Last active
January 1, 2023 18:31
-
-
Save vadimkantorov/0f4a2060f1dc245e6ad0e019b99d59ee to your computer and use it in GitHub Desktop.
Finds recursively all videos in a directory, and produces a webpage with a frame for every 30 minutes of video. A Python primer of gathering basic video information using ffprobe, and reading JPEG-encoded frames using ffmpeg.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| # Usage: python thumbnail_video_collection.py /path/to/video/dir > thumbnail.html | |
| import os | |
| import sys | |
| import json | |
| import subprocess | |
| import datetime | |
| import argparse | |
| parser = argparse.ArgumentParser() | |
| parser.add_argument('video_dir') | |
| parser.add_argument('--skip_seconds', type = int, default = 30 * 60) | |
| parser.add_argument('--thumb_size', default = '320x240') | |
| parser.add_argument('--video_extension', default = ['mp4'], action = 'append') | |
| args = parser.parse_args() | |
| video_paths = sorted([os.path.join(root, file) for root, _, files in os.walk(args.video_dir) for file in files for ext in args.video_extension if file.endswith(ext)]) | |
| print '<html><body><table border="1px">' | |
| for video_index, video_path in enumerate(video_paths): | |
| print >> sys.stderr, video_index + 1, '/', len(video_paths), os.path.basename(video_path) | |
| ffprobe = subprocess.check_output(['ffprobe', '-v', 'quiet', '-print_format', 'json', '-show_streams', video_path]) | |
| frame_width, frame_height, duration = [(stream['width'], stream['height'], int(float(stream['duration']))) for stream in json.loads(ffprobe)['streams'] if stream['codec_type'] == 'video'][0] | |
| print '<tr>' | |
| print ''.join(map('<td style="white-space:nowrap">{0}</td>'.format, ['#%d' % (video_index + 1), 'W%d' % frame_width, 'H%d' % frame_height, datetime.timedelta(seconds = duration)] + video_path.split('/'))) | |
| for ss in range(0, duration - args.skip_seconds, args.skip_seconds): | |
| jpeg = subprocess.check_output(['ffmpeg', '-v', 'quiet', '-ss', str(ss), '-i', video_path, '-s', args.thumb_size, '-an', '-f', 'image2pipe', '-c:v', 'mjpeg', '-q:v', '1', '-frames:v', '1', '-']) # this explicit iteration with fast seeks is 10x faster than "-skip_frame nokey -vf fps=1/1800" | |
| print '<td><img src="data:image/jpeg;base64,%s" /></td>' % jpeg.encode('base64').replace('\n', '') | |
| print '</tr>' | |
| print '</table></body></html>' |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment