Retrieve documents using BM25 text scoring.
Usage
DuckDBDocumentStore.retrieve_by_bm25_score( query, top_k=10, )