REVAI-4324: Multichannel transcript grouping

[dmtrrk]

Overview

Add new optional parameters to get_transcript_xx functions that are available for multichannel media files:

1. group_channels_by
   Specifies how to group multiple channels in the transcript. This parameter determines the atomic entity for breaking down the transcript into monologues. Only applicable when the submitted media has multiple channels (speaker_channels_count > 1):

• speaker - groups by speakers
• word - groups by individual words
• sentence - groups by complete sentences

2. group_channels_threshold_ms
   Threshold in milliseconds for handling speaker interruptions. When a speaker interrupts another speaker, this parameter determines how to group the segments. Only applicable when the submitted media has multiple channels (speaker_channels_count > 1):

• If the interruption occurs within this threshold, preference is given to the most recent speaker
• If the interruption occurs after this threshold, a new segment is created

Usage

from rev_ai import apiclient, GroupChannelsType

client = apiclient.RevAiAPIClient(token)
job = client.submit_job_local_file(filePath, speaker_channels_count=2)

# default (word, 1000ms)
transcript = client.get_transcript_text(job.id)

# specific WORD params
transcript_word = client.get_transcript_text(job.id, group_channels_by=GroupChannelsType.WORD, group_channels_threshold_ms=1000)

# specific Sentence parameters
transcript_sentence = client.get_transcript_text(job.id, group_channels_by=GroupChannelsType.SENTENCE, group_channels_threshold_ms=2000)

# Speaker parameter
transcript_speaker = client.get_transcript_text(job.id, group_channels_by=GroupChannelsType.SPEAKER)

[alexsku]

i would use type hints if possible

[dmtrrk]

I don't see we use type hints in this code. I consider this to be python 2.x compatibility

[alexsku]

@dmtrrk i think you were saying you have doubts about this, i thiunk this is right, we are dealing with these two parameters independently

[alexsku]

looks good to me