Skip to content

Instantly share code, notes, and snippets.

@alexeygrigorev
Created September 17, 2016 09:09
Show Gist options
  • Save alexeygrigorev/a1bc540925054b71e1a7268e50ad55cd to your computer and use it in GitHub Desktop.
Save alexeygrigorev/a1bc540925054b71e1a7268e50ad55cd to your computer and use it in GitHub Desktop.
Downloading segmented video from vimeo
import requests
import base64
from tqdm import tqdm
master_json_url = 'https://178skyfiregce-a.akamaihd.net/exp=1474107106~acl=%2F142089577%2F%2A~hmac=0d9becc441fc5385462d53bf59cf019c0184690862f49b414e9a2f1c5bafbe0d/142089577/video/426274424,426274425,426274423,426274422/master.json?base64_init=1'
base_url = master_json_url[:master_json_url.rfind('/', 0, -26) + 1]
resp = requests.get(master_json_url)
content = resp.json()
heights = [(i, d['height']) for (i, d) in enumerate(content['video'])]
idx, _ = max(heights, key=lambda (_, h): h)
video = content['video'][idx]
video_base_url = base_url + video['base_url']
print 'base url:', video_base_url
filename = 'video_%d.mp4' % video['id']
print 'saving to %s' % filename
video_file = open(filename, 'wb')
init_segment = base64.b64decode(video['init_segment'])
video_file.write(init_segment)
for segment in tqdm(video['segments']):
segment_url = video_base_url + segment['url']
resp = requests.get(segment_url, stream=True)
if resp.status_code != 200:
print 'not 200!'
print resp
print segment_url
break
for chunk in resp:
video_file.write(chunk)
video_file.flush()
video_file.close()
@Somstradamus
Copy link

I also have a problem with the new format of the vimeo vids. I usually use youtube-dl-web vercell online site and when the code is “master.json”, i just replace json with mpd and it works perfectly. But new one is “playlist.json?omit……” and it does not work at all. Is there any simple method using replacement as it was before? Thx!

Use the script posted above by @kbabanov

You need to install Python on your system tho, ChatGPT is your friend to compile it.

@Gnumaru
Copy link

Gnumaru commented Jul 5, 2024

merging the streams with moviepy is slow because it does unnecessary conversions. You can use ffmpeg to just assemble an mp4 file copying the streams without conversion
subprocess.run(['ffmpeg', '-i', 'video.mp4', '-i', 'audio.mp4', '-c:v', 'copy', '-c:a', 'copy', name])

then you can delete these lines

video_clip = VideoFileClip(video_tmp_file)
audio_clip = AudioFileClip(audio_tmp_file)
video_clip_with_audio = video_clip.set_audio(audio_clip)
video_clip_with_audio.write_videofile(name)

Of course it does create a dependency to an external command but it is worth the time saving.

Also, I've just noticed that downloading a master.json type stream (the one that just offloads the download to youtube-dl, generates a video file without audio =(.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment