🌿 LeafCompass - Smart Farming AI

LeafCompass is an intelligent agricultural assistant designed to empower farmers with data-driven insights. It leverages Machine Learning and Generative AI to diagnose plant diseases, recommend crops, predict yields, and provide real-time farming advice via a chatbot.

🚀 Live Demo

• Frontend (Vercel): https://leafcompass.vercel.app/

• Backend API (Hugging Face): https://jain-mayukh-lc-api.hf.space/docs

✨ Key Features

🍂 Plant Disease Detection

• Input: Upload an image of a crop leaf.

• Model: Convolutional Neural Network (CNN) built with TensorFlow/Keras.

• Output: Identifies the disease name and confidence level.

🌾 Crop Yield Prediction

• Input: Rainfall, Temperature, Soil Type, Region, Weather, etc.

• Model: Random Forest Regressor (Scikit-Learn).

• Output: Predicted yield in kg/ha.

🌱 Crop Recommendation

• Input: Soil N-P-K values, pH, Humidity, Rainfall, State.

• Model: Random Forest Classifier.

• Output: The most suitable crop for the given conditions.

🧪 Fertilizer Recommendation

• Input: Soil stats, Crop type, Moisture.

• Model: Classification Model.

• Output: Recommended fertilizer type.

🤖 AgroBot (AI Chat)

• Engine: deepseek-ai/DeepSeek-V3.2 (via Hugging Face Inference API).

• Function: Answers general farming questions in real-time.

🛠️ Tech Stack

Frontend

• Framework: React.js
• Styling: Tailwind CSS
• Icons: Lucide React
• Routing: React Router DOM
• HTTP Client: Axios

Backend

• Framework: FastAPI (Python)
• ML Libraries: TensorFlow, Scikit-learn, Pandas, NumPy
• AI Integration: Hugging Face Inference Client (huggingface_hub)
• Deployment: Docker (Hugging Face Spaces)

⚙️ Local Installation & Setup

Follow these steps to run the project locally on your machine.

Prerequisites

• Node.js & npm installed.
• Python 3.9+ installed.
• Git installed. 1 . Clone the Repository

git clone https://github.com/your-username/leaf-compass.git
cd leaf-compass

2 . Backend Setup

The backend handles the ML models. Note that you will have to download the large model files from Google Drive.

# Navigate to backend folder
cd backend

# Create a virtual environment (Optional but recommended)
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install Python dependencies
pip install -r requirements.txt

# Run the Server
python main.py

The server will start at http://localhost:8000.

3 . Frontend Setup

Open a new terminal for the frontend.

# Navigate to frontend folder
cd frontend

# Install dependencies
npm install

# Start the React App
npm start

The app will open at http://localhost:3000.

📂 Project Structure

leaf-compass/
├── backend/
│   ├── models/                # ML Models (.h5, .pkl) - Auto-downloaded
│   ├── main.py                # FastAPI entry point & endpoints
│   ├── requirements.txt       # Python dependencies
│   ├── Dockerfile             # Docker config for Hugging Face
│   └── README.md              # Backend specific info
├── frontend/
│   ├── src/
│   │   ├── components/        # Reusable UI (Header, Footer, Forms)
│   │   ├── pages/             # Page Views (Home, Disease, Yield, etc.)
│   │   ├── services/          # API.js (Axios configuration)
│   │   └── App.js             # Main Router
│   ├── tailwind.config.js     # Tailwind settings
│   └── package.json           # Node dependencies
└── README.md                  # Main documentation

🧠 Model Details

Feature	Algorithm	Input Data	Accuracy (Approx)
Disease Detection	CNN (MobileNetV2/Custom)	Leaf Images (224x224)	~92%
Yield Prediction	Random Forest Regressor	Weather & Soil Data	~88% (R2 Score)
Crop Recommender	Random Forest Classifier	NPK, pH, Rainfall	~99%
Fertilizer Recommender	XGBoost/Random Forest	Soil & Crop Data	~95%

🔑 Environment Variables

This project requires a Hugging Face Token for the AI Chatbot.

1 . Locally: You can add HUGGING_FACE_TOKEN in a .env file or hardcode it in main.py (not recommended for public repos).

2 . Hugging Face Spaces: Set it in the "Settings" -> "Variables and secrets" tab.

🚀 Deployment Guide

Backend (Hugging Face Spaces)

The backend is Dockerized to handle large dependencies.

• Create a Space on Hugging Face (SDK: Docker).

• Enable Git LFS if pushing models directly (or use the Google Drive downloader script provided in main.py).

• Push the backend folder contents to the Space.

Frontend (Vercel)

• Push the frontend folder to GitHub.

• Import the repo into Vercel.

• Set the Root Directory to frontend.

• Deploy!

🤝 Contributing

Contributions are welcome! Please follow these steps:

1 . Fork the repository.

2 . Create a new branch (git checkout -b feature-branch).

3 . Commit your changes.

4 . Push to the branch.

5 . Open a Pull Request.

📄 License

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

Made with 💚 for the Farming Community.