Rasoi Planner is a shared kitchen workflow app for Indian households. It helps an Owner plan meals and manage pantry inventory, while a Cook updates stock status in real time (including AI-assisted pantry updates).

• Household Owner: plans meals, manages pantry, invites/removes cook access, verifies anomalies.
• Household Cook: checks daily menu, marks inventory as low/out/in-stock, adds quantity notes, uses AI assistant.

• Contributor: develops features, fixes bugs, updates rules/tests.
• QA / Reviewer: validates role behavior, security rules, and end-to-end journeys.

• Frontend: React 19 + Vite + Tailwind
• Backend endpoint: Vercel serverless function at POST /api/ai/parse
• Data/Auth: Firebase Auth (Google) + Firestore
• AI: Gemini model via @google/genai
• Tests:
  • Firestore rules tests via Firebase Emulator
  • Browser E2E via Puppeteer + local mocks

Install these before local setup:

• Node.js 18+ (LTS recommended)
• npm (comes with Node.js)
• Java 17+ (required by Firestore Emulator used in rules:test)
• A Chromium/Chrome browser (for local sign-in popup and E2E)

No global Firebase CLI install is required because firebase-tools is in project devDependencies.

1. Install dependencies:

npm install

2. Create local env file from example:

cp .env.example .env.local

3. Set GEMINI_API_KEY in .env.local.

4. Start the app:

npm run dev

App runs at http://0.0.0.0:3000 (or http://localhost:3000).

Before validating sign-in flows, confirm in Firebase Console:

1. Authentication -> Sign-in method -> Google is enabled.
2. Authentication -> Settings -> Authorized domains includes localhost and 127.0.0.1 (if you use that host).
3. Google popup sign-in works in local browser.
4. Owner can invite/remove cook.
5. Invited cook gets access.
6. Removed cook loses access immediately.

All scripts below are defined in package.json.

• Start dev server:

npm run dev

• Start E2E-focused dev server config:

npm run dev:e2e

• Type-check:

npm run lint

• Production build:

npm run build

• Preview build output:

npm run preview

• Clean build output:

npm run clean

• Firestore rules tests (emulator-backed):

npm run rules:test

• End-to-end tests:

npm run e2e

• Full local verification before release:

npm run verify:local

• households/{householdId}
• households/{householdId}/inventory/{itemId}
• households/{householdId}/meals/{YYYY-MM-DD}
• households/{householdId}/logs/{logId}

Owner permissions:

• create/update/delete inventory
• create/update meals
• invite or remove cook (cookEmail)
• read all household data

Cook permissions:

• read household, inventory, meals, logs
• update inventory status and request quantities
• cannot modify meals or delete inventory

These constraints are enforced in firestore.rules and validated by test/rules/run.ts.

• POST /api/ai/parse

{
  "input": "tamatar khatam ho gaya",
  "inventory": [{ "id": "9", "name": "Tomatoes", "nameHi": "टमाटर" }],
  "lang": "hi"
}

{
  "understood": true,
  "message": "optional",
  "updates": [{ "itemId": "9", "newStatus": "out", "requestedQuantity": "1kg" }],
  "unlistedItems": [{ "name": "Dhania", "status": "low", "category": "Veggies", "requestedQuantity": "2 bunch" }]
}

• Invalid request body: HTTP 400
• Missing GEMINI_API_KEY: HTTP 503
• AI/runtime failures: HTTP 500 with safe fallback message
• Client uses safe fallback message if response parsing/validation fails

• Firestore rules tests entry: test/rules/run.ts
• E2E runner: test/e2e/run.mjs
• E2E mock server config: test/e2e/vite.e2e.config.ts
• E2E summary output: test/e2e/artifacts/summary.json

• Ensure rewrites in vercel.json are preserved: /api/* -> /api/* and /* -> /index.html.
• Set GEMINI_API_KEY in Vercel project environment variables.

• Firestore security rules source: firestore.rules
• Local emulator config: firebase.json
• App uses Firestore named database: ai-studio-3900af62-0bf5-496a-a136-d1c8a0c4b8bd
• Confirm production Firebase Auth domain setup before release (Google provider and authorized domains)

1. Mainline path (default):
   • Merge to main; CI deploys firestore.rules automatically when Firestore files change.
2. CI deploy preconditions:
   • verify-local must pass.
   • npm run rules:target:check must pass (project/database target integrity gate).
3. Post-deploy smoke checks:
   • Owner smoke user must read households/{householdId}/unknownIngredientQueue.
   • Non-member smoke user must receive PERMISSION_DENIED.
4. Emergency/manual deploy only:
   • npm run rules:deploy:prod
   • npm run rules:smoke:prod
5. Optional deploy diagnostics:
   • npm run rules:deploy:prod:dry

This project uses GitHub as the deployment source of truth.

1. Create a feature branch from main.
2. Commit and push branch changes.
3. Open a pull request.
4. Wait for CI check verify-local to pass.
5. Merge PR into main.
6. Vercel auto-deploys merged main commit to production.

• Workflow file: .github/workflows/ci.yml
• Triggers:
  • every pull request
  • every push to main
• Required check name for branch protection: verify-local
• CI command chain:
  • npm ci
  • npm run verify:core
  • verify-core includes npm run rules:target:check
• main push additional automation:
  • Detect Firestore-related file changes.
  • Deploy Firestore rules automatically when changed.
  • Run production smoke test for owner-allow and non-member-deny unknown queue reads.
• E2E is executed as a separate non-blocking CI job (e2e) to avoid blocking production rules deploy on browser flakiness.

• Husky install hook is configured via npm run prepare.
• Pre-push hook path: .husky/pre-push
• Pre-push command: npm run verify:local
• If checks fail, push is blocked.

• Require pull request before merging.
• Require status checks to pass before merging.
• Add required status check: verify-local.
• Require branches to be up to date before merging.
• Add repository secrets for Firestore deploy/smoke workflow:
  • FIREBASE_TOKEN
  • SMOKE_OWNER_EMAIL
  • SMOKE_OWNER_PASSWORD
  • SMOKE_OWNER_HOUSEHOLD_ID
  • SMOKE_NON_MEMBER_EMAIL
  • SMOKE_NON_MEMBER_PASSWORD

• Git repository connected to this GitHub repo.
• Production branch set to main.
• Preview deployments enabled for pull requests.
• GEMINI_API_KEY configured for Preview and Production environments.

• Open Vercel dashboard.
• Find the last known-good production deployment.
• Redeploy that deployment to production.

• Set GEMINI_API_KEY in .env.local (local) or Vercel environment settings (deploy).

• Install Java 17+ and confirm java -version resolves correctly in shell.

• Confirm the visible build id is the latest deployment.
• Validate CI Firestore deploy and smoke test status on latest main run.
• For emergency recovery, deploy + smoke manually:
  • npm run rules:deploy:prod
  • npm run rules:smoke:prod

• Add localhost / 127.0.0.1 to Firebase Auth authorized domains.
• Ensure browser popup blocking is disabled for local app.

• Install Chrome/Chromium, or set PUPPETEER_EXECUTABLE_PATH to a valid browser binary.

• Check browser network call to /api/ai/parse.
• Confirm request has input, inventory, and lang.
• Confirm response passes schema validation (understood, updates, unlistedItems).