Speaker confidence missing from PrerecordedResponse, and utterance confidence not set
barapa opened this issue · 1 comments
barapa commented
What is the current behavior?
- Speaker confidence is not defined in the PrerecordedResponse model
- When dumping to JSON, confidence field for utterance is not present
Steps to reproduce
- Call with diarization:
options = PrerecordedOptions(
model=MODEL,
language="en",
smart_format=True,
punctuate=True,
paragraphs=True,
diarize=True,
keywords=keywords,
)
source: UrlSource = UrlSource(url=audio_url)
response await self._deepgram.listen.asyncprerecorded.v("1").transcribe_url(
source,
options,
)
- Dump the response
logger.info(response.to_json())
Expected behavior
Both confidence and speaker_confidence to be present in the dumped json, as indicated in the docs
Please tell us about your environment
Python 3.12.1
barapa commented
Sorry, was looking at the wrong field. Closing as this was raised in error.