>>108857640
I use Hydrus.

>>108857640
i've been using TMSU lately. it's CLI, but i'm also using an Emacs package that integrates it with Dired, and it's a match made in heaven tbhdesu

>>108857640
flat structure
tags as part of names
all software that uses a DB / sidecars for it should be avoided

>>108859240
>all software that uses a DB / sidecars for it should be avoided
why?

Hydrus.
It's rough around the edges but it is the only thing that does this even somewhat well. The dev is nice too.

>>108857640
I'm working on it. I'll call it CHUDTagger v1.

>>108860253
ipod's database is one example of it thats dog shit if that little database is corrupt backing up the rats nest of the filesystem holding your files is a pain in the dick

>>108861994
id prefer it use metadata tags on the file that can work outside of the things database, so you can take the file, import it elsewhere and it just works and also not creating its own cryptic filesystem leaving said files in their current location/filename if hydrus worked in such a way i might be interested otherwise its trash

>>108859107
TMSU is pretty good but it doesn't look like it's maintained anymore. I've been considering using git annex since it supports tags as metadata and I like the idea of managing a file system in a way like git, but it is kinda overly complicated.

What are you looking for? Something for all your shit? I've never found a solution to centralize everything. I use hydrus for pictures, lanaragi for doujins, sonarr radarr and jellyfin for movies and series, beets and navidrome for music, calibre zotero and emacs for books and documents. Stash is good for porn too.

>>108862021
Main problem with that is that the metadata tags aren't cross-platform/cross-filesystem. You can't tag something on Windows and have it work on Linux.

Tags are mostly a meme. You can append any string to a filename. Multi-tag filenames would be tag1_tag2_tag3.jpngif

>>108862156
I tried that too. On Windows, you'll run into path length limits.

Ai would actually be beneficial here, where images and files are scanned and indexed. Then you just ask ai where that Jurassic park pepe is and it'll pull up the most likely matches statistically

>>108862053
>TMSU is pretty good but it doesn't look like it's maintained anymore.
yeah the author has been rewriting it in Rust (it's written in Go originally). for me needs it's pretty feature complete, so i don't really care about the lack of updates

>>108862156
A dedicated software makes it easier for you to manage tags.

>>108862228
I have tried AI tagging and it's very inaccurate.

>>108864236
The given examples are hydrus (garbage that mangles names, folder structure and locks you into hydrus forever) and tmsu which is completely useless without its sidecars
as long as you cannot put the metadata on the files itself it is not only a waste of time but an actual downgrade, unless you somehow stick to the same system for the next however-long-until-you-die.

>>108857640
>What do anons use to organize a large amount of data?
Hydrus

>>108864562
Works on my machine

hydrus with ocr workflow.

>spending time to make your organization WORSE instead of better
retards

>>108865308
My organization is pretty good

>>108864562
>completely useless without its sidecars
All tagging is going to involve a sidecar of some kind, whether it's an actual file you can edit yourself or an abstraction within the filesystem. I've researched this topic thoroughly and can give you some options.
There's a tagger called filetags that appends tags to the filename but you have a limited space to work with. Also one may not want to mess with a file's original filename in case it holds useful information. I used this for a while and it's perfectly fine as long as your tags fit within the filename limit and you don't mind long ugly filenames.
Another potential way is to use xattrs (extended attributes) but iirc only some file systems on Linux support it. Mac OS and Windows have something similar but exclusive to their own operating systems. No support on the BSD's. Also many programs do not respect xatters (ie you edit a pic in krita and it overwrites the file without your xattrs). And you have to take care to use the correct arguments when copying them to a new disk since they are not stored in the file itself, it's only metadata that the OS filesystem stores.

>>108864562
>>108867004
A method I'm looking more into now is git_annex's metadata feature. Git_annex brings all the functionality of git to your filesystem, so you can keep track of all file operations and revert them as needed, with everything being stored in a git database. It's somewhat clunky to use and has a learning curve but it is extremely powerful and includes built in metadata tagging and searching. My only real problem is the lack of automatic implicit tags (or "parent" tags) like TMSU where if you tag something with "cat" you can make it auto tag "animal" or any other tag you like.
Finally there's file-specific metadata that actually stores info in the file, such as exif. The format will be different for every file type and some types don't even support it. It alters the checksum of the file, which can make it difficult to find the file again or deduplicate, and there can be restrictions like all metadata needing to fit on one line with no linebreaks. I tried going this route once because I'm mostly tagging images, but I already had a large collection of webps which use a super retarded metadata format that is difficult to work with.