/hydrus/ - Hydrus Network

hopefully simple feature request:

saving downloader tabs

specifically, i sometimes download a lot of chan threads at once, and have to manually set up the tag settings and "ignore deleted files" options and so on for each tab every time… because session-saving doesn't include the tagging/etc options on a downloader that has not yet been activated.

essentially i want to be able to save and restore six un-triggered downloader tabs, *with* their tagging/deleted settings intact

>>5719

Nice! I think I was one of the people who originally suggested tab shuffling, this is a simple and convenient solution. I'd also been thinking about a way to show the duplicate scrolling easier, and changing background tint a little sounds ideal.

Question, I know that when you mark a file as better it commutes the tags of the one to the other, but is there any way to add tags during the comparison process to the list that will then be added to the better of the two? A lot of times a worse-quality image will have a watermark or URL that's not present on a higher quality image, but may yet be the original source, and it would be nice to have a way to get rid of the lower res image but preserve that metadata as a tag. Since I can't right-click to edit tags like in a normal viewer pane, I'm a little lost.

>>5720

So what is all of that? Porn?

>>5719

I'm currently downloading from gelbooru (not using a subscription) and havent had any problems so far. What are some of the problems others have been having?

>>5726

It works now, but they changed what the thumbnails link to on gallery pages. They now link to sketchy redirect links (assumedly to facilitate ads) which hydrus couldn't resolve. It can now, but with the feature that shows URLs of files, there's no way to tell if those URLs will work in the future. Similarly, subscriptions rely on urls never changing, so with the change they suddenly think all the images are new (because they have a different url). Basically, your gelbooru subscriptions will suddenly try to redownload themselves entirely.

Is there a way to disable the bg color change when filtering dupes? That stark change outside the images makes it more difficult to spot subtle changes within them.

>>5728

Thanks brah.

>>5720

My 50k micropenis salute you

Bug: shortcuts bound to two actions (say, "c" to "set tag "catgirl"" and "set tag "cute"") do not work, only the newest one is saved when you click "apply". At least in the same shortcut set.

Minor feature request OP:

If convenient, can you make the default minimum tab width a little smaller or better yet add an options box to set the size by integer or something? I know it's inconvenient to have near-blank tabs too tiny for most users, but working mainly off a 1336x768 screen I often find myself having to shift the tab bar back and forth with the directionals just to see the tabs I want.

>>5733

That might be by design to avoid your commands getting FUBAR. Have you considered assigning "cute" as a subtag of "catgirl" since you seem to want all that are tagged one to be tagged the other anyways? I have rating:NSFW set up to automatically get added when I enter board:(certain boards) tags.

>>5725

90% is.

For the most part it's the image entire databases of the following:

Gelbooru as of May 1st 2017(This took about a month to add, with around 100k files being added daily)

MyFigureCollection 2016

e621 2015

Rule34Hentai as of April 2017

Rule34Pahael 2015

HentaiFoundry 2015

And then around 500+ pixiv artists. There's about 20 other boorus/sites too but they have less than 300k images. It probably won't grow much bigger than this.

>>5727

>Do you actually tag stuff or just rely on the PTR and the tag look up?

I don't tag anything myself. I make tag archives of the sites I rip which already contains their tags, so nearly every image should be tagged(unless it wasn't tagged in the site)

>Which tag ranks the highest

As in like across all my files? I've no idea since I have tag counts and such disabled

>>5740

>image entire databases

Entire image databases*

Also I haven't had the time to run a dupe check, so that'll probably trim down 200k files or so.

>>5735

>That might be by design to avoid your commands getting FUBAR. Have you considered assigning "cute" as a subtag of "catgirl" since you seem to want all that are tagged one to be tagged the other anyways?

I don't want that. I just want to do it sometimes for when I scrap an artist that makes cute catgirl pics. Not all catgirl pictures are cute.

>>5742

Maybe a different key for each tag then? They could be right next to each other. I'm not sure Hydrus supports running macros from a single keypress yet.

Can you add this https://github.com/ipfs/go-ipfs/issues/3397#issuecomment-284337564 as an option for ipfs in the next release? It could be enabled when you check the checkbox in ipfs service settings.

>>5747

Sure, I will try to fit this in. I want to do some other IPFS work as well. If I can't do that option this week, I'll make sure to have it in for v257.

Getting an interesting bug; with the duplicate filter up, I can't move to the B image from the A. I click the button, but nothing happens. I'm on Arch Linux, fyi

Do the known urls in the dupe filter show up on the white background? Either they don't or every single image that ends up as B in mine has no known urls.

>>5752

>Maaaybe a 'open manage tags for both after the decision' checkbox?

Not him, but that would be much appreciated.

After resolving all the possible dupes, Hydrus will spam a popup window indefinitely. Forgot exactly what it said, but basically "no more dupes".

Why not add some more buttons to the top hover window? Open manage tags, remove all tags, for example. Just to make it easier for newbies, also someone like me who mostly uses that hover window instead of hotkeys.

Also the ability to select multiple images in the main browser and set them as alts to one another would be useful and help speed up the dupe checking process.

>>5761

>Also the ability to select multiple images in the main browser and set them as alts to one another would be useful and help speed up the dupe checking process.

That's a feature I would enjoy quit a lot considering the many variation set i fetch from pixiv.

>>5753

Thank you

Some kind of automated system to delete one of the two images set as an "Exact duplicate" would be grand. Look at how Dupeguru does its picture sort, allowing you to set the smaller of two images first or the smallest filesize. Having that be done automatically instead of trying to go through your list of extact dupes and picking one or the other, but simply based on which of them is xier or yier then the other would be very cool.

>>5720

>archive SIX MILLION

>inbox SEVEN

Good man. What kind of scrapers do you have running?

Since the last update, some 4chan threads have been giving me this for a few files, in one case it was for like 20+ files in a thread but usually just one or two:

TypeError: buffer object expected… (Copy note to see full error)
Traceback (most recent call last):
  File "include\ClientImporting.py", line 2817, in _WorkOnFiles
    HG.client_controller.WriteSynchronous( 'content_updates', service_keys_to_content_updates )
  File "include\HydrusController.py", line 386, in WriteSynchronous
    return self._Write( action, HC.LOW_PRIORITY, True, *args, **kwargs )
  File "include\HydrusController.py", line 117, in _Write
    result = self._db.Write( action, priority, synchronous, *args, **kwargs )
  File "include\HydrusDB.py", line 824, in Write
    if synchronous: return job.GetResult()
  File "include\HydrusData.py", line 1668, in GetResult
    raise e
DBException: TypeError: buffer object expected
Database Traceback (most recent call last):
  File "include\HydrusDB.py", line 517, in _ProcessJob
    elif job_type in ( 'write' ): result = self._Write( action, *args, **kwargs )
  File "include\ClientDB.py", line 9471, in _Write
    elif action == 'content_updates': result = self._ProcessContentUpdates( *args, **kwargs )
  File "include\ClientDB.py", line 6621, in _ProcessContentUpdates
    hash_id = self._GetHashId( hash )
  File "include\ClientDB.py", line 3426, in _GetHashId
    result = self._c.execute( 'SELECT hash_id FROM hashes WHERE hash = ?;', ( sqlite3.Binary( hash ), ) ).fetchone()
TypeError: buffer object expected

wat do?

>>5791

It's currently failing like that all already deleted pictures. It effectively does not impact an user because it's just marking them as failed rather than "already deleted".

>>5720

I only have ~1/3 of that now. Would be more if everything worked smoothly, but I get UI lockups when generating phashes and so on.

Didn't get to add more.

> >>5752

> Do you have to straddle multiple hard drives, or does it all fit into a 4TB or whatever?

Since he didn't answer, my set is actually under 1TB. His might still easily still fit on a single 4TB drive, yea.

And it's not like 8x6TB or whatever already requires particularly specialized hardware or a fortune. Just throw ten 6TB drives into a pretty standard Linux / big tower machine with one-two add-on SATA cards, RAID6 over it all, and it you got 8 drives worth of storage without much trouble in setting it up or operating it.

As a small remark: At a bigger LAN party around 15 years ago, there was a guy who had a stack of 40 drives on simple rails, and he apparently had left another 20 or 40 or whatever at home. All connected to bigger 3ware enterprise cards cards. [That was fairly impressive to look at, even though it's not really clever - they even apparently performed pretty shit under his Win 2k or whatever it was back then.]