Explain bind_host in sample configuration
Remove `with requests` when it is unnecessary
Do not actually write out pages.
Report errors when an account has been terminated
Stream responses on /vi and /ggpht endpoints

The chunk_size=None parameter to iter_content lets us consume data as
soon as it arrives
Fix regular captions

This removes all of the code that was previously used to get them from
/timedtext, and instead, always uses whatever is extracted from the
video page.

This does unfortunately now require a whole video fetch just for the
captions. But assuming captions are only requested by a frontend, this
won't be a problem due to the memory cache. The captions link will be
in memory because the just-requested video is in memory too.
Do not extract storyboards into v/a formats
Extract fact check notices to second__clarification
Fix recommended extraction when fact check notice
Fix recommended videos extraction on IDs starting with - and _

Let's just leverage yt_dlp instead of rolling our own algorithms and fix
this kind of issue (not finding yt_dlp dump file for a given video) once
and for all

Example videos:
* https://www.youtube.com/watch?v=-q78QXpSL2M
* https://www.youtube.com/watch?v=_4SKG5uUEqs
Update link to NewLeaf installation
Fix file detection and recommendations
Update blurb in README to reflect current state
Fix link to documentation

As mentioned by @lesnake:matrix.org on matrix channel
Add many venv names to gitignore
Retrieve the first 20 comments of a video on /api/v1/comments/:videoid

Got some inspiration from https://github.com/nlitsme/youtube_tool (for the x-youtube-client-X
This is not a complete reimplementation of Invidious API as continuation is not implemented
(to retrieve more than the first 20 comments and comments replies), likes and replies count
are also missing.
Support premiere videos on channel
Quote json keys correctly