leverage entry points for enumerating new extractors

[darkdragn]

Checklist

• I'm reporting a feature request
• I've verified that I'm running youtube-dl version 2019.10.29
• I've searched the bugtracker for similar feature requests including closed ones

Description

Topic

Code management, and new feature introduction.

This is going to sound a little pretentious, and I am sorry before hand. I love the work your team does, and I'm in the process of writing a few extractors, but I started to get curious about one point in your code base. That point being, has the team considered using entry points to load extractors and separating the codebase to more easily track issues or allow custom extractors?

If anyone is unfamiliar with entry points, python's documentation is sufficient, but here's a humorous walk through: Snek

Proposal

Separating codebase in to multiple projects. A quick patch to youtube_dl.extractors.init.py adding something similar to:

import pkg_resources
for entry_point in pkg_resources.iter_entry_points('youtube_dl_extractors'):
    _ALL_CLASSES.append(entry_point.load())

Could be sufficient enough to enable this capability.

Benefits

• Tracking extractor issues separate from core code issues
• Separating the extractors into different bundles (I.E. nsfw, anime, live, news...)
• Allowing for developers to test / maintain their extractors separate from the youtube-dl codebase (For example, instead of forking youtube-dl to write an extractor, just provide a cookiecutter template with an entrypoint and install it in a venv to play with an test)
• Meta packages: you could either make bundles of extractors be extras (youtube-dl[nsfw]) or provide metapackages that are just a setup.py and a requirements.txt, the requirements file listing the extractor plugins (youtube-dl-nsfw)

Afterword

Sorry if this has been proposed before. I spend a good 30+ minutes searching the issues and couldn't find a proposal like this, but I really think the community and your team could benefit from a restructuring like this. If your team supports this, I would be more than willing to kick everything off, and put forward a beta, I just don't want to run the leg work, if the team has any objections.

Thank you for your time. Thank you and your team for a wonderful project, and I with the project all the best going forward.

[dstftw]

This has been discussed numerous times. For now there are neither plans nor resources for that.