>>7326
>pull tags from various booru sites.
i'm deathly afraid of dealing with network code. i know it's probably a simple affair (GET then parse the returned webpage), but i know from experience that for whatever reason, shit don't work.
plus i don't want to add more bloat to it, especially since other things could do it faster and more efficiently than a single unified program
>Osmosis only works on a hardcoded image directory
this is true, reduces complexity drastically of code, and makes locating files an easier affair. it does make it a bit inconvenient if you want to use multiple folders, but writing code that stores location that won't freak out with duplicate file/directory names is a bit daunting, not to mention the frontend for a user to set such values would be a "bit of a bitch".
again, i'd like to keep it simple, my code is already a mess as-is (i'm relying on what is essentially controlled arbitrary code execution to make the tag filtering system work)
> but it also doesn't keep the files themselves as copies in an internal database
i don't see why one would want to do that, seems like a waste of file space and further convolution to the end user.
<CrazyAmphibian
ding ding
i have heard of hydrus, but haven't used it myself. i do have a friend who has previously used it, and i have taken some notes and gripes with what little i've seen.
so i more or less thought "yeah, i could make something like that. a sidegrade".
so that's partially why the program is the way it is.
images are in a clear defined place, which makes adding and removing them easy. same with tags, which makes reading and modifying them in a script really easy.
filtering fit for an autist, where tags can contain any character such as spaces or commas
also hydrus uses python, which i do not like for various reasons.
>inb4 Lua user shit talking python
that and it's a huge file size (>200MB uncompressed). Osmosis win64 is 34.8MB (really 7.9MB, the pdb file is completely optional and as far as i can tell, useless).