Open source video conferencing app powered by LiveKit. Built with Django and React.
π(summary) relax whisperX payload format
Florent Chehab451be40bb77a
diff --git a/CHANGELOG.md b/CHANGELOG.mdindex 364194c1..1ad163df 100644--- a/CHANGELOG.md+++ b/CHANGELOG.md@@ -32,6 +32,7 @@ and this project adheres to- β¬οΈ(dependencies) update django to v5.2.13 [SECURITY]- π(backend) rely on backend to allow participant update their metadata- π(summary) fix failure webhook notification #1233+- π(summary) relax whisperX payload format #1233## [1.13.0] - 2026-03-31diff --git a/src/summary/summary/core/shared_models.py b/src/summary/summary/core/shared_models.pyindex 340a9959..043ea759 100644--- a/src/summary/summary/core/shared_models.py+++ b/src/summary/summary/core/shared_models.py@@ -9,21 +9,33 @@ class WordSegment(BaseModel):"""Word segment model for transcription tasks."""word: str = Field(title="Word")- start: float = Field(title="Start Time", description="Start time in seconds.")- end: float = Field(title="End Time", description="End time in seconds.")+ start: float | None = Field(+ default=None, title="Start Time", description="Start time in seconds."+ )+ end: float | None = Field(+ default=None, title="End Time", description="End time in seconds."+ )score: float | None = Field(- title="Confidence Score", description="Confidence score for the word segment."+ default=None,+ title="Confidence Score",+ description="Confidence score for the word segment.",)speaker: str | None = Field(- title="Speaker", description="Speaker identifier for the word segment."+ default=None,+ title="Speaker",+ description="Speaker identifier for the word segment.",)class Segment(BaseModel):"""Segment model for transcription tasks."""- start: float = Field(title="Start Time", description="Start time in seconds.")- end: float = Field(title="End Time", description="End time in seconds.")+ start: float | None = Field(+ default=None, title="Start Time", description="Start time in seconds."+ )+ end: float | None = Field(+ default=None, title="End Time", description="End time in seconds."+ )text: str = Field(title="Segment Text", description="Transcribed text for the segment.")@@ -31,7 +43,7 @@ class Segment(BaseModel):title="Word Segments", description="List of word segments within the segment.")speaker: str | None = Field(- title="Speaker", description="Speaker identifier for the segment."+ default=None, title="Speaker", description="Speaker identifier for the segment.")diff --git a/src/summary/tests/models/__init__.py b/src/summary/tests/models/__init__.pynew file mode 100644index 00000000..7c3c4e3c--- /dev/null+++ b/src/summary/tests/models/__init__.py@@ -0,0 +1 @@+"""Tests for models."""diff --git a/src/summary/tests/models/test_whisper_x_response.py b/src/summary/tests/models/test_whisper_x_response.pynew file mode 100644index 00000000..86c33323--- /dev/null+++ b/src/summary/tests/models/test_whisper_x_response.py@@ -0,0 +1,60 @@+"""Tests for WhisperXResponse."""++from summary.core.shared_models import WhisperXResponse+++def test_whisper_x_response_partial_response_is_valid():+ """A partial response from WhisperX is valid.++ Sometimes WhisperX Api doesn't return all the data and timestamps,+ we don't want to fail the whole process.+ """+ WhisperXResponse.model_validate(+ {+ "segments": [+ {+ "start": 1.135,+ "end": 7.3,+ "text": " Test 1, 2, 3... Est-?",+ "words": [+ {+ "word": "Test",+ "start": 1.135,+ "end": 2.216,+ "score": 0.654,+ "speaker": "SPEAKER_00",+ },+ {+ "word": "1,",+ "start": None,+ "end": None,+ "score": None,+ "speaker": None,+ },+ {+ "word": "2,",+ "start": None,+ "end": None,+ "score": None,+ "speaker": None,+ },+ {+ "word": "3...",+ },+ ],+ "speaker": "SPEAKER_00",+ }+ ],+ "word_segments": [+ {+ "word": "Test",+ "start": 1.135,+ "end": 2.216,+ "score": 0.654,+ "speaker": "SPEAKER_00",+ },+ {"word": "1,"},+ {"word": "?"},+ ],+ }+ )