AI Revision Platform

A revision-planning web app: users register subjects and topics, and the platform generates a day-by-day study schedule that spreads their topics evenly across however many days they have before an exam.

Note: the plan generation is currently a deterministic scheduler (even distribution of topics across days), not an LLM-based planner — see generate_plan in backend/app.py.

What it does

• Add user & subjects — POST /adduser stores a user's subjects/topics in Firestore
• Generate a revision plan — POST /generateplan takes an email + number of days, fetches the user's topics, and splits them evenly across the days, saving the result to Firestore
• Fetch a user's saved plans — GET /getplans/<email>
• Lookup a user — GET /getuser/<email>

Tech Stack

Layer	Tools
Backend	Flask, Flask-CORS
Database	Firebase Firestore (via firebase-admin)
Frontend	React (Create React App)
Hosting	Backend on Render, frontend on Vercel

Project Structure

backend/
  app.py        Flask API: user CRUD + plan generation
frontend/
  src/components/
    AddUser.js       Form to register a user's subjects/topics
    GeneratePlan.js  Form to request a revision plan
    ViewPlans.js     Displays saved plans

Testing

Unit tests cover the plan-distribution logic (backend/ml/distribute_topics) — no Firebase credentials needed:

cd backend
pip install pytest
pytest tests/ -v

Tests run automatically on every push via GitHub Actions.

Running locally

Backend

cd backend
pip install -r requirements.txt

Set GOOGLE_APPLICATION_CREDENTIALS_JSON in a .env file to the JSON contents of a Firebase service account key (Firebase Console → Project Settings → Service Accounts → Generate new private key). The app reads this from the environment — do not commit the key file itself.

python app.py

Frontend

cd frontend
npm install
npm start