Combine Everything

main.py

# Previous Code (Unchanged)

def combine_segments(segment_files, output_file_path):

    # How to concatenate files in FFMPEG: https://trac.ffmpeg.org/wiki/Concatenate
    
    input_list = []
    filter_list = []

    for idx, file_name in enumerate(segment_files):
        input_list.append(f"-i file:{file_name}")
        filter_list.append(f"[{idx}:v][{idx}:a]")

    filter_list.append(f"concat=n={len(segment_files)}:v=1:a=1[v][a]")

    input_args = " ".join(input_list)
    filter_ags = "".join(filter_list)

    ffmpeg_cmd = \
        f"ffmpeg {input_args} -filter_complex '{filter_ags}' -map '[v]' -map '[a]' -vsync 2 file:{output_file_path}"

    return os.system(ffmpeg_cmd)


def main():
    file_path, input_language = sys.argv[1:]
    input_language = str(input_language).lower()
    transcript = transcribe_audio(file_path=file_path)
    segments = transcript.segments

    # Cut out the first segment from the input video for voice cloning
    initial_segment = segments[0]
    initial_segment_path = get_video_segment_path(file_path, initial_segment)
    cut_segment(file_path, initial_segment.start, initial_segment.end, initial_segment_path)
    voice_object = clone_voice(initial_segment_path)
    delete_files([initial_segment_path])

    processed_segments = []

    for idx, segment in enumerate(segments):
        translated_text = translate_segment(segment.text, input_language)

        audio_file_path = get_audio_file_path(file_path, segment, input_language)
        generate_speech(translated_text, voice_object, audio_file_path)

        video_segment_path = get_video_segment_path(file_path, segment)
        cut_segment(file_path, segment.start, segment.end, video_segment_path)

        before_segment_path = get_before_video_segment_path(video_segment_path)
        if idx == 0 and segment.start > 0.00:
            cut_segment(file_path, 0.00, segment.start, before_segment_path)
            processed_segments.append(before_segment_path)
        elif idx > 0:
            cut_segment(file_path, segments[idx - 1].end, segment.start, before_segment_path)
            processed_segments.append(before_segment_path)

        dubbed_video_segment = layer_video_audio(video_segment_path, audio_file_path)
        processed_segments.append(dubbed_video_segment)

        if idx == len(segments) - 1 and segment.end < get_media_file_length(file_path):
            after_segment_path = get_after_video_segment_path(video_segment_path)
            cut_segment(file_path, segment.end, get_media_file_length(file_path), after_segment_path)
            processed_segments.append(after_segment_path)

        delete_files([audio_file_path, video_segment_path])


    dubbed_video_file_path = get_dubbed_video_file_path(file_path, input_language)
    combine_segments(processed_segments, dubbed_video_file_path)
    delete_files(processed_segments)

    return dubbed_video_file_path


if __name__ == "__main__":
    main()