Build a Simple Python LLM Q&A Prototype for One Document

I need a very small proof-of-concept Python application that allows me to ask questions about the content of a text document. This is only a prototype, not a production application. I will provide one small text/document file. The application should: * Read the provided document * Split the content into manageable chunks * Allow the user to enter a question * Retrieve the most relevant text/chunks * Send the relevant context to an LLM * Return a concise answer based on the document A simple command-line interface or minimal Streamlit interface is sufficient. You may use a lightweight retrieval approach, embeddings/vector search, or another reasonable method. Deliverables: 1. Working Python code 2. requirements.txt 3. Simple instructions for running the application 4. One example showing a question and generated answer Please keep the implementation simple and well organized. No authentication, database, cloud deployment, or advanced UI is required.