sagarsambhwani/RetrievLAB ? reverse-engineered prompt

Build me a research focused retrieval lab in Python that lets me load documents, chunk them, make embeddings, and compare different retrieval methods like dense search and BM25 on the same dataset.

I want the project to be organized in a clean, modular way so each part can be swapped out later, like loaders, chunkers, embedders, retrievers, and evaluation. Please include a simple benchmark dataset, a way to run experiments from the command line, and reports that show retrieval quality with metrics like Recall at K, Precision at K, and MRR.

Keep it practical and simple first, with a solid baseline using linear search before anything more advanced. If you need to look up current docs online for any library details, go ahead. Also add a basic setup so the project is easy to run locally, with tests and a clear entry point. The goal is not a flashy app, it is a small but real experimental framework for understanding retrieval from first principles.