Does anyone have llama-index working with groq?

Hi, is there a way I can predict how many API calls a given query is going to use before I send it? I am presently watching a list index / accumulate query go over 214 nodes and it is taking a very long time (hours) and has used more than 214 api calls (which I was not expecting and worries me).

There are loads of structured PDF extraction tools. A recent one is https://www.e2enetworks.com/blog/nougat-neural-optical-understanding-for-academic-documents

When I load an index from storage, I want to see some simple statistics about it--number of bytes, number/name of docs, created/modified dates, etc. How can I do that?

example:

[for this node?] using the context provided by the index summary [previously calculated], identify all the distinct narrative scenes in this passage, along with all the characters who participate in each scene.