Smart Supermarket Web IoTC 2025

Smart Supermarket Web IoTC 2025 is an IoT-enabled supermarket platform that combines a web POS system, barcode and weight integration from edge devices, payment and invoice flows, and an AI chatbot service.

Overview

The project is organized as a multi-service repository:

• client-web: React frontend for login, shopping, cart, checkout, and admin dashboard.
• server-web: Node.js + Express API for authentication, products, PLU items, orders, users, settings, and invoice generation.
• server-gemini: Flask service that provides streaming chat responses with Google Gemini.
• pi-model: Raspberry Pi-side scripts and Flask API for barcode scanning and weight reading.
• ram-model: Research/training code for recommendation and ad-insertion models.

Main Features

• User, admin, and guest access flows.
• Barcode-based shopping and PLU/weight-based product handling.
• Cart management and order checkout (cash and MoMo flow).
• PDF invoice download after payment.
• Admin management for products, users, orders, and store settings.
• AI shopping assistant chat (Gemini-powered streaming responses).
• IoT bridge endpoints for receiving scanned barcodes and reading weight from edge devices.

Tech Stack

• Frontend: React, Material UI, Axios
• Backend API: Node.js, Express, MongoDB (Mongoose), JWT
• AI Service: Flask, google-genai, Server-Sent Events (SSE)
• IoT Side: Python, Flask, Raspberry Pi device scripts
• ML Research: Python, PyTorch

Repository Structure

Smart-Supermarket-Web-IoTC-2025/
|-- client-web/
|-- server-web/
|-- server-gemini/
|-- pi-model/
|-- ram-model/
`-- README.md

Prerequisites

• Node.js 18+
• npm 9+
• Python 3.10+
• MongoDB instance (local or Atlas)

Environment Variables

server-web (server-web/.env)

MONGO_URI=your_mongodb_connection_string
JWT_SECRET=your_jwt_secret
PORT=5000

server-gemini (system env or .env)

GEMINI_API_KEY=your_google_ai_key
GEMINI_MODEL=gemini-2.0-flash-exp
PORT=5000
FLASK_DEBUG=false

client-web (for React, typically .env)

REACT_APP_BACKEND_URL=http://localhost:5000
REACT_APP_FLASK_API=http://localhost:5001
REACT_APP_NGROK_URL=http://localhost:5002
REACT_APP_MOMO_SERVER=http://localhost:5000

Run Services Locally

Open separate terminals for each service.

1) Start backend API (server-web)

cd server-web
npm install
node server.js

Backend default URL: http://localhost:5000

2) Start Gemini service (server-gemini)

cd server-gemini
python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS/Linux
# source .venv/bin/activate
pip install -r requirements.txt
python app_flask.py

Recommended URL: http://localhost:5001

3) Start Pi API (pi-model/api)

cd pi-model/api
pip install flask flask-cors pymongo
python api.py

Recommended URL: http://localhost:5002

4) Start frontend (client-web)

client-web source files are present, but package.json is not included in this repository snapshot. Add/restore the frontend package manifest first, then run:

cd client-web
npm install
npm start

API Surface (High Level)

server-web exposes routes under /api:

• /api/auth
• /api/users
• /api/products
• /api/orders
• /api/plus
• /api/admin
• /api/payment
• /api/bill
• /api/settings

pi-model/api exposes:

• POST /scan-barcode
• GET /last-barcode
• POST /start-weigh

server-gemini exposes:

• POST /api/chat (streaming SSE)
• GET /api/test-model

Notes

• Keep all secrets in environment variables. Do not hard-code credentials.
• Use different ports for server-web, server-gemini, and Pi API when running locally.
• The ram-model folder is training/research oriented and not required for basic web app operation.

License

This project is currently provided for academic and demonstration purposes.