Dear Thomas, Edward, and the Verba Team,
First of all, thank you very much for creating Verba. This tool looks really good.
I like how the code is structured, the way you have split the components, and how simply yet efficiently you have organized the user interface.
I recently cloned the Verba repository and have been experimenting with it to determine if this tool might be applicable for my use case.
I have a few questions, and any help you can provide in finding the answers would be very much appreciated.
My goal is to be able to split the RAG data for different users so that user “Bob” might have his set of documents separate from “Alice’s” set of documents. I am reading the Weaviate documentation, and the first feature that seems suitable for my use case is Weaviate multitenancy. It looks very promising.
To try Verba with multitenancy, I have made the following updates:
- Updated the dependency in the docker-compose.xml file from semitechnologies/weaviate:1.24.2 to semitechnologies/weaviate:1.26.0, enabling the use of multitenancy in Verba.
- Updated the Verba code everywhere it manipulates classes (now called collections in Python Client v4 according to your documentation) and introduced a “test” tenant to see whether this approach works in practice.
I noticed that Verba currently uses the previous version of the Weaviate client library, “Python Client v3.” It appears that new multitenancy features were implemented in Weaviate after version 1.25+. Therefore, I wanted to ask your opinion on whether it is worth the time to try the new multitenancy features with the older “Python Client v3” library, or it is mandatory to migrate the code to the new “Python Client v4” before. What is your perspective on this? Additionally, do you think that multitenancy is the correct solution for my use case?
I also wanted to ask whether you have any plans to implement “multi-user” or “multi-tenant” capabilities in Verba.
Thank you very much for any reply.
Cheers,
Mikhail