refactor: improve interruption detection word count logic #809
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
- Remove condition that skipped word count check for empty/undefined STT text - Apply minInterruptionWords filtering uniformly to all speech scenarios - Normalize undefined/null transcripts to empty string for consistent handling - Update onEndOfTurn to use same splitWords logic as onVADInferenceDone - Add comprehensive test suite with 23 test cases covering: * Empty and undefined transcript handling * Word count threshold logic * Punctuation and whitespace handling * Integration scenarios between both methods This ensures consistent interruption behavior regardless of transcript content, preventing unwanted interruptions from silence or very short utterances. All 23 tests pass successfully.
|
|
|
…nsive interruption detection tests
…nsive interruption detection tests - Refactored onVADInferenceDone() to normalize undefined/null text and apply word count check consistently - Refactored onEndOfTurn() to use splitWords() for consistent word splitting with onVADInferenceDone() - Added 23 comprehensive unit tests covering all scenarios: empty strings, undefined, short/long speech, thresholds - All tests passing (23/23) - Added SPDX headers to REFACTORING_SUMMARY.md for REUSE compliance
Author
|
please review this @toubatbrian :) |
toubatbrian
reviewed
Nov 10, 2025
| this.cancelPreemptiveGeneration(); | ||
| this.logger.info('skipping user input, new_transcript is too short'); | ||
| return false; | ||
| const wordCount = splitWords(info.newTranscript, true).length; |
toubatbrian
reviewed
Nov 10, 2025
toubatbrian
reviewed
Nov 10, 2025
toubatbrian
reviewed
Nov 10, 2025
toubatbrian
requested changes
Nov 10, 2025
Contributor
toubatbrian
left a comment
toubatbrian
left a comment
There was a problem hiding this comment.
Overall LG! Left some comments
Removed tests for interruption threshold logic.
Removed tests for undefined and null handling normalization.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
Description
This PR resolves an inconsistency in the interruption detection logic for
minInterruptionWordswithin LiveKit agents. Previously, the word count check was skipped for empty or undefined STT transcripts, leading to unwanted interruptions triggered by silence or very short utterances.Now, interruption filtering based on
minInterruptionWordsis uniformly enforced across all speech scenarios, ensuring robust and predictable turn detection regardless of transcript content.Changes Made
minInterruptionWordsfiltering uniformly to all speech scenarios.onEndOfTurnandonVADInferenceDoneto use the samesplitWordslogic.All 23 tests pass successfully.
Pre-Review Checklist
Testing
restaurant_agent.tsandrealtime_agent.tsfunction properly after changes/fix #785