utils.py

def convert_segments_object_to_text(data):
    result = []

    for segment in data['segments']:
        words = segment['words']
        segment_speaker = segment.get('speaker', None)
        segment_start = segment.get('start', None)
        segment_end = segment.get('end', None)
        current_speaker = None
        current_start = None
        current_end = None
        current_text = []

        # Forward fill speaker, start and end if missing
        for i, word_info in enumerate(words):
            if 'speaker' not in word_info:
                if i > 0 and 'speaker' in words[i - 1]:
                    word_info['speaker'] = words[i - 1]['speaker']
                elif i < len(words) - 1 and 'speaker' in words[i + 1]:
                    word_info['speaker'] = words[i + 1]['speaker']
                else:
                    word_info['speaker'] = segment_speaker
                    
            if 'start' not in word_info:
                if i > 0 and 'end' in words[i - 1]:
                    word_info['start'] = words[i - 1]['end']
                else:
                    word_info['start'] = segment_start
                    
            if 'end' not in word_info:
                if i < len(words) - 1 and 'start' in words[i + 1]:
                    word_info['end'] = words[i + 1]['start']
                elif i == len(words) - 1:
                    word_info['end'] = segment_end
                else:
                    word_info['end'] = word_info['start']

        for word_info in words:
            word = word_info.get('word', '')
            start = word_info.get('start', None)
            end = word_info.get('end', None)
            speaker = word_info.get('speaker', None)

            if current_speaker is None:
                current_speaker = speaker
                current_start = start

            if speaker == current_speaker:
                current_text.append(word)
                current_end = end
            else:
                # Finish current segment
                if current_start is not None and current_end is not None:
                    formatted_text = f'{current_speaker} ({current_start} : {current_end}) : {" ".join(current_text)}'
                else:
                    formatted_text = f'{current_speaker} : {" ".join(current_text)}'
                result.append(formatted_text)

                # Start new segment
                current_speaker = speaker
                current_start = start
                current_end = end
                current_text = [word]

        # Append the last segment
        if current_text:
            if current_start is not None and current_end is not None:
                formatted_text = f'{current_speaker} ({current_start} : {current_end}) : {" ".join(current_text)}'
            else:
                formatted_text = f'{current_speaker} : {" ".join(current_text)}'
            result.append(formatted_text)

    return '\n'.join(result)