How to sort and label large image libraries?

[grayestofghosts]

Hi, I've been hoarding images and other data for a while but this is the first time I've gotten around to really trying to organize them and am looking for some help. I've been working on a library/archive/time capsule project that includes a large amount of images and I feel like some of them need context to understand and want to preserve that information for possible future users. I have a Mac computer so I've started labeling things with tags but I understand that this feature doesn't transfer across file systems. While poking around I found that .jpegs have a comments field in their metadata but this isn't reliably preserved, even on the same Mac it seems -- if I put a comment on a .jpeg on a flash drive, disconnect the drive and reattach it, I can't see it anymore. It seems like the only reliable method of attaching data would be through the directory tree and file name, and maybe a text document in the folder explaining some things that need more explanation cross referencing by the file name? Are there any methods I'm missing that aren't too fragile/unreliable/non-transferrable? Thanks.

Comments

[luckyzukky]

i've heard good things about hydrus network for this exact use case. i haven't personally tried it (i instead deployed a booru kinda site but it has its downsides for stuff like this) but i might give it a shot. i'd look into it if i were you, it seems quite capable and adaptable to your needs

[anomalymonster]

I have personally used it and I use it a lot! I sync it between a Windows and Linux, and can confirm it works between devices (though it needs to be set so one can make changes and the other only reads it, otherwise the database can corrupt). I haven't personally tried it on Mac though and can't vouch for how well it would work.

The biggest downside to me is that it organizes folders based on the SQL database and it assigns them names that are random strings, which makes it harder to find files outside of the Hydrus client. It can also be a bit fragile so if you fuck around with it too much, the database can corrupt. Though as long as you make regular backups you should be fine (there's a portable version and you can literally just copy and paste the whole folder to a backup drive).

It also has a feature where you can include the links to where you got the image, which makes it very handy for archiving art and giving credit if you plan on sharing it later.

[doranwen]

Recently I heard about something called TagStudio for dealing with large image libraries. It's not exactly transferable, but as the developer points out in an intro video, all it takes is for the metadata "carrier" file to become separated from the actual image, and suddenly you've lost the info you were depending on. An open library format along with the pictures is possibly more useful in the long run. May not fit your situation, but I thought I'd mention it since I just saw the intro video this past week.

[grayestofghosts]

I'm not familiar with an "open library format". It looks like a JSON file?

[doranwen]

Right, all I meant by "open" is "not proprietary". Like, anyone can edit or read a JSON file. Wording is hard sometimes, lol.

[grayestofghosts]

Oh, so not from the Open Library organization but just like a database file, or even like a csv or xml.

[doranwen]

Right. Then you can always find a program to open or work with it, or code one later if need be. But again, I have never used the program, so this isn't a plug for it, lol. I was just impressed by the video that I saw and thought it looked really useful, so when I saw this post it came to mind and I figured I'd mention it.

Thoughts

[ysabetwordsmith]

This isn't so much about mechanics as about theory:

Think about how you most often search for things when trying to find something. Are you looking for a topic? A size/shape? A time period? That can help you figure out how to design the main structure of your organizational system.