Tesseract

After two weeks of re-writing the filtering engine 3 times, I'm finally back to where I started. Yay!

Well, a little better than where I started. The more filters I added, the more complex and ugly the code to make them work became. After two re-writes, I'm finally happy with the 3rd iteration. It's pretty modular and will allow me to easily add and tune additional filters later.

Other than that, what's new since the last update?

1. Can now filter (or hide entirely) posts and comments if they are below a set score. This can be done either globally or a per-community group basis.
2. I ripped out the user tagging feature and am going to re-write it closer to how the community groups are done. The main reason is the original code was just a test/proof-of-concept. Additionally, once I do this, it would also allow creating filter policies that apply to user tags the same way they apply to community groups.
3. Instead of only showing the first policy hit, all policies are evaluated, and each reason is shown. I wanted to avoid this, but since different policies/options can have different effects (hide vs filter), it was possible to "short-circuit" a more strict filter if it hits a less strict one first. Additionally, it is nice to see all the reasons a post was filtered rather than just the first hit.
4. Editing the filter preferences is MUCH less of a fustercluck now. The global policy and the policies attached to community groups are now standardized, and there is now a standardized policy edit component (seen in post image). Not only is it more organized, it's much more modular and allows editing filters from the quick settings menu.
5. The policy name in the stub/placeholder (e.g. [Default Policy]) is clickable and will bring up a modal to edit that policy.
6. The global filters that apply to everything are now called "Default Policy". Currently, the names of the community groups that have filtering enabled are also the policy names. User tagging will likely be similar once I get that re-implemented.
7. I'm adding user exceptions to the policies (work in progress). That should allow you to exempt specific users (or user regex patterns) from the policy (global or community-group based) while filtering or hiding everything else that the policy flags.
8. I removed the "Help Farmer" filter option. I don't want to have to maintain a list of "ask" communities, and you can achieve the same effect by putting those communities into a group yourself, enabling filtering for it, and setting a minimum account age to filter.

As you can see, the biggest areas of work in this version are to the filtering options.

Wish I had more to report, but re-writing the whole filtering engine was a huge grind, and I had to take several days off to avoid burning out completely. Now that that's over with, I'm hoping the development pace picks back up on the remaining 45 items on my "to do" list for this release.

Ideas I'm Toying With. Thoughts?

1. Option to hide post/comment scores until after you've voted. Basically to coach you to vote based on the content rather than joining bandwagon.
2. Provisions to assign filter policies to user tags. Seems like it could be useful, but trying to think if that would do anything the existing global and community group-based filters don't already do.
3. ???

Screenshots

Please do not take anything personally if you show up in the screenshots of my filtering preferences. These were either chosen at random because they were at the top of the feed at the time of testing or were chosen because they have both Lemmy and Piefed versions/alts and were used to test wildcard and regex based filtering.

Also, all of the filtered items can be optionally hidden completely. They're shown as the placeholder stubs here since that is more useful for a screenshot.

Filtered Post Placeholders

• The option to filter low score posts is enabled with a threshold score of -5.

• Ozma is filtered in my "News and Politics" group as well as anything with the keyword 'trump'
• I filter communities from feddit.org since I don't speak German.
• I don't want to see slop from Mastodon in my feed

Misc Filtering Options for the Policy

Filter Users Directly or by Regex Pattern

Filter Communities Directly, By Regex, or by Instance

Filter Specific Keywords

Filter by Instance/Platform

Filter by Post's URL Domain

Exempt Specific Users or User Regex Patterns from the Policy

Note: This is still work-in-progress