How swebi/orderly works
Project Type and Purpose
Orderly is a full‑stack, user‑centric scheduling app. Its purpose is to let users log in with Google, save their class timetable, and then automatically scrape SRM Academia each midnight to push updated events into Google Calendar via background jobs.
Main Architectural Components
a. Frontend (Vercel‑hosted React app)
• Pages & Components (Login, Dashboard, TimetableForm, PrivateRoute, etc.)
• UI library (Shadcn / MagicUI) + routing + state/hooks (useAuthenticate, use-toast)
• HTTP client (axios wrapper) calling the backend API
b. Backend (AWS EC2–hosted Express server)
• REST API Layer
– Controllers: auth.controller (Google OAuth), api.controller (subject/timetable CRUD), admin.controller (manual trigger, job control)
– Middlewares: userVerification, adminVerification, errorHandler
– Routes: /auth, /api, /admin
• Job Queue & Workers
– BullMQ configuration (queue.js, procedure.js)
– Worker process (worker.js)
– BullBoard UI (bullBoard.js) for real‑time monitoring
• Web Scrapers (Puppeteer scripts in /scrapers)
– dayOrder.js, timetable.js, planner.js
• Business Logic & Helpers
– Google API utilities (googleUtils.js)
– Data transformation (generateEvent, timeConverter)
– Crypto, Prisma client for Postgres
c. Datastores & External Services
• PostgreSQL (via Prisma ORM) – persistent storage of users, subjects, timetables
• Redis – BullMQ’s backing store for job queues
• External HTTP services:
– SRM Academia website (Puppeteer scraping)
– Google OAuth & Calendar APIs
Relationships & Data Flows
1. User → Frontend: interactive UI in browser
2. Frontend → Backend (HTTP/JSON): login, CRUD operations, manual trigger
3. Backend → PostgreSQL (Prisma): read/write user data, timetables
4. Backend → Redis: enqueue scraping/calendar‑sync jobs
5. BullMQ Worker → Redis: fetch job, run scraper (Puppeteer) → generate events → call Google Calendar API
6. BullBoard ← Backend: serve dashboard on /admin/queues
Architectural Patterns & Principles
• Separation of Concerns:
– UI vs. API vs. background jobs
• Layered Architecture:
– Controllers → Services/Utils → Data Access
• Asynchronous Processing:
– BullMQ for reliable, retryable, scheduled tasks
• Modularization:
– /api, /scrapers, /bull, /utils each encapsulate distinct logic