YouTube playlist to transcription for accessibility and ai assisted processing etc

yt_playlist_2_transcription.py

import subprocess
import os
import re

# Fast draft by Kilian Lindberg
# Purpose: Accessibility, ease transcript processing for AI in case of limited visual video interaction limitation
# Disclaimer: Code as is, no claims but an inspirational aspiration

# Install yt-dlp using pip
subprocess.run(['pip', 'install', 'yt-dlp'])

# URL of the playlist
playlist_url = 'https://youtube.com/playlist?list=PLf-tvI4zeKLIaHFaIfvla03TAedPHUsln'

# Command to get video URLs in the playlist
command = ['yt-dlp', '--flat-playlist', '-i', '--print-to-file', 'url', 'file.txt', playlist_url]
subprocess.run(command)

with open('file.txt', 'r') as file:
    links = file.read().splitlines()

transcript_file = 'YouTube-transcripts.txt'
if os.path.exists(transcript_file):
    os.remove(transcript_file)

for link in links:
    video_id = link.split('=')[-1]
    transcript_command = ['yt-dlp', '--write-auto-sub', '--skip-download', '--sub-lang', 'en', '-o', '%(id)s.%(ext)s', link]
    subprocess.run(transcript_command)
    transcript_filename = f'{video_id}.en.vtt'
    try:
        with open(transcript_filename, 'r') as tf:
            transcript = tf.read()
            raw_text = re.sub(r'\d{2}:\d{2}:\d{2}.\d{3} --> \d{2}:\d{2}:\d{2}.\d{3}.*\n', '', transcript)  # Remove time codes
            raw_text = re.sub(r'<[^>]+>', '', raw_text)  # Remove HTML tags
            raw_text = raw_text.strip()
            if raw_text:
                with open(transcript_file, 'a') as af:
                    af.write(raw_text)
                    af.write('\n\n')
    except FileNotFoundError:
        print(f'Transcript for {link} not found.')

print('All transcripts have been collected in YouTube-transcripts.txt')