Add a complete RAG example

[GabrielBianconi]

[GabrielBianconi]

Start with a brief proposal

[virajmehta]

Proposal: take the Wikipedia dataset from pinecone datasets and use it with either Qdrant or PGVector in a local container. Then run BeerQA against it with 2 functions for query rewriting and generation and then 1 more for evaluation.

Much like a single-iteration version of the work-in-progress beam search example but with a vector store instead of the live Wikipedia API

[GabrielBianconi]

pinecone-datasets is unlicensed so we can't use it as is. We should find something similar, or get permission from them.

Or we just generate them ourselves and share.