binael/AfricanFoodBot

AI-powered chatbot for answering questions about African foods, recipes, ingredients, and cooking techniques. # AfricanFoodBot A conversational Flask application for exploring African recipes and cuisine. ## Overview AfricanFoodBot combines a retrieval-backed recipe store with an LLM-powered assistant. Users can ask questions about ingredients, preparation steps, regional dishes, and cooking techniques, and receive responses grounded in recipe metadata. ## Features - Flask web interface with a simple chat UI - Recipe retrieval using ChromaDB - Prompt generation and response delivery using Google Gemini via the GenAI client - Local persistence of indexed recipe content in `chromadb_data/` ## Repository structure ``` app.py chromad_db_retriever.py food_bot.py llm.py prompts.txt README.md recipes.json requirements.txt chromadb_data/ static/ templates/ ``` ## Getting started ### Prerequisites - Python 3.10+ recommended - `pip` installed - A Google API key for the Gemini model if you want LLM responses ### Installation ```bash python -m venv venv venv\Scripts\activate pip install -r requirements.txt ``` ### Configuration Create a `.env` file in the project root with the following value: ```env GOOGLE_API_KEY=your_api_key_here ``` ### Run the application ```bash python app.py ``` Open `127.0.0.1` in your browser to start chatting. ## Usage - The home page provides a chat box for asking about African recipes. - The backend exposes `/api/chat`, which accepts JSON payloads like: ```json { "message": "How do I make jollof rice?" } ``` - Successful responses return: ```json { "response": "..." } ``` ## Notes - `app.py` gracefully handles missing backend dependencies by falling back to a stub response, so the UI can still be demonstrated without a complete integration. - `chromad_db_retriever.py` stores indexed recipe data in the `chromadb_data/` folder. - `recipes.json` contains the recipe source data used for retrieval. ## Recommended improvements - Add more recipes to `recipes.json` for broader knowledge coverage - Enhance prompt design in `ll …