Thanks for sharing. Can you share your opinion for my use case? "My use case involves retrieving user data from db(eg.sql),create embeddings for it and should be able to store the data and embeddings in chromadb with ID being the user ID, so that when each user performs a search query through gpt(I am using openAI lib),search should be made only on the corresponding embedding which holds that particular user data(should be able to identify it through user ID)". Note: My document data cannot be a text file since it needs to be fetched from db.