NFL RAG is a Retrieval-Augmented Generation (RAG) model designed to provide real-time NFL news updates without requiring users to manually search for the latest information. The model extracts news links from the official NFL news page, builds a vector database using the collected articles, and generates the most relevant responses based on user queries while maintaining conversational context.
- Real-Time News Extraction: Automatically retrieves the latest news articles from the NFL official news page.
- Vector Database Storage: Utilizes ChromaDB to store and index extracted articles for efficient retrieval.
- Conversational Memory: Utilizes LangGraph to remember the user's previous interactions, allowing for more natural and contextual conversations.
- Utilizes the BeautifulSoup library to extract and process the most recent articles from the webpage.
- Contains the crawl function, which is responsible for parsing news links from the official NFL news page.
- Builds a vector database using LangChain's Chroma to store and retrieve news articles efficiently.
- The vector database will be constructed in the vector_database directory.
- Implements a RAG model with LangGraph to maintain conversational context and improve response relevance over time.
- Utilizes Streamlit to build a RAG interface.
- To run NFL RAG, ensure you have all the libraries included in
requirements.txtinstalled:pip install -r requirements.txt - Also, you have to create an environment file that contains your LangSmith API key.
- Feel free to delete the outdated 'vector_database' directory after you clone the project.
- Run
streamlit run RAG.pycommand to start the RAG model. - Provide your OpenAI API Key in the sidebar of Streamlit interface.
- Interact with RAG using natural language queries through Streamlit interface.