set_up.py

from ui import *
from file_name import *
from diarization import start_diarization
from transcribe import start_transcribe
from video_tool import convert_video_to_audio, add_subtitle_to_video
import gradio as gr
import re
import os
import utils


def prepare_input(input_file, start_time, end_time, lang, model_size, use_summary, progress=gr.Progress()):
    gr.Info(ui_lang["progress_starting_process"])

    if input_file is None or not os.path.exists(input_file):
        gr.Warning(ui_lang["input_video_warning"])
        return [None, None, [None, None]]
    if validate_time_format(start_time) is False:
        gr.Warning(ui_lang["start_time_warning"])
        return [None, None, [None, None]]
    if validate_time_format(end_time) is False:
        gr.Warning(ui_lang["end_time_warning"])
        return [None, None, [None, None]]
    if lang is None:
        gr.Warning(ui_lang["lang_radio_warning"])
        return [None, None, [None, None]]
    if model_size is None:
        gr.Warning(ui_lang["model_dropdown_warning"])
        return [None, None, [None, None]]

    print(f"SOURCE: {input_file}")
    print(f"AUDIO FILE: {audio_file}")

    # Convert video to audio
    progress(0.2, desc=ui_lang["progress_preparing_video"])
    convert_video_to_audio(
        input_file, audio_file, start_time, end_time)

    # Start diarization
    progress(0.4, desc=ui_lang["progress_acquiring_diarization"])
    start_diarization(audio_file)

    # Start transcribing
    progress(0.6, desc=ui_lang["progress_transcribing_audio"])
    start_transcribe(input_file, lang, model_size, progress)

    # Add subtitle to video
    progress(0.8, desc=ui_lang["progress_add_subtitle"])
    add_subtitle_to_video(input_file, subtitle_file,
                          video_subtitle_file, start_time, end_time)

    # Return video file link, transcribe string, transcribe.txt, subtitle.txt
    transcribe_txt_list, _ = utils.read_transcribe_subtitle_file(
        input_file, False)
    transcribe_txt = "\n".join(transcribe_txt_list)
    return [
        video_subtitle_file,
        transcribe_txt,
        [transcribe_file, subtitle_file]
    ]


def validate_time_format(input_string):
    pattern = re.compile(r'^\d{2}:\d{2}:\d{2}$')
    return pattern.match(input_string) is not None