The new comment format is similar to the description's commandRuns.
This should fix the issues with most comments but there are still
some more changes that would need to be made like adding support for
formatting (bold, italic, underline) and channel emojis.
Fixes issue 4566
Returns the 'isPostLiveDvr' field in the videos API when the video
is a post-live DVR (= ended livestream that hasn't been reprocessed
into VOD yet).
Example taken 10 minutes after that livestream ended:
/api/v1/videos/euqnWk-uP6M
{
...
"isPostLiveDvr": true,
...
}
Partially fixes 4421
YouTube added a new format for like data in December 2023.
This PR adds support for parsing saod new format.
Related: LuanRT/YouTube.js issue 557
Closes 4419
Before this PR, Invidious assumed that fetching the DASH manifest from
YouTube will always be successful and didn't check the status code.
That meant that if YouTube gave a rate-limiting page, invidious would
return an HTTP 200 response with the 'application/dash+xml' Content-Type
header and the YouTube ratelimiting page as the body.
No associated issue
This handler should no have been removed in 4276, as it adds the required CORS
header (Access-Control-Allow-Origin) for public acces to the API.
Thanks to iBicha for noticing this!
This handler should no have been removed in 4276, as it adds the required CORS
header (Access-Control-Allow-Origin) for public acces to the API.
Thanks to iBicha for noticing this!
The function names `count_users_active_6m` and `count_users_active_1m` were
swapped. As the names were swapped on both sides (declaration and use), this
had no functional impact.
No related isse was tracked.
This PR fixes an issue where the `scopes` parameter would see its wildmark
characters (*) removed during the login page redirection, after that a call
to `/authorize_token` was made while the user was not logged in.
Closes issue 4200
Multiple users have reported that the fields parameter is slowing down API
response times significantly. As most API endpoints are already optimized to
make as few requests as possible to Youtube, there is no point in limiting the
output. Furthermore, the added processing might be part of the broader memory
leak problem (See 1438).
In addition, the small increase in data output is not much of an issue compared
to the huge video proxy that lies next to this API.
No related issue tracked
This PR parses the start and end time for clips.
It also adds a new, dedicated API endpoint (`/api/v1/clips/{id}`) for
retrieving the start and end time of a clip.
Here is a sample response from that new endpoint (`video` is a video object,
as described in https://docs.invidious.io/api/common_types/#videoobject):
GET `/api/v1/clips/UgkxxPM3BRphCAPLP88YoUGuj79KXPfpNNO_?pretty=1`
Response:
```
{
"startTime": 8842.645,
"endTime": 8855.856,
"clipTitle": "✂️ Kirby is pink!",
"video": {}
}
```
Closes issue 3921
Some playlist author's don't have a YouTube channel, so does movies.
This caused various extractors (related videos, search) to fail.
Closes the following issues:
2530, 3349, 3766, 3812, 4133
* Add option to disable force_resolve in make_client
Some websites such as archive.org and textcaptcha.com
does not support IPv6 and as such requests fail when Invidious requests
with IPv6 to those services.
* Reenable force_resolve on pubsub subcribe request
* Make force_resolve false by default in make_client
* Remove missed explicit force_resolve=false
