/hydrus/ - Hydrus Network

>>14438

here

Just wanted to confirm the problem is indeed resolved. Thanks for all your hard work.

>>14475

Great, thanks for letting me know.

I had a great week doing small work. I cleaned a heap of old code, fixed some bugs, and added some quality of life.

The release should be as normal tomorrow.

I had an ok week, though a little short on work time. I tweaked how subs work for the new system, getting them booting a little faster, re-added import/export/duplicate for subs (including importing+conversion from the old format), and cleaned up some bugs, including recent problems with linking certain downloader components together.

The release should be as normal tomorrow.

I got delayed on messages this week, I'll try and catch up a bit now this evening.

>>14439

>>14435

>>14440

Thanks lads. Keep on pushing.

>>14443

>>14446

Yep, I'm just a soulless madman. I typically put out a release every Wednesday, trying for by 8pm EST. Tomorrow will be v401.

The code is imageboard tier, so while it can do some wacky fun stuff, it is also permanently a bit shit and always in beta. The ride never ends.

If you missed the help, check it out in your release under the client's help menu or here: https://hydrusnetwork.github.io/hydrus/ It has an extensive 'getting started' section that talks about updates and the other basics.

>>14438

Damn, thank you for this report. I believe I have fixed this for tomorrow, please let me know if you have any more trouble.

>>14436

Thank you for this report. I am sorry for the trouble. There are actually three things going on here:

1 - it turns out subscriptions fail to gracefully deal with gallery url tests when the url has no current definition

2 - e621 recently changed their default gallery url format, and the new default downloader did not still support the old one

3 - the new subscription tech sometimes randomly samples a slightly older previously visited gallery url for bandwidth and login tests

For 401, I have fixed 1 and also added a separate definition in to more nicely cover 2. This problem should be completely gone, please let me know if you have any more trouble.

>>14451

Thanks. v401 seems to have fixed the problem

>>14415

>I am not totally sure, but I think that 'remove arrows' options causes the tab bar to just grow in width, which would cause the whole hydrus client to end up growing massively wide. I had a quick look, but I didn't see a mode to tell it to do nicer multi-row tabs–do you know if this is possible, maybe with QSS? I'll happily turn it on if there is.

Oh, it grows in width, that's dumb. I looked a bit more and yeah it seems Qt doesn't support multi row tabs. I'm more familiar with C# and the like where it's as simple as changing a property to true, thought it would be the same here. That's a shame, it's so cumbersome now when you have like 50 tabs open.

Maybe there could be button or menu with a dropdown list somewhere listing all the open tabs and you could select which one to switch to there? Just a thought.

>>14417

Yeah, a dropdown would work, if I am allowed to fit it in. Some of the 'notebook' code is still a bit borked by wx->Qt hacks, so once I get to clearing it out, I'll have a look at how much I can edit it. I think I can replace the QTabBar with my own subclassed control, so that's a possible. In my searching, it looked like there is some code floating around that may do multirow already, so I'll see if that is available to use in python.

>>14429

Shortcut keys for next/previous tab might also help make it faster to navigate

>>14431

Thanks, I will add this to my todo.

>>14442

Accidentally created a new thread. Please delete this.

>A better alternative

seems to be closed source so automatically inferior.

> Call me what you want.

This is a non-discrimination zone. If your media is within the prescribed formats you are welcome here.

>>14331

Hey, making downloaders is quite advanced, but if you know web tech like HTML and JSON, it will be easier. My help starts here:

https://hydrusnetwork.github.io/hydrus/help/downloader_intro.html

That said, there are many user-created downloaders here:

https://github.com/CuddleBear92/Hydrus-Presets-and-Scripts/tree/master/Download%20System/All-in-Ones/Single-Sites

This looks like it could be for the site you want:

https://github.com/CuddleBear92/Hydrus-Presets-and-Scripts/blob/master/Download%20System/All-in-Ones/Single-Sites/allthefallen-2019.08.10.png

But I am afraid I have no idea if it works. If you haven't done it before, get the png and drop it onto Lain under network->downloaders->import downloaders.

>>14344

Thank you. I got it on there, but now every file I try to download from there is ignored. Is there a way to disable ignored files? I'm having the same issue with another downloader as well. Its only on the ones I added though.

>>14342

>>14340

I came here to ask how I can import/export files using their tags as their filename while preserving namespaces, but I may have found my answer. Could this be used to filter export names to use semicolons in place of colons and vice versa for incoming files?

>>14327

>>14330

>>14337

Thanks lads, keep on pushing.

>>14340

Thanks. I grew up in the UK, where you use s or c depending on if it is a verb or a noun, like advise/advice. I get mixed up a lot myself, so I'll just change it to the US spelling to be clear.

>>14342

>>14343

In the past two weeks, I have updated the downloader system to do much more flexible string processing. This new stuff can go nuts, doing filtering and converting and splitting in different orders. My next step is to copy that ability to the existing filename tag parsing panel, which is all specific old hardcoded solutions. If you like the regex, I hope soon to really improve it, and I'll appreciate your feedback when it happens. I want to keep the simple 'just tag the filename m8' workflow but also let users do some neat regex stuff if they want.

I hadn't considered this tech for exporting tag filenames, but it may make sense. The current export filename 'phrase' system is another hardcoded thing, and I like some of it and dislike other bits. I don't think you can export actual namespaces unless you do something weird like:

creator;[creator] - {file_id}

But that only works if you have one creator tag, to get 'creator;whatever - 123456.jpg', otherwise you'll get I think 'creator;whatever, other - 123456.jpg'.

I was going to update the 'phrase' to the 'tag summary generator' tech youPost too long. Click here to view the full text.

>>14357

In order to change fixed character/characterset keys to different values without that error appearing,

1. regex should be selected

2. example text changed to match intended key value

3. appropriate key type should be selected.

The example text field isn't visible outside of the regex key definition, so it cannot be changed from the default example or previous example it complains about.

It doesn't work that way. You have to get the files yourself. Hydrus is mostly for organization of files you already have, but it does have features to facilitate getting more.

Also, from the way you talked about "following the best and most active tags", it seems like you don't really understand how tags work either. Take a look at the hydrus manual. It'll explain everything, and, with some exceptions, it's worded simply enough for beginners to understand.

https://hydrusnetwork.github.io/hydrus/help/index.html

>>14306

>>14285

>>14308

I found a solution that works.

It's all in the client already, too!

So, first you go to your trash, hit CTRL+A to select every file.

Next, you right click and go to "share > copy > hashes > sha256".

Now you have all your trash files in the system clipboard.

Next, you open up a new page, use "system: similar to", select your parameters (hamming), paste all of the hashes you copied from the trash, and there are all your similar files!

Theoretically, at least. It will take a couple of minutes for 200 files, but it's working!

Thanks!

>>14283

Thanks m8, keep on pushing.

>>14289

Yeah, I have search and import/export+client_api in my big notes job. The import/export will have some note renaming capability, so when we come to it, please let me know if that does what you want, or if you absolutely need regex capability. It might be that is easier to do with a client api script than writing it into the program.

I don't know when I will get all this done, but I hope to keep moving forward a step at a time in normal weekly work.

>>14292

Hey, great, thanks for letting me know. I reworked the whole pipeline a couple weeks ago and have been cleaning it up since, so it times queries and caches results much better.

>>14293

Seconding modified time features, would be really useful. Would also be nice to be able to store different times from different sources and use the earliest by default.

>>14311

Thanks. I am thinking about it. I don't really want to add the additional data structure of a per-source 'posted time' timestamp, since it is a bunch of extra overhead, but it does make it easier to present a modified date.

Also, if I were to build this sort of thing and it turned out most boorus had upload times within a few days, it might not actually prove that useful. I like this idea of bumping the decision to file import options. If a merged modified date fails to account for an interesting situation where different sites' modified dates differ greatly, maybe it could be extended.

I am also thinking about the issue that most of us have huge collections with fairly bad modified dates already, and going back to sites to reparse this data would be a big old waste of bandwidth. I'm thinking of writing a small import/export system for modified dates, so users can share this info. I'll get good new modified dates going first though, and we'll solve the retroactive problem later. Just need to find time for it.

>>14318

This is cool!

>>14209

Thanks m8, keep on pushing.

>>14210

Ah, thank you. As it happens, my test machine only had three tags there, and they happened to sort by accident. The old manual sort call must have been lost in the rewrite. I'll sort them using the normal lexicographic sort there for 397.

>>14214

Thank you. I have several reports on this now. It is always the bottom for some reason, even on long lists. I will fix this for 397.

>>14216

That's the hope. It'd be nice to have some sort of .txt import/export, and mass import/export with an external Hydrus Notes Archive and a new 'notes migration' panel, and Client API access for notes. I hope to slowly push on this in future normal weekly work.

>>14217

>>14231

>>14243

The server is still up. Some weird dude got upset and started spamming TOS-violation stuff to the server along with some friends, so the old invite url got turned off. Assuming the users who run the discord server are ok at that point about putting a new invite up, I'll fold the new on into 397 and all my help links.

>>14219

>>14223

>>14246

Thanks. I agree. The reason for that sibling hard-replace change was because the text on those menu items was too long, and it was getting text…elided in a way where you couldn't see the difference between the two options. I split it up as a quick patch so users (particularly new users) could actually read the difference.

And I totally agree on the super nested thumbnail menus. There are an increasing number of commands in hydrus now, and for things like the duplicates system the available menus are flexible depending on the thumb or selection status, so you can get annoying stuff like in your image with a submenu with only one submenu item, just wasting space and time.

I am happy that that big thumbnail menu is now listed essentially by Post too long. Click here to view the full text.

>>14248

>As someone who uses this a lot, what do you think you would like?

Either would work, but since it already shows a confirmation window with suggested changes maybe a description and single/all could be integrated into it?

>What other menus do you find annoying?

Mainly things in the thumbnail menu that used to be easy accessible in older versions being moved into submenus and lower on the menu, for example "remove > selected" "manage > tags" and "open > in external program". And when you right click a tag "copy > x".

Maybe it's because I learned the older layouts first that it was annoying for me. The menu is better organized now and a little more cumbersome, but perhaps better for new users. I just learned to use shortcut keys instead where possible.

Menus aren't great in other places but it's less of an issue there since it's not ones you use often.

>>14208

>Beyond that, I'd love some note 'preview' on the media viewer and comprehensive import/export/Client API support for notes.

Does that mean there's a chance that we could get something similar to how it works on Danbooru or am I misunderstanding? I don't know if that's feasible but it would be amazing. I'm talking about stuff like this:

https://danbooru.donmai.us/posts/3900728

I had an ok week. I fixed some bugs, improved some quality of life, and added a new system predicate to search for notes by name.

The release should be as normal tomorrow.

Thanks. Some weird dude got upset and started spamming TOS-violation stuff to the server along with some friends, so the old invite url got turned off. Assuming the users who run the discord server are ok at that point about putting a new invite up, I'll fold the new on into 397 and all my help links.

My list of favourites tags in the manage tags dialogue is in complete disarray and no longer sorted alphabetically for some reason

>>14175

Same goes for the recent tab

>>14175

>>14176

Hydev on tablet: SHIT. Three steps forward, two back. The new filter must be killing the sort. Sorry, I will fix this for 396.

I had a good week. The hydrus file 'notes' system is improved, with all files now supporting multiple notes, icons on files to show when they have some, and the ability to search for files by the number of notes they have. Also some improvements to unusual tag autocomplete and a variety of bug fixes.

The release should be as normal tomorrow.

>>13997

Hey Dev, thanks for the release. What's up with the Favorite tags disappearing when you use them, and not coming back when you go to the next image? Really puts a damper on my workflow.

I think it's great that they disappear when you used them, but they really should come back every time you switch images. Unless I'm missing something.

>>13999

Thank you, this has slipped past me a couple times now. I will make sure this happens this week, and generally improve all the suggested tag panels. Sorry for the trouble.

I had a good work week. I fixed a variety of bugs, improved the feel of the suggested tags columns, and made zooming in the media viewer nicer.

The release should be as normal tomorrow.

If you get an error with your subscriptions about 'DomainOK()', please pause your subs under the 'network' menu. I am sorry, I will have this fixed for 394!

>>13959

Sorry, I will make sure I put time into this this week.

>>13970

Thank you, I will fix this so it re-synchronises when you set a new media.

>>13974

Ah, are you running from source? I got this problem (and plenty more) when I tried out Qt 5.14. I now use Qt 5.13.2 on all platforms.

>>13986

>>13974

To follow up, you want to do:

(get into your venv)

pip3 install PySide2==5.13.2

Or for PyQt5, whichever you are using.

>>13987

>are you running from source

Yes, but I had no luck running by installing the dependencies through pip (Shiboken error). I just use the dependencies listed from the Arch AUR package and then git pull your repository and run it.

Just tried running the compiled binary and the same thing happens. Looks like it some distro-specific shit so don't bother with it.

I had an ok week. I mostly cleaned code, particularly improving how tag autocomplete text is parsed and the found results cached and filtered. The whole thing feels just that bit snappier now. I also fixed the accidental 'resuming half-complete subscription’ bug from last week's release.

The release may be late tomorrow.

I had a good week catching up on messages and general todo. As well as bugfixes and quality of life, I also prototyped a solution for basic cloudflare problems.

The release should be as normal tomorrow.