How dhanushsaicoder/project-iris works
Project Type and Purpose
• This is a full-stack Progressive Web App (PWA) called Project Iris.
• It provides real-time object detection and depth estimation in the browser (using TensorFlow.js models) and records those sessions in a backend service for review and analytics.
• Admins can manage users and view overall session statistics via a React-based dashboard.
File Structure
• /backend
• Express server (server.js)
• Routes → Controllers → Models (MVC pattern)
• Auth, session and user management
• Uses Mongoose to persist to MongoDB
• /frontend
• React app bootstrapped with CRA + PWA template
• /public/tfjs_models serves TensorFlow.js models (fastDepth, COCO-SSD)
• /src/components, /src/pages, /src/hooks, /src/context → Component-based architecture + Context API for state
• /src/workers/depth.worker.js → Web Worker for heavy depth computation
• Service worker, PWA manifest
• Root
• install_all scripts, plan.md, project_context.md, testing_steps.md
• package.json orchestrates dev:mobile and proxy setups
README for Additional Insights
• Tech Stack
• Backend: Node.js + Express, MongoDB (Mongoose), JWT auth, bcrypt, CORS
• Frontend: React, Material-UI, Recharts, TensorFlow.js (COCO-SSD, fastDepth), Web Worker
• PWA capabilities, offline-first service worker
• Features
Real-time detection, haptic/audio feedback, session tracking, admin dashboard, user settings