Flahti : The Moroccan Farmers Companion Chatbot

🌍 GenAI Challenge : Moroccan agriculture

📖 Project Description

Flahti is an AI-powered chatbot designed to assist Moroccan farmers by providing comprehensive agricultural education and guidance. Flahti offers information on various aspects of farming, including crop selection, pest management, fertilization, water management, and optimal growing conditions. The chatbot aims to enhance farming practices, increase productivity, and promote sustainable agriculture in Morocco.

🌿 Key Features

Flahti provides personalized advice on:

• 🌾 Crop Selection: Region-specific and seasonal crop recommendations.
• 🐛 Pest and Disease Management: Identification and treatment advice.
• 🌞 Optimal Growing Conditions: Insights on ideal environmental factors for various crops.
• 🌱 Fertilization Guidance: Nutrient needs and proper application techniques.
• 💧 Water Management: Efficient irrigation methods and water conservation tips.
• 🧑‍🌾 Farm Management and Planning: Crop rotation and resource planning.
• 🌱 Sustainable Farming Practices: Organic farming and soil health maintenance.
• 📊 Market Information: Up-to-date market prices and demand trends.
• 🏛️ Government Schemes and Support: Information on subsidies, loans, and training programs.

🎥 Demo

The demo showcases:

• Interaction with Flahti.
• Features such as crop selection guidance, pest management advice, and water management tips.

WhatsApp.Video.2024-05-19.at.07.22.11_df7c3ded.mp4

🤖 Trained Models / Baselines or APIs Used

Flahti leverages several trained models and APIs to deliver accurate and relevant advice to farmers. These include:

• 📝 Retriever-Augmented Generation (RAG) with LlamaIndex: For retrieving and generating contextually accurate responses from a large corpus of agricultural data.
• 🧠 Large Language Models (LLMs) Mistral: For natural language processing and understanding.
• 📄 Data Sources: Includes PDF reports from the Food and Agriculture Organization (FAO), World Bank reports, Moroccan Ministry of Agriculture, and research from local agricultural institutions.
• 🦙 Ollama: For using LLMs locally.

🛠️ Technical Stack

Flahti is developed using a robust technical stack to ensure reliability and scalability. The main components include:

• 🔙 Backend: Flask framework for server-side logic.
• 🌐 Frontend: HTML, CSS, JavaScript, and Bootstrap for user interface design.
• 🧠 Language Models: Utilized Mistral for natural language processing.
• 📄 Data Retrieval: RAG with LlamaIndex for effective data retrieval from PDF sources.
• 💻 Local Models: Ollama for using LLMs locally.