feat(blacklist): Automatically add media with blacklisted tags to the blacklist

[benbeauchamp7]

Description

This pull request introduces blacktags, which automatically blacklist media based on their tags (keywords). Blacktags are configured on the settings page and processed with the Process Blacktags job, which is configured to run weekly by default.

The job queries discover with each sort option round-robin style, with each sort being queries up to "Limit" times, where limit is configured in the settings just under where tags are entered. This is done so that media is blacklisted as much as possible no matter the sort used on the discover page (I don't want media with blacktags to appear when I sort popularity ascending for example). The default limit is 50 pages, resulting in a maximum of 32,000 blacklist entries per tag (2 media types * 50 pages * 20 entries per page * 16 sort options). The maximum limit is 250, which would cap out at 160,000 entries per tag.

Screenshot (if UI-related)

UI

Settings:

Job:

To-Dos

• Add copy and paste functionality to keyword setting to allow sharing of blacktag configurations
• Add blacklist filter to hide media added by blacktag
• Successful build pnpm build
• Translation keys pnpm i18n:extract
• Database migration (if required)
• More bugtesting before ready for review

Issues Fixed or Closed

• Fixes Add Tag-Based Content Filtering System in Blacklist Settings #1076

[benbeauchamp7]

That covers the functionality I was planning to implement. The only request I didn't cover from the original issue is an easy way to enable or disable the feature from the settings with a checkbox. I kind-of implemented a way to easily mass-unblock media by running the job without any blacktags configured, which will drop the blacktagged media.

Hoping to open everything up for review tomorrow night, want to do one last pass for bugs and code review

[gauthier-th]

Genuine question, I didn't look at the code yet: why do you have to implement it using a job? Why can't this be done the same way as the standard blacklist is done, i.e. dismissing the item when it is displayed if the condition matches?

[benbeauchamp7]

TMDB doesn't return keywords on their endpoints, it actually looks like keywords are one of the only things they *don't* return. In order to not stuff the blacklist, I would have needed to hit the API again *for each movie* to get their keywords before actually showing the media. That felt slow and bad, so I went with the job.

…

On Mon, Jan 27, 2025 at 4:03 AM Gauthier ***@***.***> wrote: Genuine question, I didn't look at the code yet: why do you have to implement it using a job? Why can't this be done the same way as the standard blacklist is done, i.e. dismissing the item when it is displayed if the condition matches? — Reply to this email directly, view it on GitHub <#1306 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKKZSHDCWCCYIW3D4LOE5IT2MYAA5AVCNFSM6AAAAABV3ARAB2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMMJVGMYTMNBXHE> . You are receiving this because you authored the thread.Message ID: ***@***.***>