IssueTracker

google-cloud-speech 1.2.0 Beta
Python 3.5

With streaming_recognize, interim_results=True and enable_speaker_diarization=True I expect to get a WordInfo list for the top alternative for every recognition result. However, I only get a word list for the final result and not for interim results.

from google.cloud.speech_v1 import enums
from google.cloud.speech_v1 import SpeechClient
from google.cloud.speech_v1 import types
client = SpeechClient()
config = types.StreamingRecognitionConfig(
    config=types.RecognitionConfig(
        enable_speaker_diarization=True, ...
    ),
    interim_results=True
)
request = types.StreamingRecognizeRequest(audio_content=b'...')
requests = [request]
for response in client.streaming_recognize(config, requests):
    if response.results:
        result = response.result[0]
        if result.alternatives:
            alternative = result.alternative[0]
            if alternative.words:
                print(result.is_final)

Console output is always: True

This seems to be backend related and not an issue of the Python library. (See https://github.com/googleapis/google-cloud-python/issues/9509)

I'm aware of the fact that interim results may change. Still, I need to do speaker detection on the fly and can't wait for the final result to determine if there was a speaker change.