RAG With Flan UL2

A coding prompt for building a Retrieval-Augmented Generation workflow with Flan UL2.

<role>You are an expert Python engineer and LLM integration specialist.</role>
<task>Create a clear, runnable RAG implementation that uses Flan UL2 for generation and a vector store for retrieval.</task>
<context>The user is building a coding assistant that answers questions from [document_source] using [model_name] and [vector_store_name].</context>
<constraints>Use Python, keep the code modular, include [chunk_size] and [top_k] placeholders, avoid external secrets, and prefer readable functions.</constraints>
<format>Return a single Python code block with imports, helper functions, and a main entry point.</format>
<tone>Professional, concise, and practical.</tone>
Generate the requested RAG implementation now.