llama_index
llama_index copied to clipboard
run-llama
Object index improvements
This PR does a few things
- Makes creating an object index easier (no need to think about mappings in most cases)
- Adds postprocessors to object index retriever (since it is hard to use post processors with agents)
- Makes saving an object index to a backend (qdrant, chroma, etc.) much easier
- Makes object index nodes have static IDs (helps with persistence)
- Updates docs/notebooks
Check out this pull request on 
See visual diffs & provide feedback on Jupyter Notebooks.
Powered by ReviewNB
![review-notebook-app[bot] avatar](https://avatars.githubusercontent.com/in/15474?v=4 "Published by a pub.dev verified publisher"){kind=small}
@jerryjliu we can add an option for that, but that really only makes sense for objects you cam define on the fly (I.e. a query engine tool). This wouldn't really work for arbitrary function tools
The main advantage of backing by a vector store is so that re-creating the index is a no-op 👍🏻
@jerryjliu we can add an option for that, but that really only makes sense for objects you cam define on the fly (I.e. a query engine tool). This wouldn't really work for arbitrary function tools
The main advantage of backing by a vector store is so that re-creating the index is a no-op 👍🏻
yeah that's fair. at small scales it doesn't really matter. it does somewhat negate the value of vector store support
doesn't have to be in this PR! to unblock
@jerryjliu Added FnNodeMapping and an example + test 👍🏻 Should be good to go