fix(pydantic-ai): set output.value for plain text responses#2944
Open
ritiztambi wants to merge 1 commit intoArize-ai:mainfrom
Open
fix(pydantic-ai): set output.value for plain text responses#2944ritiztambi wants to merge 1 commit intoArize-ai:mainfrom
ritiztambi wants to merge 1 commit intoArize-ai:mainfrom
Conversation
When a PydanticAI agent uses the default `output_type=str`, the LLM responds with a plain text message (no `final_result` tool call). The `_extract_from_gen_ai_messages()` function was only setting `output.value` when a `final_result` tool call was present, leaving it unset for plain text responses. This caused downstream consumers (e.g. Langfuse) that rely on `output.value` to show null output for the most common PydanticAI usage pattern. The fix captures the first assistant text part as `output_value`, mirroring how `input.value` is derived from the last user text part. If a `final_result` tool call is also present, it still takes precedence (overwrites `output_value`), preserving existing behavior for structured output agents.
dosubot
bot
added
the
size:XS
Contributor
|
CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅ |
Author
|
I have read the CLA Document and I hereby sign the CLA |
Author
|
recheck |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes #2943
Summary
output_type=str, the LLM responds with plain text (nofinal_resulttool call).output.valuewas never set in this case, causing downstream consumers (e.g. Langfuse) to show null output.output_valuein_extract_from_gen_ai_messages(), mirroring howinput.valueis derived from the last user text partfinal_resulttool call is also present, it still overwritesoutput_value, preserving existing behavior for structured output agentsChange
One-line addition in
semantic_conventions.py— after yieldingllm.output_messages.*.message.contentfor a text part, also setoutput_valueif not already set.