/hydrus/ - Hydrus Network

>>7015

>APNG imports are also fixed. The recent decompression bomb stuff was accidentally breaking them.

Oh nice, I thought that was a problem with pixivutil

Parsing namespaces means what? English is my second/third language, so it might be really obvious. I think it might mean what namespaces I use, but I'm not sure, can anyone give some explanation?

> - the manage tags dialog will now protest with a yes/no dialog if you attempt to cancel it with uncommitted changes

Thanks for this dev!

>>7020

Ok, after importing a shit ton more tabs, the program slowed down a bit, it is importing the urls, but it tells me 0 images found, I think that's because it already has the pages urls, however being told point blank instead of "0 new urls" only it would be better to say "0 new urls - XXX already imported urls"

My Christmas wish this year is for you to accept third-party contributions and considering porting to Python 3, this would open the doors for much quicker progress and at the same time guarantee that Hydrus continues thriving with an actively developed programming language.

I know you have read this a few times probably already, I am also aware that it's probably quite a bit of work to porty everything, especially if any third-party libs are only compatible with Python 2 – however I think those cases should be extremely rare, and for most of them a similar alternative with a Python 3 version should exist.

So, I'm not sure if this is on your todo list yet, but since you keep one I thought I'd atleast mention my wish.

Thanks again dev, and please don't take this as me being ungrateful or not appreciating the work that goes into that.

>>7022

I hear the python 3 thing a bit, Is there any actual reason to move to 3 over 2?

From the bit I gleam, everything good about 3 got backported and there is no real advantage to performance…

>When you select which namespaces to parse, do you:

I sometimes parse all but generally a few.

>>7018

I suppose it means what namespaces you import when downloading from e.g. a booru.

>>7023

This is wrong, that might have been the case with the first few releases of Python 3, but for example a core data structure, the dict(ionary) is much faster in Python 3 and only the most crucial features have been backported.

Even if Python 3 _was_ (I am not saying it is) slightly slower than Python 2: for an application like this it would be irrelevant. Things that need to be fast: the GUI toolkit, ffmpeg etc. are written in C/C++ and Python just makes use of them. In other words, if you chose Python 2 specifically for any speed advantage over Python 3, your application clearly doesn't have to be super fast, or you wouldn't even be considering Python.

The standard library is much improved, Unicode support is much more sensible with saner defaults and around 2020 support for Python 2 will be completely dropped, including future security patches (currently only important security patches are backported).

There's no other choice than to switch to Python 3 sooner or later, and it will only be advantageous. A Google search could explain this probably much more completely and eloquently than I can though.

>>7026

I did the google search and up to 2 years ago people were still finding it hard to justify using python 3 over 2, that and documentation, even 6 years on after 3 was introduced was still not near python 2.

from the way it sounds, the only real advantages python 3 has is development side, and by the time you are to the point those advantages would help, they stop being things you would even care about.

but this is all stuff that was about 2 years old as of now.

From the sound of it, on the user end, nothing matters, there is no clear performance gain one way or the other, and on the development side, documentation is an issue.

For what it's worth if everything we both said is true, it would probably be best to get all the features into the program, and then work backwards with the porting from there rather than doing it now.

Ok found an issue when I went to do an experiment.

With the duplicate detector I wanted to see something and know for a fact it works

I personally do not use tags as its cumbersome currently, instead I use ratings, and for various reasons with the duplicate detecting I have halted doing this en mass, and it seems like that was the right call.

———————–

The default action on setting a better/worse pair is to move all local tags from the worse file to the best (i.e. adding them to the better file and then deleting them from the worse) and then sending the worse file to the trash.

———————–

Took a while to find 2 images in dups that were tagged different and both duplicates, selecting one as the better one did not add the tags to the better one

for ratings I went into settings and saw there was a way to add these to a move/copy but only one at a time instead of all.

the issue I can see here is when I get to phase 2 or further of sorting needing to add more and more ratings to be able further sort needing to remember to go back to the duplicate and sort them out there.

Not that I even know if I added said rules they would even take effect seeing the tag ones don't/arent on my db at least.

Not the biggest issue for me personally right now but it is still a concern.

Thank you based dev.

>>7015

>- tags imported or exported via neighbouring .txt files are now correctly tag censored

Can you elaborate on this? I'm not sure what it means exactly. Is there a syntax for censoring stuff in the sidecar txt files?

Thanks as always hydrus.

>>7022

>accept third-party contributions and considering porting to Python 3

Bang this both out in 1 go with a third party port ;^)

Wasen't there a project someone had written in Golang that was compatible with hydrus or was that something else entirely?

>>7039

>I am perpetually neck-deep in work, so scheduling a convenient week or three to do these rewrites will prove tricky

>still bangs out weekly updates, fixes and improvements for Hydrus on 3 platforms

Fucking how? How do you do this? Even neets don't have this level of productivity

>>7043

>>NEETs

>> productive

I think you might want to rework your comparison there buddy.

>>7047

Point taken.

>>7039

I'm in the process of closing tabs, I don't have enough closed to do a remotely large scale test,

for what its worth on tabs pausing, they were tabs that weren't 404'ed and were still 'active' just not getting images, they paused and when I pressed check they doubled the image count.

I don't know if it's a by product of an older version or if they were pausing even with my settings that should make pausing impossible.

As for the move to newer things, I would suggest getting everything done that you need to do immediately, and any major issue that needs to be hammered as I can't imagine anything user side will happen for a month or so while you do the full port.

>>7037

I use the pixiv artist id downloader and check "creator" and "title", but "creator" doesn't seem to do anything so I tag the artist name myself. "unnamespaced" on Pixiv would be mostly Japanese tags and I can't read those so I don't choose it.

manga/ugoira support when

Thanks for the answer!

> That said, if you are keen to write a python module I could import, there are plenty of neat things I'd love to not have to do myself. I am particularly interested in neural network and audio stuff at the moment, if you have any experience.

I'm not sure if I could be of help with that, I haven't done much in either area. However if you could outline something like this in more details (think a "public todo wishlist"), maybe I or some other people could take a look at it.

i just downloaded and dumped my furry lewds folder into the program, its importing the pictures now. the thing is, its a nearly 70k-pictures folder, with several subfolders within.

is there a newbie guide thread or something? i dont think i can tag 70k pictures on my own even if i spent my whole life on it, there's probably a better way to do it but im totally new to this.

I made an alternate Linux build for people having trouble with the official build. A few things are still weird, but overall it works. Run from source works slightly better, but who wants to muck around with pip? Built on Linux Mint 18, so it should also work on Ubuntu 16.04.

http://www.mediafire.com/file/gbrg4sk0a5kshdx/hydrus-278.tar.gz

I'll post one of these every week until the official Linux build works on my machine. If anything is broken, tell hydrus_dev (and thank him while you're at it). He was kind enough to give me his build command, but I don't know Python. If something is broken, I'm just as clueless as you are.

I wish there was a way to right-click a tag in the tag manager window and select a "Replace with…" option, which will open a window where you enter a new tag to replace all the occurrences of the selected tag with. Most useful when you have multiple files selected of course. Siblings are sometimes a too broad alternative since it doesn't only affect the selected files.

Dev?

>>7057

Connect to the public tag repository and you'll probably get tags for most of them

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

>>7057

Ratings

my god ratings.

here are mine

real/drawn

this sorts things into 2 piles

safe/suggestive/explicit

this further divides them into 6 piles

funny

now this now goes further to 12 piles

art

this takes the pile list to 24

and pony

this pushes things to 48 piles

with modifiers of favorite and tag further signifying something is there I want to keep track of.

from this list I will tag things like drawn into a further 3 piles of

stand alone/one page/multi page

which takes 24 of the piles and puts it to 72

western and eastern

which puts it to 144 piles

so on so forth.

The current way tagging works is cumbersome, I cant say that I like it, but with ratings, and very simple things you can sort out a fuckload of images in short order

If the program does not hang on me I can sort out 1 image every 1-2 seconds. some of this can be done far more rapid then others.

just focus on a few small things and make sure you have at least one tag per phase of sorting that signals the images were dealt with. for phase one with my, its safe/suggestive/explicit as every image that passes through me has to be rated one of those while most of the other tags can be unrated.

also have something in the background playing when you do this, otherwise the task will be far to tedious.

>>7068

I have a question, have you ever given thought to converters in the program?

let me lay this out, I wanted to see what my largest images were, so I went 15mb+ the biggest thing in the program is a fucking 17,000 by 24,000 image that some retard thought was perfectly reasonable to make a png ,I just had a look at it and it IS detailed enough, kind of… its obvious it was an artist source image but this is stupidly large file size for what it is.

taking it to photoshop even a 10 puts it down to 15mb if jpeg and 8 gets it down to 10mb and that's not even touching the fucking stupid dimensions it uses.

The next image is 50mb, another should have never been a png image.

There seem to be a substantial amount of what the fuck file dimensions and are you retarded file type choices.

oh my god, 40k by 16k png at 84mb

So yea, an in program converter for some files like this. something user activated, that only does one image at a time, then leaves both files, flags itself something for dup detector so you can make sure you are ok with what you got before you remove one.

jesus christ, I thought the top of the list would all be video, not an image.

On another note, is there a way in the downloaders to sort by order they were found?

let's say I downloaded a /a/ thread that had a manga uploaded and was in order,

Hydrus before you click on anything, at least for me, defaults into the drop down

this means If I scroll up or down, the order in which the images were 'acquired' changes.

If I go by sort by time imported, it skips the current import order and defaults to the first time the system found it, so with one window where I downloaded I think 67 threads, each thread has its own theme to walk into, if I wasn't careful I would have 3330 images in the order I want them in to go through, but I would have 1145 images technically in order, but also wildly out of order at the same time.

on a final question, is there any way to indicate with a webm or whatever else the program can handle if it has sound? not even a play it in program, which would honestly be awesome, but just something that would tell you 'hey, you may want to open this one externally'

>>7074

Holy shit, is that windows 2000?!

ran into a convenience problem.

When I made my ratings list, i'm damn sure they weren't on/off buttons otherwise I wouldn't have gone through the hassle of remove X buttons on

so, here I am, i have a good 150 images to go through, I already looked at everything's thumbnail, this was a thread where I loved every image, and I want to rapid fire the 'favorite' rating… 20 images already in archive…

so now instead of images loaded alt+b right rinse repeat, I have to have the mouse in the upper right corner, wait for image to load, identify if I favorited it, then go through motions.

Is there a way in program to make them explicitly what I tell them to be, without an on/off state unless I want it to be that way?

I think this could be useful for when I have a macro pad full of shit to tag, but for the rapid fire I want to get shit done, it's more of a detriment to have to check if It was ever rated before.

>>7037

sorry for the late reply, I almost never use that, as I only have 'real' girls and lewds in my databases. So no input from me