This has gota be simple and im just missing it.

All the samples show building the vectorStoreIndex.from_documents, but what's the right way to have take an existing already built index and simply create the object so that you can do other things like, index.as_query_engine()

Is it just me, or is there something about the llamaindex git repo that makes forking/cloning it really slow?

Has anyone tried building a bot similar to kapa.ai? I’m curious if what they do is achievable purely / mostly - by storing the documentation in markdown directly, thus allowing the bot to respond to queries with markdown code.

I want to do something similar for a hobby project, currently the source corpus isn’t markdown, but I want nicely formatted responses and special things like class names called out with code tags.

Anywho, any thoughts or references I should look into are appreciated

Has anyone tried writing a more involved node parser? It seems like you'd probably get much better results this way, for example. I've been thinking about having a MarkdownNodeParser than understands headings and such, and breaks on those before it splits on chunks, and perhaps keeps the section hierarchy you're currently nested within as metadata text, tbd there.