i was a bit overwhelmed with all the different indexes. Is there some documentation on when i want to use a certain index over the other?
My usecase is I have 1000s of tweets collected and I want to ask the llm to find ones with certain subjects or summarize tweets based on a certain search criteria.
My 2cnts, if you are working on a single document that is relatively short, like a 2 page pdf, start with the simplest list index. Then move to the simple vector index and you'll get a feeling how the results change. The list index won't scale to a large corpus (or it will but cost is big), if applied naively over your data.
some of this can be for you to experiment with! a keyword index looks up content based on keywords, whereas a vector index looks up stuff based on embedding similarity