The community member is experiencing a ZeroDivisionError when trying to use the BM25Retriever with a Weaviate vector store. The comments suggest that the BM25Retriever requires a docstore or nodes, and when using Weaviate (or other vector databases), the docstore is disabled by default. The community members recommend passing the nodes directly to the BM25Retriever instead. They also discuss the possibility of retrieving all the nodes from the index, and the use of Qdrant for hybrid or fusion search, where the community member can implement their own custom ranking function.
Hi, while trying to use bm25_retriever = BM25Retriever.from_defaults( docstore=index.docstore, similarity_top_k=2 ), getting this error - ZeroDivisionError I am using weaviate as vector store and the loaded index does have a good amount of documents, How to resolve this error?
