destreamer/README.md
kylon 83fecf2894
Fixes and clean up (#51)
* Simplify main

* Fix init

* Cleaner output for the end user

* Fix extractVideoGuid after sync with dev

* TokenCache: Make variable private

nit: switch to import

* Add option to disable video thumbnails

* Create a unique file name to avoid overwrite

* Remove os dependency

* Reimplement simulate

* Update README

Co-authored-by: @kylon
Co-authored-by: @snobu
2020-04-10 19:35:57 +03:00

4.5 KiB

Destreamer

CI build status

Saves Microsoft Stream videos for offline enjoyment

Alpha-quality, don't expect much. It does work though, so that's a neat feature.

It's slow (e.g. a 60-min video takes 20-30 minutes to download). Not much i can do about it for now unless i find a better way than ripping HLS.

NEWS

  • We now have a token cache so we can reuse access tokens for their one hour lifetime. What this really means is that within one hour you only need to login via the popup browser once.

This project is now looking for contributors

Roadmap -

  • Token cache (so you don't have to log in every time you run destreamer)
  • Download closed captions if available
  • Performance improvements (via aria2c maybe?) // This is under consideration, we're not sure if this borders on abusing the streaming endpoints or not.
  • Single static binary (for each major OS)

Send a quality PR first and i'll add you as a contributor to the repository.

DISCLAIMER

Hopefully this doesn't break the end user agreement for Microsoft Stream. Since we're simply saving the HLS stream to disk as if we were a browser, this does not abuse the streaming endpoints. However i take no responsibility if either Microsoft or your Office 365 admins request a chat with you in a small white room.

PREREQS

  • Node.js: anything above v8.0 seems to work. A GitHub Action runs tests on all major Node versions on every commit.
  • ffmpeg: a recent version (year 2019 or above), in $PATH or in the same directory as destreamer.ts.

Destreamer takes a honeybadger approach towards the OS it's running on, tested on Windows, macOS and Linux, results may vary, feel free to open an issue if trouble arise.

USAGE

  • npm install to restore packages
  • npm run -s build to transpile TypeScript to JavaScript
$ node ./destreamer.js

Options:
  --help                    Show help                               [boolean]
  --version                 Show version number                     [boolean]
  --username, -u                                                     [string]
  --outputDirectory, -o                          [string] [default: "videos"]
  --videoUrls, -V           List of video urls or path to txt file containing the urls
                                                           [array] [required]
  --simulate, -s            Disable video download and print metadata
                            information to the console
                                                   [boolean] [default: false]
  --noThumbnails, --nthumb  Do not display video thumbnails
                                                   [boolean] [default: false]
  --verbose, -v             Print additional information to the console
                            (use this before opening an issue on GitHub)
                                                   [boolean] [default: false]

Make sure you use the right escape char for your shell if using line breaks (as this example shows).

For PowerShell your escape char is the backtick (`) instead of backslash (\), for cmd.exe use caret (^).

$ node destreamer.js --username username@example.com --outputDirectory "videos" \
    --videoUrls "https://web.microsoftstream.com/video/VIDEO-1" \
                "https://web.microsoftstream.com/video/VIDEO-2" \
                "https://web.microsoftstream.com/video/VIDEO-3"

You can create a .txt file containing your video URLs, one video per line. The text file can have any name, followed by the .txt extension. Run destreamer as follows:

$ node destreamer.js --username username@example.com --outputDirectory "videos" \
    --videoUrls list.txt

Passing --username is optional. It's there to make logging in faster (the username field will be populated automatically on the login form).

You can use an absolute path for --outputDirectory, for example /mnt/videos.

RANDOM NOTE

IMPORTANT NOTE

Just ignore this error, we already have what we need to start the download, no time to deal with collaterals -

image

EXPECTED OUTPUT

<<<< OUTPUT >>>>

The video is now saved under videos/, or whatever the outputDirectory const points to.