My question at Github https://github.com/googleapis/python-speech/issues/52 has been active for 9 days and the only two people to have attempted an answer have both failed but now I think it might be possible for someone to answer it who understands how Google Cloud Buckets work even though they do not understand how Google's Speech Api works. In order to convert long audio files to text they first must be uploaded to the Cloud. I was using some syntax that now appears to be broken and the following syntax might work except that Google does not explain how to use this code in coordination with files uploaded to the Cloud. So in the code below published here:
https://cloud.google.com/speech-to-text/docs/async-recognize#speech_transcribe_async-python
The content
object has to be located on the cloud and it needs to be a bytes object. Suppose the address of the object is: gs://audio_files/cool_audio
What syntax would I use such that the content object refers to a bytes object?
from google.cloud import speech
from google.cloud.speech import enums
from google.cloud.speech import types
client = speech.SpeechClient()
audio = types.RecognitionAudio(content=content)
config = types.RecognitionConfig(
encoding=enums.RecognitionConfig.AudioEncoding.LINEAR16,
sample_rate_hertz=16000,
language_code='en-US')
operation = client.long_running_recognize(config, audio)
print('Waiting for operation to complete...')
response = operation.result(timeout=90)
GCS stores objects as a sequence of bytes. If your object has a Content-Encoding header that can cause the content to be transformed while downloading (e.g., gzip content will be uncompressed if the client doesn't supply an Accept-Encoding: gzip header); and if it has a Content-Type header the client application or library may treat the information differently.