KrishiChakra

Multi-Agent RAG model for Crop Rotation Solutions using Satellite Data inputs from ISRO's Bhuvan and AI-Powered Agricultural Management System with RAG-based Expert Recommendations

Overview

KrishiChakra is an intelligent agricultural management platform designed to help Indian farmers optimize crop rotation strategies using AI-powered recommendations based on ICAR research, IEEE papers, and agricultural handbooks.

Key Capabilities:

AI-Driven Crop Rotation Planning - 3,944 document knowledge base
Farm Management Dashboard - Track fields, crops, yields
Voice Input Support - Multilingual support for farmers
Profit Estimation - Market-based revenue calculations
Soil Health Monitoring - Track benefits over rotation cycles
Mobile-Friendly Interface - Responsive design for field use
Blockchain Integration - GrainTrust certification system

Features

1. Intelligent Crop Rotation Planning

Generate 1-5 year rotation plans based on:
- Soil type (Clay, Loam, Sandy, etc.)
- Climate zone (Tropical, Subtropical, Temperate)
- Field size and location
- Previous crop history
AI recommendations backed by:
- 13 ICAR agricultural handbooks
- 30 IEEE research papers
- Proven crop rotation principles

2. RAG-Powered Knowledge System

3,944 document chunks indexed with FAISS
Local embeddings (all-MiniLM-L6-v2) - No API limits
LLM generation (GPT-4o-mini) - Research-based answers
Sub-second query response time

3. Farm Management

Multi-field management
Crop tracking (Kharif, Rabi, Summer seasons)
Yield monitoring
Profit/loss analysis
Historical data visualization

4. User Experience

Clean, modern UI with Tailwind CSS
Dark/Light theme support
Mobile-responsive design
Voice input for text fields
Multilingual support (Hindi, English)

Architecture

┌─────────────────────────────────────────────────────────────┐
│                     Frontend (Next.js)                      │
│                    Port: 3000/3001                          │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐  ┌──────────┐  │
│  │Dashboard │  │ Auth     │  │ Custom   │  │Knowledge │  │
│  │  Pages   │  │ System   │  │  Plans   │  │   Base   │  │
│  └──────────┘  └──────────┘  └──────────┘  └──────────┘  │
└────────┬────────────────────────┬───────────────────┬──────┘
         │                        │                   │
         ▼                        ▼                   ▼
┌─────────────────┐    ┌──────────────────┐    ┌──────────────┐
│   Backend API   │    │   RAG AI Service │    │  Blockchain  │
│   Port: 8000    │    │   Port: 8001     │    │  GrainTrust  │
├─────────────────┤    ├──────────────────┤    ├──────────────┤
│ • Auth (JWT)    │    │ • FAISS Index    │    │ • NFT Certs  │
│ • Farm CRUD     │    │ • 3,944 Chunks   │    │ • QR Codes   │
│ • PostgreSQL    │    │ • Local Embed.   │    │ • Traceability│
│ • Voice Input   │    │ • LLM (GPT-4o)   │    │              │
│ • Analytics     │    │ • Knowledge Base │    │              │
└─────────────────┘    └──────────────────┘    └──────────────┘

Component Breakdown:

Component	Technology	Purpose
Frontend	Next.js 14, TypeScript, Tailwind	User interface, routing, state management
Backend	FastAPI, SQLAlchemy, PostgreSQL	API, database, authentication
RAG Service	FastAPI, FAISS, Sentence Transformers	AI knowledge retrieval & generation
Blockchain	Ethereum, Hardhat, Solidity	Certification & traceability

Tech Stack

Frontend

Framework: Next.js 14 (App Router)
Language: TypeScript 5.0
Styling: Tailwind CSS, shadcn/ui
State Management: React Context API
HTTP Client: Fetch API
Auth: JWT tokens, secure cookies

Backend

Framework: FastAPI 0.115
Language: Python 3.11
Database: PostgreSQL / SQLite
ORM: SQLAlchemy 2.0
Auth: JWT (python-jose)
Validation: Pydantic v2

RAG AI Service

Framework: FastAPI 0.115
Embeddings: Sentence Transformers (all-MiniLM-L6-v2)
Vector DB: FAISS (IndexFlatIP)
LLM: OpenAI GPT-4o-mini via api.a4f.co
Document Processing: PyMuPDF, NumPy
Knowledge Base: 43 documents (13 handbooks + 30 IEEE papers)

Blockchain

Network: Ethereum (Sepolia testnet)
Smart Contracts: Solidity 0.8.x
Framework: Hardhat
Library: Ethers.js

Getting Started

Prerequisites

Node.js 18+ and npm/pnpm
Python 3.11+
PostgreSQL 14+ (or SQLite for development)
Git

Installation

1. Clone Repository

git clone https://github.com/Kushal-Raj-G-S/KrishiChakra.git
cd krishichakra

2. Setup Frontend

cd frontend
npm install
# or
pnpm install

# Copy environment file
cp .env.example .env.local

# Start development server
npm run dev
# Frontend runs on http://localhost:3000

3. Setup Backend

cd backend

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Setup database
python setup_database.py

# Start server
uvicorn main:app --reload --port 8000
# Backend runs on http://localhost:8000

4. Setup RAG AI Service

cd rag_model

# Create virtual environment
python -m venv venv
venv\Scripts\activate  # On Linux/Mac: source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Start RAG API server
python api_server.py
# RAG service runs on http://localhost:8001

# Or use the batch file (Windows)
start_api.bat

Environment Variables

Frontend (`.env.local`)

NEXT_PUBLIC_API_URL=http://localhost:8000
NEXT_PUBLIC_RAG_API_URL=http://localhost:8001

Backend (`.env`)

DATABASE_URL=postgresql://user:password@localhost/krishichakra
SECRET_KEY=your-secret-key-here
ALGORITHM=HS256
ACCESS_TOKEN_EXPIRE_MINUTES=30

RAG Service (already configured in code)

# config.py or local_rag.py
API_KEY = "your-api-key-here"
BASE_URL = "https://api.a4f.co/v1"
EMBEDDING_MODEL = "all-MiniLM-L6-v2"  # Local, no API needed
PRIMARY_LLM = "provider-5/gpt-4o-mini"

Project Structure

krishichakra/
├── frontend/                    # Next.js Frontend
│   ├── app/                    # App Router pages
│   │   ├── dashboard/         # Dashboard pages
│   │   ├── auth/              # Authentication
│   │   ├── knowledge/         # Knowledge base
│   │   └── page.tsx           # Landing page
│   ├── components/            # React components
│   │   ├── ui/               # shadcn/ui components
│   │   ├── dashboard-*.tsx   # Dashboard components
│   │   └── protected-route.tsx
│   ├── contexts/              # React contexts
│   │   ├── auth-context.tsx
│   │   └── app-context.tsx
│   ├── lib/                   # Utilities
│   │   ├── auth.ts
│   │   ├── rag-service.ts    # RAG API client
│   │   ├── supabase.ts
│   │   └── types.ts
│   └── package.json
│
├── backend/                     # FastAPI Backend
│   ├── app/
│   │   ├── api/              # API routes
│   │   │   ├── auth.py
│   │   │   ├── farms.py
│   │   │   └── voice.py
│   │   ├── models/           # Database models
│   │   │   └── database.py
│   │   ├── services/         # Business logic
│   │   └── config.py         # Configuration
│   ├── main.py               # FastAPI app entry
│   ├── requirements.txt
│   └── setup_database.py
│
├── rag_model/                   # RAG AI Service
│   ├── api_server.py          # FastAPI server (Port 8001)
│   ├── local_rag.py           # RAG system (FAISS + Local embeddings)
│   ├── extract_ieee.py        # PDF extraction script
│   ├── test_api_live.py       # API testing
│   ├── start_api.bat          # Windows startup script
│   ├── requirements.txt
│   ├── files/                 # Knowledge base documents
│   │   ├── ICAR-*.txt        # Agricultural handbooks (13)
│   │   └── *.txt             # IEEE papers (30)
│   ├── faiss_index_local/     # FAISS vector database
│   │   ├── index.faiss       # 3,944 document chunks
│   │   └── metadata.pkl      # Document metadata
│   └── ieee_papers/           # Original PDF papers
│
├── blockchain/                  # Smart Contracts (Optional)
│   ├── contracts/
│   │   └── GrainTrust.sol
│   ├── scripts/
│   └── hardhat.config.js
│
├── README.md                    # This file
├── QUICK_START.md              # Quick setup guide
├── RAG_MIGRATION_SUMMARY.md    # RAG system documentation
└── TEST_RAG_INTEGRATION.md     # RAG testing guide

API Documentation

Backend API (Port 8000)

Authentication

POST /api/auth/register
POST /api/auth/login
GET  /api/auth/me

Farm Management

GET    /api/farms/              # List user's farms
POST   /api/farms/              # Create new farm
GET    /api/farms/{id}          # Get farm details
PUT    /api/farms/{id}          # Update farm
DELETE /api/farms/{id}          # Delete farm

Fields

GET    /api/fields/             # List fields
POST   /api/fields/             # Create field
GET    /api/fields/{id}         # Field details

RAG AI API (Port 8001)

Query Knowledge Base

POST /api/v2/query
Content-Type: application/json

{
  "question": "What are the best crop rotation practices for wheat?",
  "user_id": "user123",
  "top_k": 5,
  "verbose": false
}

Response:
{
  "question": "...",
  "answer": "Based on ICAR guidelines...",
  "sources": ["ICAR-Kharif.txt", "wheat-rotation.txt"],
  "context_chunks": [...],
  "metadata": {
    "total_chunks_in_db": 3944,
    "chunks_used": 5,
    "embedding_model": "all-MiniLM-L6-v2 (local)",
    "llm_model": "provider-5/gpt-4o-mini"
  }
}

Health Check

GET /health
GET /api/v2/stats

RAG System

Knowledge Base Composition

Category	Count	Size	Source
ICAR Handbooks	13	5.4 MB	Agricultural advisories, crop management
IEEE Papers	30	1.6 MB	Research on crop rotation, ML agriculture
Total Documents	43	7.0 MB	-
Total Chunks	3,944	-	2000 chars/chunk, 200 overlap

RAG Pipeline

Document Ingestion
- Extract text from PDFs (PyMuPDF)
- Clean and normalize text
- Chunk into 2000-character segments
Embedding Generation
- Model: all-MiniLM-L6-v2 (384 dimensions)
- Batch processing for speed
- Local execution - No API limits!
Vector Storage
- FAISS IndexFlatIP (cosine similarity)
- Persistent storage
- Fast retrieval (<100ms)
Query Processing
- Embed user query
- Retrieve top-5 relevant chunks
- Generate answer with LLM
- Format for farmer-friendly display

Performance Metrics

Build Time: 80 seconds (3,944 chunks)
Query Time: 2-5 seconds (including LLM generation)
Accuracy: Based on peer-reviewed research
Coverage: 28+ crop varieties, 3+ climate zones

Supported Crops

Cereals: Wheat, Rice, Maize, Sorghum (Jowar), Pearl Millet (Bajra)
Legumes: Chickpea, Green Gram (Mung), Pigeon Pea (Arhar), Lentil (Masoor), Pea, Cowpea, Soybean
Cash Crops: Cotton, Sugarcane
Vegetables: Tomato, Potato, Onion, Cabbage, Brinjal
Oilseeds: Mustard

Testing

Frontend Testing

cd frontend
npm run test        # Unit tests
npm run test:e2e   # E2E tests (if configured)

Backend Testing

cd backend
pytest              # Run all tests
pytest -v          # Verbose output

RAG System Testing

cd rag_model
python test_api_live.py     # Test RAG API
python query_production_rag.py  # Interactive queries

🔧 Troubleshooting

Common Issues

Port 8001 already in use

# Windows
netstat -ano | findstr :8001
taskkill /F /PID <PID>

# Linux/Mac
lsof -ti:8001 | xargs kill -9

RAG API not loading documents

# Rebuild FAISS index
cd rag_model
python local_rag.py  # Ensure reset=True in __main__

Frontend can't connect to backend

Check .env.local has correct API URLs
Verify backend is running on port 8000
Check CORS settings in FastAPI

Contributing

We welcome contributions! Please follow these steps:

Fork the repository
Create a feature branch (git checkout -b feature/AmazingFeature)
Commit changes (git commit -m 'Add AmazingFeature')
Push to branch (git push origin feature/AmazingFeature)
Open a Pull Request

Code Style

Frontend: ESLint + Prettier
Backend: Black formatter, PEP 8
Commits: Conventional Commits format

License

This project is licensed under the MIT License - see the LICENSE file for details.

Team

BioBloom - BMSIT Hackathon Team

Frontend Development
Backend Development
RAG System Architecture
Blockchain Integration
UI/UX Design

Acknowledgments

ICAR - Agricultural research and advisories
IEEE - Research papers on precision agriculture
Hugging Face - Sentence Transformers models
OpenAI - LLM capabilities
shadcn/ui - Beautiful UI components

Support

For issues or questions:

📧 Email: support@krishichakra.com
🐛 Issues: GitHub Issues
📖 Docs: Documentation

Name		Name	Last commit message	Last commit date
Latest commit History 8 Commits
backend		backend
frontend		frontend
rag model		rag model
.gitignore		.gitignore
LICENSE		LICENSE
RAG_MIGRATION_SUMMARY.md		RAG_MIGRATION_SUMMARY.md
README.md		README.md
TEST_RAG_INTEGRATION.md		TEST_RAG_INTEGRATION.md
VOICE_DIALOGUES.md		VOICE_DIALOGUES.md
flatten_conflicts.json		flatten_conflicts.json
flatten_to_move.txt		flatten_to_move.txt
flatten_to_move_filtered.txt		flatten_to_move_filtered.txt

Folders and files

Latest commit

History

Repository files navigation

KrishiChakra

Table of Contents

Overview

Key Capabilities:

Features

1. Intelligent Crop Rotation Planning

2. RAG-Powered Knowledge System

3. Farm Management

4. User Experience

Architecture

Component Breakdown:

Tech Stack

Frontend

Backend

RAG AI Service

Blockchain

Getting Started

Prerequisites

Installation

1. Clone Repository

2. Setup Frontend

3. Setup Backend

4. Setup RAG AI Service

Environment Variables

Frontend (.env.local)

Backend (.env)

RAG Service (already configured in code)

Project Structure

API Documentation

Backend API (Port 8000)

Authentication

Farm Management

Fields

RAG AI API (Port 8001)

Query Knowledge Base

Health Check

RAG System

Knowledge Base Composition

RAG Pipeline

Performance Metrics

Supported Crops

Testing

Frontend Testing

Backend Testing

RAG System Testing

🔧 Troubleshooting

Common Issues

Contributing

Code Style

License

Team

Acknowledgments

Support

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Frontend (`.env.local`)

Backend (`.env`)

Packages