[ ] I’m using docker compose setup, with modifications (modified .yml files, third party plugins/themes, etc.) - if so, describe your modifications in your post. Before reporting, see if your issue can be reproduced on the unmodified setup.
[ ] I’m not using docker on my primary instance, but my issue can be reproduced on the aforementioned docker setup and/or official demo.
According to the UI, the similarity ratings are showing up, so the PG plugin seems to be working
I have my preferences set up properly for the similarity plugin (pretty sure) see the first followup comment.
However, when I go through my articles, I’m not seeing the numbers in “all articles” drop by more than one when I’m going through articles that are flagged as > my “minimum similarity” and have title lengths > the “minimum title length”.
I’m not sure how to enable debugging with the docker stack to search for these log messages to do any more diagnosis on my own unfortunately.
Refreshing the feed (F,R) or re-loading the browser page completely still keeps the total number of “all articles” the same, so I’m guessing that these similar articles aren’t getting marked as read.
The plugin is described as: Marks similar articles as read (requires pg_trgm)
So when I mark an article as read, the total number of unread articles (in the box by “all articles”) drops by one (normal behavior). If the plugin is working and marking some of the other articles as read in the background (the ones that are similar to the current article), then I’m assuming that I should see the total number of unread articles drop by more than one. Does that make sense, or perhaps I’m mis-understanding how the plugin works?
example:
I read an article that doesn’t have any similar articles: “all articles” unread count drops by one.
I read an article that has two similar articles: “all articles” unread count should drop by 3 (one for the original, two for the two duplicates being marked as read)
My understanding is: If I read an article about Apple’s new iPhone once, all of the other “Apple unveiled a new iPhone” articles that match the similarity values, should also be marked as read and I won’t see them (I only want to read about the Apple iPhone once, not 20 times. )
as far as i remember, the plugin has no concept of ‘origin’ articles: it works on the feed it’s enabled on and marks articles as read if they have necessary similarity score to other articles in the database.
the idea is that you don’t enable it for the feed which is the authoritative source on whatever you’re interested in.
Does the function that marks similar articles as read happen at read-time, or when polling the RSS feed, or when? For example, the feed debugging log - is that happening when the user is reading an article, or when the feed gets updated?