"Smart @Docs", meta-index docs, and automatically decide which ones are relevant based on query #3627
Open
2 tasks done
Labels
area:context-providers
Relates to context providers
kind:enhancement
Indicates a new feature request, imrovement, or extension
priority:medium
Indicates medium priority
Validations
Problem
Sometimes having to think of/search the specific keyword for relevant
@Docs
wastes some (not always too much, but at least some) amount of time.Solution
Since the whole purpose of the AI assist system is to spare a few brain-cycles, I figure it would make sense to have some sort of meta-index of all the indexed docs that basically provides the system the gist of what each of the docs is about and what kind of thing it might have inside, sorta similar to how right now within each specific doc it can figure which pages are relevant; and with that, the user would be able to do something like add the
@SmartDocs
"context" and the system would figure out which@Docs
to actually include in the search for specific content based on what the user is talking about in the rest of the prompt.I imagine this hierarchical indexing/search approach would likely be less compute intensive than actually performing the search on all indexed docs.
This wouldn't necessarily replace the current
@Docs
system, since there could be situations where either due to ambiguity, or imperfections in the AI or in the RAG algorithm, the wrong docs might be picked up, or some important ones might get left out; users should still be able to call for specific docs directly like how it's currently done, the SmartDocs context would just be an additional option.ps: Obviously, the name "SmartDocs" is just a placeholder, I'm sure someone can think of something better; I forgot to go sleep many hours ago.
The text was updated successfully, but these errors were encountered: