AI Smart Crop Advisory Prototype

A demo-friendly prototype of KrishiMitra with the original three-tier architecture:

• Frontend: React + Vite + Tailwind + React Router + Axios
• Backend: Spring Boot (Web, Security + JWT, Data JPA) with endpoints for auth, crop detection, advisory, weather, and market prices
• AI microservice: FastAPI stub with a focused /predict endpoint returning seeded PlantVillage classes
• Database: MySQL for the demo (H2 in-memory for local dev fallback)

Demo flow (2–3 minutes)

1. Register/Login
2. Upload crop image (tomato/potato/corn)
3. See disease result with confidence and treatment
4. View rule-based advisory + weather snapshot
5. Check seeded mandi prices and trend

Quick start (local dev)

Backend

cd backend
mvn spring-boot:run

This backend targets Java 17 and is configured for MySQL by default. Set SPRING_DATASOURCE_URL, SPRING_DATASOURCE_USERNAME, and SPRING_DATASOURCE_PASSWORD for your local database. AI service URL can be set via AI_SERVICE_URL (defaults to http://localhost:8000/predict).

If mvn is not available in PATH on Windows, run Maven with full path:

Set-Location backend
& "C:\Users\admin\.maven\maven-3.9.14\bin\mvn.cmd" spring-boot:run

For backend test and build commands, see backend/README.md.

AI service

cd ai-service
pip install -r requirements.txt
uvicorn main:app --reload --port 8000

The AI service now builds a lightweight color-texture classifier from seeded leaf samples on first use and caches the trained weights at ai-service/disease_model.joblib, so inference runs fully offline.

Frontend

cd frontend
npm install
npm run dev -- --host

Set VITE_API_BASE_URL to point to the backend (default http://localhost:8080).

Testing

Backend tests

cd backend
mvn test

Windows PowerShell with explicit Maven path:

Set-Location backend
& "C:\Users\admin\.maven\maven-3.9.14\bin\mvn.cmd" test

Frontend tests

cd frontend
npm run test:run

Run only the integration test:

cd frontend
npm run test:run -- src/App.integration.test.jsx

Docker Compose demo

docker-compose up --build

Services:

• mysql on 3306
• ai-service on 8000
• backend on 8080 (uses MySQL + AI service inside the network)
• frontend on 80 (proxies /api/* to backend)

API overview

• POST /api/auth/register — create prototype user, returns JWT
• POST /api/auth/login — authenticate and return JWT
• POST /api/crop/detect — secured; sends image + crop to AI service and stores report
• GET /api/advisory?cropType=&city=&season= — rule-based recommendations
• GET /api/weather?city= — normalized weather snapshot using Open-Meteo (with fallback)
• GET /api/market?cropType=&city= — seeded mandi prices + mini trend
• FastAPI POST /predict — returns top disease class with confidence for target crop

Notes

• Crops limited to tomato, potato, and corn for the demo.
• Advisory and market data are lightweight seeded mappings to keep the prototype reliable.
• Multilingual support (English, Hindi, Marathi) for main UI labels.
• JWT is prototype-grade and should be replaced for production.