augo-amos/kenya-travel-agent

AI-powered travel assistant for Kenyan tourism destinations # Kenya Travel Agent An AI-powered travel assistant for Kenyan tourism destinations. This project scrapes travel data from various Kenyan tourism websites and uses LangChain with HuggingFace models to provide intelligent responses about Kenyan travel destinations. ## Features - **Web Scrapers**: Collect data from Kenyan tourism websites - **Vector Database**: FAISS-based storage for efficient similarity search - **AI-Powered Q&A**: Flan-T5 model for natural language responses - **Interactive Chat**: Command-line interface for conversations - **Multi-Destination Support**: Covers Masai Mara, Diani Beach, Amboseli, and more ## Tech Stack - Python 3.12 - LangChain 0.1.0 - HuggingFace Transformers - FAISS Vector Store - Selenium & BeautifulSoup for scraping - Streamlit (optional, for UI) ## Prerequisites - Python 3.12+ - pip (Python package manager) - Git ## Installation 1. **Clone the repository** ```bash git clone github.com cd kenya-travel-agent ``` 2. **Create a virtual environment** ```bash python3 -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate ``` 3. **Install dependencies** ```bash pip install -r requirements.txt ``` ## Data Collection Run the scrapers to collect travel data: ```bash # Basic web scraping python3 scrapers/run_scrapers.py # Advanced scraping with Selenium python3 scrapers/run_advanced_scraper.py ``` ## Build the Knowledge Base Create embeddings and vector store: ```bash python3 models/vector_store.py ``` ## Run the Travel Agent ### Interactive Mode ```bash python3 agent/travel_agent.py ``` ### Single Query Mode ```bash python3 agent/travel_agent.py --query "Tell me about Masai Mara" ``` ## Project Structure ``` kenya-travel-agent/ ├── agent/ │ └── travel_agent.py # Main AI agent ├── scrapers/ │ ├── run_scrapers.py # Basic scrapers │ └── advanced_scraper.py # Selenium scrapers ├── models/ │ └── vector_store.py # V …