasses createOrUpdatePatientByInsuranceId which would overwrite subscriber)
- Name extraction anchored to "Relationship:" to scope Patient Info column only
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Extract patient first/last name from Patient Information DOM section
(scoped to avoid duplicate Subscriber Information column values)
- Switch to latest tab at start of step2 (Eligibility Identifier opens in new tab)
- DOB: double-click + ActionChains.send_keys (no pyperclip, avoids Chrome crash)
- BCBS MA button changed to variant="default" to match nearby buttons
- Backend processor uses extracted names from selenium result
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- New Selenium worker (fresh Chrome per run, no persistent session)
login → OTP modal → eTools → ConnectCenter → Verification →
New Eligibility Request → fill form (NPI, member ID, DOB) →
Expand All → CDP PDF back to app
- Backend route fetches BCBS_MA credentials + provider NPI from settings
- Frontend OTP modal with 6-digit code entry
- BCBS MA added to insurance credentials dropdown in settings
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Move Select Procedures above Check Eligibility in appointment right-click menu
- Show 3 blank service lines by default when opening Select Procedures with no saved procedures
- Fix serviceLines not being preserved when API returns empty procedures list
- CDT combo buttons no longer auto-fill price (only fill codes); user maps price via Map Price button
- Overlap detection in schedule: shorten earlier appointment display span when a later one starts within its range
- Procedures dialog: replace single manual-add row with 3 pre-filled blank rows grid + Add Line / Save Lines buttons
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Rewrote UnitedDH claim worker to navigate via eligibility page → Selected Patient → Submit Claim button flow
- Updated helpers_uniteddh_claim.py step names to match new 9-step workflow
- Changed payer selection in both eligibility and claim workers to type + Enter
- Updated patient table column header from 'DOB / Gender' to 'DOB'
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Using paymentGroupId label as the page-load wait caused TimeoutException when
the claim flow's page structure differed, returning ERROR and triggering
_minimize_browser (about:blank + minimize). Now waits for the btn-primary
Continue button (original safe indicator) and wraps both dropdown selections
in try/except so the step never fails due to a missed dropdown.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Mirrors the same Provider & Location page pattern from the eligibility worker:
- wait for paymentGroupId label visibility before interacting
- Treatment Location: ng-arrow-wrapper click + ARROW_DOWN/ENTER
- Billing Entity: ng-arrow-wrapper click + ARROW_DOWN/ENTER
- re-find Continue button after selections to avoid stale element
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- United SCO eligibility: add Treatment Location step before Billing Entity
on Provider & Location page; both use ng-arrow-wrapper ActionChains click
+ ARROW_DOWN/ENTER keyboard selection to handle upward-opening panels
- Use visibility_of_element_located for Billing Entity label wait so code
waits for page to fully render after Select Insurance modal closes
- DOB (and all dates) now display as MM/DD/YYYY instead of Mon DD, YYYY
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
DentaQuest stores its MFA trust as a persistent cookie, not in
LocalStorage like DDMA/Azure B2C. Deleting the Cookies file on startup
wiped that trust token, forcing OTP on every app restart. Now only the
credentials tracking file is reset on startup; if credentials change,
_force_logout() still calls delete_all_cookies() for a full reset.
Covers both eligibility and claim flows (both share the same
DentaQuestBrowserManager singleton).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add appointment type categories matching insurance claim form (recall, filling, pedo, dentures, implant, endo, crown, perio, extraction, ortho, consultation, emergency, other)
- Auto-infer appointment type from CDT codes with priority rules (endo > implant > crown > ...)
- typeLocked flag prevents auto-overwrite when user manually sets type
- Show appointment type label and procedure codes on schedule cards
- Background sync on /day route retroactively fixes stale appointment types
- Fix PUT /api/claims/:id to save claimFiles (previously silently dropped)
- Auto-link AppointmentFile records to ClaimFile when claim is created or updated
- Fix D5750 (denture reline) CDT range to map correctly to dentures category
- Fix typeLocked Zod rejection in appointment update route
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Fix Express route ordering in appointments-procedures so /prefill-from-appointment
is matched before /:id (D0140 and other codes now always reach the claim)
- claim-form: always fetch AppointmentProcedure records when an existing claim loads
so post-save procedure edits (e.g. adding D0140) are reflected immediately
- appointments-page: replace sessionStorage with React state (newApptPrefill) for
slot-click prefill so columns B-F correctly carry their staffId into the form
- add-appointment-modal / appointment-form: thread prefillData prop; add
NewAppointmentPrefill interface; useEffect applies values via setValue
- appointments upsert: remove per-patient dedup so the same patient can have
multiple appointments on the same day in the same column
- DentaQuest / TuftsSCO: navigate to about:blank and minimize instead of
quit_driver after each run — session cookie stays in memory so OTP is only
required once per app startup, not on every eligibility or claim check
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- claims.ts: batch-column now routes each patient to the correct portal
(MH/CCA/DDMA/TuftsSCO/UnitedSCO) based on patient.insuranceProvider
- appointments-page.tsx: eligibility badge falls back to patientStatus
for all insurance types, not just MassHealth
- unitedDHClaimProcessor/ddmaClaimProcessor/tuftsSCOClaimProcessor:
auto-save claim PDF when no socketId (batch-column path)
- unitedSCOEligibilityProcessor: unknown eligibility no longer stored as INACTIVE
- queues.ts: add tuftssco-claim-submit and uniteddh-claim-submit to SeleniumJobType
- selenium_UnitedSCO_eligibilityCheckWorker.py:
- step1: add Select Insurance OK click with staleness wait; Provider &
Location page just clicks Continue; skip first/last name input
- step2: extract name from eligibility details tab (ALL CAPS → title case);
skip "not available" guard; fix name regex to match only all-caps words
- .gitignore: ignore all chrome_profile_* directories
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds stop-app.sh to kill all services by port (5000-5003, 3000/3001)
and updates setup-desktop-shortcut.sh to create both Start and Stop
desktop shortcuts in a single run. README Step 13 updated accordingly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Each office sets VITE_CLOUDFLARE_HOST and CLOUDFLARE_HOST in their
local .env files instead of hardcoding the subdomain in source code.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Deleting users with patients/appointments/claims was blocked by FK
constraints. Now reassigns those records to the requesting admin before
deletion, and cleans up user-specific data (backups, cloud files).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Removed per-user patient filtering so all staff accounts see the same
patient pool. Previously each user only saw patients they created.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace the prisma migrate deploy subprocess with a custom
applyMissingMigrations() that reads each migration SQL file and runs it
via prisma.$executeRawUnsafe. This avoids the Prisma CLI aborting when
the restored _prisma_migrations table contains orphaned entries (migration
names that no longer have a corresponding file in the migrations folder).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- After importing a backup, run prisma migrate deploy so any schema
migrations the backup is missing are applied automatically. This
prevents pages from failing due to missing tables/columns when the
backup was taken on an older version of the app.
- Force logout and redirect to login after a successful restore so the
JWT is refreshed against the restored database (prevents userId
mismatch causing user-scoped queries to return empty results).
- Fix getTotalPatientCount() in /status route to pass userId so it
counts only the current user's patients instead of all patients.
- Add prisma.$connect() after $disconnect() to ensure a clean reconnect.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Route DDMA eligibility through InProcessQueue (concurrency=1) so it
queues behind other selenium jobs instead of running concurrently
- New ddmaEligibilityProcessor: starts Python session, polls for OTP/
completion via socket events, saves PDF and updates patient DB
- Frontend ddma-buton-modal now uses shared app socket + job:update
pattern (drops private socket connection)
- SeleniumService: upgrade ddma_browser_manager with credential hash
tracking, anti-detection options, and startup session clearing;
upgrade DDMA worker with firstName/lastName support, PDF via
printToPDF, force-logout on credential change; upgrade helpers with
dual OTP strategy (app API + browser polling); add /clear-ddma-session
endpoint; reduce fixed sleeps with smart WebDriverWait
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Reset public schema before psql restore so existing tables don't
block CREATE TABLE statements (pg_dump without --clean has no DROPs)
- Disconnect Prisma connection pool after restore so next request
gets fresh connections against the restored data
- Auto-create the USB backup subfolder if it doesn't exist instead
of failing; only fail if the destination drive itself is missing
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds ordered setup steps (Node.js, Python, Chrome, PostgreSQL, Redis,
pip dependencies) so a fresh machine can be fully configured from
git clone to running the app without guessing.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
