Some scripts to download videos from Microsoft Stream Classic

ffmpeg_proxy.go

package main

import (
	"github.com/elazarl/goproxy"
	"net/http"
	"regexp"
)

const Cookie = "" // TODO

func main() {
	proxy := goproxy.NewProxyHttpServer()
	proxy.Verbose = true
	proxy.OnRequest(goproxy.ReqHostMatches(regexp.MustCompile("^.*:443$"))).HandleConnect(goproxy.AlwaysMitm)
	proxy.OnRequest(goproxy.ReqHostMatches(regexp.MustCompile("^.*$"))).DoFunc(func(req *http.Request, ctx *goproxy.ProxyCtx) (*http.Request, *http.Response) {
		req.Header.Set("Cookie", Cookie)
		resp, err := ctx.RoundTrip(req)
		if err != nil {
			ctx.Warnf("error: %s", err)
			return req, nil
		}

		return req, resp
	})

	panic(http.ListenAndServe("0.0.0.0:7777", proxy))
}

microsoft_stream_download.py

import json
import os.path
import re
import subprocess

import requests

msft = requests.Session()
msft.headers['Cookie'] = ''  # TODO


def parse_duration(val):
    if 'H' not in val:
        m = re.match(r'PT(\d+)M([\d.]+)S', val)
        assert m is not None
        return int(m[1]) * 60 + float(m[2])
    else:
        m = re.match(r'PT(\d+)H(\d+)M([\d.]+)S', val)
        assert m is not None
        return int(m[1]) * 3600 + int(m[2]) * 60 + float(m[3])


VIDEO_IDS = []


def main():
    total_duration = 0
    names_map = {}
    download_tasks = []
    for video in VIDEO_IDS:
        r = msft.get(f'https://euwe-1.api.microsoftstream.com/api/videos/{video}?api-version=1.0-private')
        r.raise_for_status()
        r = r.json()

        names_map[video] = {'name': r['name']}

        remote_duration = parse_duration(r['media']['duration'])
        total_duration += remote_duration

        if os.path.exists(f'videos/{video}.mkv'):
            local_duration = float(subprocess.check_output(
                ['ffprobe', '-v', 'error', '-show_entries', 'format=duration', '-of',
                 'default=noprint_wrappers=1:nokey=1', f'videos/{video}.mkv']).decode().strip())

            if abs(remote_duration - local_duration) < 15:
                names_map[video]['ok'] = True
                continue

            assert False, f'Video {video} downloaded partially, remote: {remote_duration}, local: {local_duration}'

        playback_urls = r['playbackUrls']

        try:
            m3u8 = next(filter(lambda x: x['mimeType'] == 'application/vnd.apple.mpegurl', playback_urls))
        except StopIteration:
            assert False, f'Unsupported video: {r}'

        download_tasks.append((video, m3u8['playbackUrl']))
        names_map[video]['ok'] = False

    with open('videos/names.json', 'w') as f:
        json.dump(names_map, f, indent=2)

    print(f'Total duration is {total_duration} seconds')

    for uid, url in download_tasks:
        ffmpeg_command = f"ffmpeg -i '{url}' -map 0:v:3 -map 0:a -c copy videos/{uid}.mkv"
        subprocess.check_output(['tmux', 'new-session', '-d', '-s', f'download_{uid}', ffmpeg_command],
                                env={'http_proxy': 'http://127.0.0.1:7777'})

        print(f'Started download for {uid}')


if __name__ == '__main__':
    main()