>install hydrus client, fuck yeah a tag system for all my

>>61563823
Why are you saying it on /g/?

>>61563874
Who else am I going to bitch to? Do you think /b/ or /a/ or /tv/ is smart enough to understand having to deal with 60,000 unorganized images? I have to write a fucking script now to delete all the thumbnails and after that there seems to be an entire browser cache of 6 months in there. I found porn thumbnails that I don't know where they came from.
On the flip side, would BASH or python be a better starting point for this? I just need to check the dimensions of an image and delete it if they are 250x250 or smaller.

Hydrus is a huge meme and so is image tagging. Just like, use folders brah.

File: f8905296.jpg (275KB, 1280x720px) Image search: [Google]

275KB, 1280x720px

So apparently I was on /o/ at some point. According to my browser history I have never visited /o/. This is some weird shit. Bumping anyway while I try to find an elegant way to parse the file command.

>>61563823
Works on my machine
how the fuck did you manage to delete your entire archive?

>>61564241
I'm guessing it was the option of what to do with the original file after importing. Default is delete the file. I think it found the file, deleted it, then prompted me with an error about how it could no longer import the file because it didn't exist.

>>61563947
Bash should be fine, use mediainfo and grep for dimensions.

>>61563823
Make symlinks in folders for each tag. Scalable and compatible with everything since it's just files

>>61563823
Just restore from a backup. You do have backups that you have confirmed working, don't you?

File: soft_link[1].png (12KB, 384x419px) Image search: [Google]

12KB, 384x419px

>>61564304
>Make symlinks in folders for each tag.
Not OP, but what would that look like? I'm thinking about stepping up my organization game.

>>61564304
What about searching for an image using more than 1 tag

>>61564309
Funny you mention that. I back up important things. Personal info, financial stuff, an SD card that has the encrypted zip file with all my passwords and their corresponding sites on it, but I never considered these images important. It's not until I lost them and flipped the fuck out that I realized how long I've been saving them and how much your shitpost picture folder can define an anon.
>>61564292
This is working much more nicely than $ file was. Even if $ file was accurate grepping the dimensions was a bitch. Thanks anon.

>>61563823
This is why you should be using BTRFS with snapshots.

>>61564304
>delete original image file
>dead links

Hydrus would be very good if it could manage an existing folder instead of copying it to another one.

>>61564445
Op here. I agree. Also i hate regular expressions. This shit will be the end of me.

>>61563823
That's why you should RTFM first, I think it openly states that it destroys your existing archive by default.

That's also why I didn't want to use Hydrus.

I've been thinking of making a Hydrus clone that doesn't behave like that, but it's a lot of work and I wonder if catalogizing my wank material is worth it.

>>61564304
>Scalable and compatible with everything
until you run out of inodes

>>61563823
Are they mostly anime images? Because if they are I have a little bash script I made a while back that uses md5 hashes to grab the tags from boorus and write them to the file with exiftool.

>>61564304
Apart from already mentioned problems, some programs might clobber symlinks

>>61564956
That sounds awful

>>61563823
Why didn't you restore the image archive from your backup, OP? Oh right.
>Complaining about data loss
>Using an OS with backup software built in
>Couldn't be bothered to set it uP
>Too cheap to spend $50 on a USB HD
You did this to yourself.

>>61564973
Shotwell will list all the tags so it works pretty well for me.

>>61564983
Once you've changed the exif, md5 sum of the whole file changes too, which means that the image is no longer searchable my md5, no?

>>61565001
The script starts by clearing exif keywords so the md5 is good clean again, so if tags are updated on the booru you can update the image. I'll post it in a bit once I clean it up a little.

>>61565018
But if the original image had exif tags, then it clears them and still changes the hash

>>61565025
Yeah, it doesn't work all the time, but it works most of the time.

>run a test shell command on an image to get dimensions and evaluate if it's too small to be a real image
>everything works
>put it in a for loop going through the directory
>nothing works
>for loop is broken
stackexchange has failed me. Fuck it I'll get back to this tomorrow. Thanks for keeping this thread alive for a few hours.

File: 1497887412730.jpg (144KB, 1300x1134px) Image search: [Google]

144KB, 1300x1134px

This actually gave me some motivation to read up on graph databases
I wonder if you can build something like hydrus using one

File: 2017-03-21--1490169405_905x230_scrot.png (11KB, 905x230px) Image search: [Google]

11KB, 905x230px

>>61565025
>>61565018
save the original md5 as a tag dumb dumb

>>61564956
>>61565910
Here's my script. It's a little messy but it works for the most part. https://pastebin.com/dDSwRUFf