download_youtube_extract_random.py

import youtube_dl
import sys
import os

sys.path.append('.')

if len(sys.argv) <= 1:
	print("Usage: {} URL-OR-ID [OUTPUT_DIR] [SECONDS-BETWEEN]".format(__file__))
	exit()

url = sys.argv[1]
output_dir = os.path.join('.', 'output') if not len(sys.argv) == 3 else sys.argv[2]
seconds_between = 10 if not len(sys.argv) == 4 else int(sys.argv[3])

if not url.startswith('http'):
	url = 'http://www.youtube.com/watch?v=' + url

if not os.path.exists(output_dir):
    os.makedirs(output_dir)

ydl = youtube_dl.YoutubeDL({
	'ffmpeg_location': os.path.join('.', 'ffmpeg.exe'),
	'format': 'bestvideo+bestaudio',
	'outtmpl': os.path.join(output_dir, '%(id)s.%(ext)s')
	})

with ydl:
	result = ydl.extract_info(url, download=True)
	video = result['entries'][0] if 'entries' in result else result

filename = os.path.join(output_dir, "{}.{}".format(video['id'], video['ext']))
if not os.path.exists(filename):
	filename = os.path.join(output_dir, "{}.{}".format(video['id'], 'mkv'))

cmd = "ffmpeg -i {} -vf fps=1/{} {}".format(
	filename, seconds_between,
	os.path.join(output_dir, 'screenshot%04d.jpg')
	)

print(cmd)
os.system(cmd)