@Zerowinger
There won’t be an automatic import of those as the script doesn’t like dupes, but I’ll manually import the ones that are missing when it happens.
Bless you.
I’d only gotten as far as saving a copy of an image if I don’t already have it. But your idea of saving comments is interesting; if I got that to work, saving tags as well should be trivial. That might bring up questions about using Derpibooru database dumps as a tag source, but let’s cross that bridge later.
My area of experience is embedded systems, so C/C++, assembly, and Python (if I’m feeling REALLY ambitious) are my comfy languages, though I’ll dabble in Perl and PHP if I have to – such is the life of a LAMP stack user. I use Arch, BTW. 😉 My “head script” was mostly in Bash 🙄, so I bet your JS would beat the socks off of it.
Regarding compatibility, the only issue I’ve ever read about on Derpibooru had to do with rate limiting, or lack thereof, by a user script. I think it might even be mentioned in the API documentation (just checked, it is, right near the top under Licensing), but it’s also common sense that you don’t want to hammer their server with requests at an audible frequency. 🤣 As long as it can finish in a “reasonable” amount of time, and it’s grabbing data at a “reasonable” rate, I don’t think anyone will have a problem.
There are currently 33,835 images with the ai generated
tag on Derpibooru. There are 27 days left until those image pages, and whatever is uploaded over the next four days, are deleted. 27 days == 38,880 minutes, so theoretically, one request per minute would do the job, IFF one were to grab just the image file. At least we’re not in a rush. :)