feat(litellm): capture tool calls in streaming responses #2646
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
Add support for accumulating tool calls from streaming delta chunks in both sync and async streaming handlers. Changes: - Add _accumulate_tool_calls() to collect tool call chunks by index - Add _build_message_from_accumulated() to format messages for attribute extraction - Update _finalize_sync_streaming_span() to accumulate tool calls - Update _finalize_streaming_span() (async) to accumulate tool calls - Add tests for streaming with single and multiple tool calls The implementation follows the pattern from OpenAI instrumentation's _ChatCompletionAccumulator but in a simpler form since LiteLLM already normalizes the streaming format.
dosubot
bot
added
the
size:L
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Outdated suggestions cannot be applied.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.
Summary
Add support for accumulating tool calls from streaming delta chunks in both sync and async streaming handlers for the LiteLLM instrumentation.
Changes
_accumulate_tool_calls()to collect tool call chunks by index_build_message_from_accumulated()to format messages for attribute extraction_finalize_sync_streaming_span()to accumulate tool calls_finalize_streaming_span()(async) to accumulate tool callsContext
The existing streaming handlers (
_finalize_sync_streaming_spanand_finalize_streaming_span) only captureddelta.contentbut misseddelta.tool_calls. This meant tool calls in streaming responses were not being recorded in spans.The implementation follows the pattern from OpenAI instrumentation's
_ChatCompletionAccumulatorbut in a simpler form since LiteLLM already normalizes the streaming format.Testing