We've currently got Category.image, Category.video, etc. The troublesome one is Category.other. Many skins support images, videos, web locations, and PDF's, but not things like Word documents or zip files. As a result, a lot of otherwise simple tests become clumsy compound affairs:

isPDF = fileCategory == Category.other && fileName.toLowerCase().endsWith(".pdf");

Ugh. And if you want to know if there are any PDF's in a directory, this won't do the job:

cc = JAlbumUtilities.countCategories(currentFolder, false);
anyPDFs = cc.getCount(Category.other) > 0;

That ends up counting all the other stuff, too, like Word docs and zip files.

It would be very handy to have a Category.pdf available.

To avoid badly breaking existing skins, Category.other should still include the PDF's.

Not sure I understand the use case for that. If you look at the list of jAlbum supported file types, once you get past images, videos, audios, web locations, and PDF's, it's a pretty motley collection - mostly things I've never encountered in a live album. And it includes some things like DjVu files, which haven't been seen in the wild in almost two decades.

Adding new categories is sadly a brittle endeavor: They are Enums, and once you've added a new one, loading such a project into an older jAlbum will break it.

Actually, I don’t need counts. I just need to know if there are any in a folder, with a Boolean to indicate whether to recurse. Hell, even a hack like me can write that!

ETA: Sometimes I forget that if the core doesn't provide the method I need, I really can just sugar my own churro.

In Groovy, you can write stuff like this to check if a folder has pdf files:

currentFolder.children.stream().anyMatch(f -> f.name.toLowerCase().endsWith(".pdf"))

In Java, it's like this:

currentFolder.getChildren().stream().anyMatch(f -> f.getName().toLowerCase().endsWith(".pdf"))

That only looks inside the 1:st level of that folder. If you want to know whether a pdf is lurking at any depth under a folder, just change getChildren() to getDescendants().

I happen to love the Stream API, but I usually need to look up the details. Add the word parallel() to the chain, and the whole thing is solved by multiple CPU cores as well That helps speeding up traversing deeper folder structures.