How kgand/shellhacks2024 works
Project type and purpose
• This is a full-stack “Lotion” application combining a React Native/Expo mobile frontend with a Python/Flask backend.
• Purpose: AI-powered note classification, query, and AR-based annotation using camera scans.
Top-level components
a. Mobile (frontend)
• App entry points: app.json, app/_layout.tsx, index.tsx
• Pages/flows: (auth) login/register, (extra) onboarding, (pages) home
• UI components: ARCamera, CreateNote, Query, Quiz, Read, SubjectDetails, onboarding/login/register subfolders
• Context: AuthProvider
• Utilities: auth, network, inference, classes
• Configs: Tailwind, Firebase
b. Server (backend)
• Flask app: app.py (defines HTTP endpoints)
• Modules:
– image_processing.py (camera image → text via Google Vision)
– sentiment_processing.py
– database_actions.py (app-level CRUD logic)
– mongo_actions.py (low-level MongoDB operations)
• Data: Test_Data
• Dependencies: requirements.txt
c. Datastores & External services
• MongoDB (persistent note storage)
• Firebase Auth (mobile user authentication)
• Google Vision API (OCR)
• ChatGPT API / HuggingFace (text classification & query)