Smart Campus Assistant – MVP

This repository contains a minimal but working skeleton for the Smart Campus Assistant MVP:

• Backend: FastAPI, async SQLAlchemy, basic RAG stub using sentence-transformers + FAISS.
• Frontend: Single-page React UI (via CDN) with Tailwind CSS and a clean chat interface.

1. Folder structure

• backend/
  • app/ – FastAPI application
    • main.py – FastAPI app + routing
    • config.py – settings and environment configuration
    • db.py – async SQLAlchemy engine & session
    • models.py – database models (timetables, bus, events, faculty, exams, FAQs)
    • schemas.py – Pydantic models for API responses & chat payloads
    • rag.py – minimal RAG engine + FAISS index handling
    • routers/core.py – main REST endpoints and /api/chat
  • requirements.txt – backend dependencies
• frontend/
  • index.html – React + Tailwind chat UI (CDN-based)
• datasets/ – place your CSV/JSON/PDF source data here (to be wired into RAG + DB import scripts)

2. Running the backend (dev)

From the backend directory:

cd backend
python -m venv venv  # if you don't already have one
venv\Scripts\activate  # on Windows
pip install -r requirements.txt

uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

The FastAPI docs will be available at:

• http://localhost:8000/docs
• http://localhost:8000/redoc

Key endpoints:

• GET /api/health – health check
• GET /api/timetable – timetable entries (filters: program, semester, section)
• GET /api/bus_schedule – bus routes
• GET /api/events – campus events
• GET /api/faculty_directory – faculty list
• POST /api/chat – chatbot endpoint

For now, tables are auto-created on startup using SQLite (smart_campus.db). You can switch to Postgres later by changing database_url in app/config.py.

3. Running the frontend (dev)

The frontend is intentionally very light for the MVP:

cd frontend
python -m http.server 5173

Then open:

• http://localhost:5173

The UI will talk to http://localhost:8000/api/chat by default. Make sure the backend is running.

4. Next steps – plugging in real campus data

1. Prepare datasets (CSV/JSON) for:
   • Timetables, bus schedules, events, exam schedules, faculty directory, FAQs.
2. Write small import scripts that:
   • Read CSVs from datasets/
   • Map rows into ORM models in app/models.py
   • Insert into the DB using an AsyncSession.
3. Build the RAG index:
   • Extract text from PDFs / notices into plain text chunks.
   • Call RAGEngine.build_index(documents) with (text, source_id) tuples.
   • Persisted FAISS index will then be used automatically by /api/chat.

Once those are in place, the MVP will answer queries both via structured tables and retrieved campus documents from the RAG pipeline.