fix: normalize empty tool results for LlamaIndex to prevent loop #464
+38
−2
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Adapts AG-UI tool result messages for LlamaIndex by converting empty results to a canonical value ("ok").
Prevents unnecessary re-invocation of tools in LlamaIndex workflows when tools return no meaningful payload.
|
I cut a prerelease version ( attn @tylerslaton @ranst91 |
brandonmcconnell
requested review from
mme,
ataibarkai,
maxkorp and
NathanTarbert
as code owners
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
TL;DR:
Adapts AG-UI tool result messages for LlamaIndex by converting empty results to a canonical value ("ok").
Prevents unnecessary re-invocation of tools in LlamaIndex workflows when tools return no meaningful payload.
Explanation
I did some testing and found that whenever I returned a value from the useCopilotAction hook's handler, the big wouldn't occur (had to do a ton of testing to avoid false positives).
This leads me to think that LlamaIndex handles empty tool result values as falsy in a sense, or a failure, and reattempts the request, tool call and all. Returning a value prevents this.
To patch this in @ag-ui/llamaindex, I reconfigured the LlamaIndexAgent class to intercept those empty values and return content: "ok" in place of the empty or undefined value.
Possible gotcha⚠️
If there are cases where a failed tool call would result in an undefined/empty value, we would need to be able to distinguish them from these successful tool calls that return the same value.