def merge_speaker_segments()

in python/ts-to-word.py [0:0]

• 17 lines of code
• 6 McCabe index (conditional complexity)

def merge_speaker_segments(input_segment_list):
    """
    Merges together consecutive speaker segments unless:
    (a) There is a speaker change, or
    (b) The gap between segments is greater than our acceptable level of delay

    :param input_segment_list: Full time-sorted list of speaker segments
    :return: An updated segment list
    """
    outputSegmentList = []
    lastSpeaker = ""
    lastSegment = None

    # Step through each of our defined speaker segments
    for segment in input_segment_list:
        if (segment.segmentSpeaker != lastSpeaker) or \
                ((segment.segmentStartTime - lastSegment.segmentEndTime) >= START_NEW_SEGMENT_DELAY):
            # Simple case - speaker change or > n-second gap means new output segment
            outputSegmentList.append(segment)

            # This is now our base segment moving forward
            lastSpeaker = segment.segmentSpeaker
            lastSegment = segment
        else:
            # Same speaker, short time, need to copy this info to the last one
            lastSegment.segmentEndTime = segment.segmentEndTime
            lastSegment.segmentText += " " + segment.segmentText
            segment.segmentConfidence[0]["text"] = " " + segment.segmentConfidence[0]["text"]
            for wordConfidence in segment.segmentConfidence:
                lastSegment.segmentConfidence.append(wordConfidence)

    return outputSegmentList