[Question] Implementing a custom vector store for the retrieval module

[williambrach]

I'm looking to replace the current Internet search (e.g., BingSearch) with my own vector store (Qdrant) in the retrieval module. Based on the README, I understand this should be possible by adding an implementation for my vector store in knowledge_storm/rm.py.

Questions:

1. Are there any existing implementations of vector stores for the retrieval module?
2. Can you confirm that adding a custom implementation in knowledge_storm/rm.py is the correct approach? Custom implementation should return documents in the same format as other retrievers.

Any guidance or examples would be greatly appreciated. Thank you!