[YouTube] Make channelId resolution follow redirects and fix YoutubeStreamInfoItemLockupExtractor building incorrect handles

[litetex]

Followup of #1344 (comment)

Fixes #1345

This fixes the following problems:

1. Fix YT handle resolution. It works now like this:
   When resolving a YT handle like @TheDailyShow, YT first returns a response with WEB_PAGE_TYPE_UNKNOWN and a redirect to thedailyshow. The redirect must then be resolved again to get the actual channel id which is UCwWhs_6x42TyRM4Wstoq8HA. However this is not always the case.
2. I built the handle url in YoutubeStreamInfoItemLockupExtractor incorrectly, however it was (sometimes?) resolved correctly due to the previous problem.

• I carefully read the contribution guidelines and agree to them.
• I have tested the API against NewPipe.

[AudricV]

When resolving a YT handle like @TheDailyShow, YT first returns a response with WEB_PAGE_TYPE_UNKNOWN and a redirect to thedailyshow. The redirect must then be resolved again to get the actual channel id with is UCwWhs_6x42TyRM4Wstoq8HA.

This is not always the case, this is the first time I see this. You can try for instance with @google or @Gronkh, you get directly the channel ID in a browseEndpoint.

Please update your comment in the code and your PR description :)

[litetex]

Please update your comment in the code and your PR description :)

Done :)

[Stypox]

Code in this PR looks good to me, and this function is only called to resolve URLs upon fetching stuff so it's not like we're going to get many more requests (it'd be bad if it was one request per info item). Thanks!

[Stypox]

Btw this (|| &&) combination might be wrong, since clarifying parenthesis are missing. What is the intended condition here @AudricV?

[litetex]

the equalsIgnoreCase should also likely be inverted for better null safety (we can also then remove the "" initialization above)

[AudricV]

What is the intended condition here

The intended condition is if it is a valid webpage type and a valid browse ID so:

(webPageType.equalsIgnoreCase("WEB_PAGE_TYPE_BROWSE")
        || webPageType.equalsIgnoreCase("WEB_PAGE_TYPE_CHANNEL"))
        && !browseId.isEmpty()

Thanks for catching this 4 year-old-bug I introduced!

[TobiGr]

Fixed in 6397b2e. It wasn't the only place where the parenthesis was missing. getChannelResponse and resolveChannelId are quite similar and might qualify for deduplication at a later point...

[Stypox]

I wonder if there's any checkstyle rule for this, but I could only find one which does the opposite 😅 https://checkstyle.org/checks/coding/unnecessaryparentheses.html

[AudricV]

Not really important, but I wonder whether the contrary is more common instead. If so, this comment should updated.

[AudricV]

What is the intended condition here

The intended condition is if it is a valid webpage type and a valid browse ID so:

(webPageType.equalsIgnoreCase("WEB_PAGE_TYPE_BROWSE")
        || webPageType.equalsIgnoreCase("WEB_PAGE_TYPE_CHANNEL"))
        && !browseId.isEmpty()

Thanks for catching this 4 year-old-bug I introduced!

[AudricV]

We should probably add tests for channels without this handle -> c/ -> browseId redirection.

[TobiGr]

added in aa70b5a. I amended the initial commit which creates the YouTubeChannelHelperTest because the other mocks were touched while recording the new ones for the class.