[youtube] Redefine uploader metadata

[coletdjnz]

IMPORTANT: PRs without the template will be CLOSED

Description of your pull request and other information

uploader metadata for videos, tabs, etc. is/was inconsistent and not well defined. With handles and the channel UCID url now being the two primary urls to a channel, this updates the metadata to reflect that.

New mapping will be strictly:

channel -> channel name
channel_id -> Only UCID
channel_url ->Only  UCID /channel url

uploader -> channel name (same as channel field)
uploader_id -> Only @handle
uploader_url - > Only @handle url 

Note this is slightly different than ytdl-org/youtube-dl@f7ce98a

Template

Before submitting a pull request make sure you have:

• At least skimmed through contributing guidelines including yt-dlp coding conventions
• Searched the bugtracker for similar pull requests
• Checked the code with flake8 and ran relevant tests

In order to be accepted and merged into yt-dlp each piece of code must be in public domain or released under Unlicense. Check all of the following options that apply:

• I am the original author of this code and I am willing to release it under Unlicense
• I am not the original author of this code but it is in public domain or released under Unlicense (provide reliable evidence)

What is the purpose of your pull request?

• Fix or improvement to an extractor (Make sure to add/update tests)
• New extractor (Piracy websites will not be accepted)
• Core bug fix/improvement
• New feature (It is strongly recommended to open an issue first)

[pukkandan]

cc @dirkf

[dirkf]

I compared the extractor test-cases. Where yt-dlp specifies uploader metadata here, the yt-dl test-cases in the latest git master generally match if present. Here are the non-matching ones.

YoutubeIE:

--- yt-dl
+++ yt-dlp
 ...
             'url': 'OtqTfy26tG0',
 ...
-                'uploader': 'The Cinematic Orchestra - Topic',
+                'uploader': 'The Cinematic Orchestra',
 ...

I didn't change this result, which is Sergey's from 0cf09c2, though the current yt-dlp result looks more plausible.

I think these remaining differences are all results that I changed in ytdl-org/youtube-dl@f7ce98a.

YoutubePlaylistIE:

--- yt-dl
+++ yt-dlp
 ...
        'url': 'PLBB231211A4F62143',
 ...
-                'uploader_id': '@WickmanVT',
+                'uploader_id': 'UCKSpbfbl5kRQpTdL7kMc-1Q',
 ...
         'url': 'https://www.youtube.com/embed/videoseries?list=PL6IaIsEjSbf96XFRuNccS_RuEXwNdsoEu',
 ...
             'id': 'PL6IaIsEjSbf96XFRuNccS_RuEXwNdsoEu',
 ...
-                'uploader_id': '@milan5503',
+                'uploader_id': 'UCEI1-PVPcYXjB73Hfelbmaw',
 ...
         'url': 'http://www.youtube.com/embed/_xDOZElKyNU?list=PLsyOSbh5bs16vubvKePAQ1x3PhKavfBIl',
 ...
-                'uploader_id': '@music_king',
+                'uploader_id': 'UC21nz3_MesPLqtDqwdvnoxA',
 ...

YoutubeYtBeIE:

--- yt-dl
+++ yt-dlp
 ...
         'url': 'https://youtu.be/yeWKywCrFtk?list=PL2qgrgXsNUG5ig9cat4ohreBjYLAPC0J5',
 ...
-                'uploader_id': '@backuspagemuseum',
+                'uploader_id': 'backuspagemuseum',
 ...

Also, ytdl-org/youtube-dl#31675 (comment).

[coletdjnz]

TODO:

• investigate channel/uploader name extraction (might be getting wrong one for topic channels)
• handle extraction fallback for channel pages (ownerUrls)
• notificationsIE (no handle info given)

[dirkf]

Reviewing these tests for a new yt-dl extractor version, I find a contradiction between the channel ("Full name of the channel the video is uploaded on.") values extracted for test videos IB3lcPjvWLA and MgNrAu2pzNs (eg).

IB3lcPjvWLA: Afrojack (yt-dlp test value) vs AfrojackVEVO
MgNrAu2pzNs: Stephen (yt-dlp test value) vs Stephen - Topic

Currently the yt-dl extractor doesn't test these values because it just sets channel from uploader, if set.

The second test is commented TODO: should be "Stephen - Topic". If so, I believe that the corresponding value AfrojackVEVO must be correct for IB3lcPjvWLA, which would (FWIW) match yt-dl's current values and the header of the corresponding channel page in Invidious in each case.

[Separately, what are these - Topic suffixes?]

The first value is here (similar to where yt-dlp gets it)

traverse_obj(initial_data, (
    'contents', 'twoColumnWatchNextResults', 'results', 'results',
    'contents', ..., lambda k, _: k.endswith('SecondaryInfoRenderer'),
    ('owner', 'videoOwner'), 'videoOwnerRenderer', 'title'),
    get_all=False)

and here

traverse_obj(initial_data, (
    'engagementPanels', ..., 'engagementPanelSectionListRenderer',
    'content', 'structuredDescriptionContentRenderer', 'items', ...,
    'videoDescriptionHeaderRenderer', 'channel'),
    get_all=False)

Everywhere else, eg traverse_obj(microformats, (..., 'ownerChannelName')) and as in the linked post, matches the second value.

What should be the correct resolution of this TODO?

I expect a similar issue applies to tab extraction of video data for playlists. The channel value extracted there should be consistent.

[coletdjnz]

An issue should be raised for this.

The topic suffix is correct unless YouTube has changed the channel. It's a topic channel which are quite messy to deal with.