/hydrus/ - Hydrus Network

Hydrus Bunker

Posting mode: Reply

Check to confirm you're not a robot
Name
Email
Subject
Comment
Password
Drawing x size canvas
File(s)

Remember to follow the rules

Max file size: 350.00 MB

Max files: 5

Max message length: 4096

Manage Board | Moderate Thread

Return | Catalog | Bottom

Expand All Images


Q&A Thread Anonymous Board owner 08/08/2019 (Thu) 00:24:05 Id: 348093 [Preview] No. 10
Please feel free to ask questions about hydrus here.

As a reminder, the help and getting started guide is here:

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


Anonymous 08/16/2019 (Fri) 03:11:11 Id: 2a6070 [Preview] No.34 del
I've been trying to scrape some things by character tags off gelbooru lately, but I notice that hydrus can only seem to find between 50-70% of the content for the character. Is this something on gelbooru's end messing with it? For example, there are 770 images tagged as 'hacka_doll_3', but hydrus only finds 464. I assume it's something to do with lolicon/shotacon tags. I've not tried making a gelbooru account yet, but I don't want to have to pay for anything just to save art that's free to access.


Anonymous 09/27/2019 (Fri) 22:12:31 [Preview] No.125 del
Hey, I tried merging my windows hydrus db with my linux hydrus DB by clicking file>import and export folders>manage export folders and then in the destination hydrus client, adding the exported files via file>import files>add folder.

However, this doesn't seem to preserve tags so I was wondering how I could do that?

Additionally, I noticed most of the .webm, .mp4, .webm files failed to import to the new client. Any help would be greatly appreciated!


Anonymous Board owner 09/28/2019 (Sat) 22:39:13 Id: 273f4f [Preview] No.137 del
>>34
Yes, I think it is. There's some cookie or something for gelb to unlock spicy content. I'll be extending the login script system to allow adding static cookies so some other users who are working on a new login script can add this, but for now, your best bet is to either add the cookie manually (check your web browser, once you apply the 'show all' or whatever it is form on gelb, and then load up in hydrus network->data->review session cookies), or use Hydrus Companion browser add-on to sync it across automatically:

https://gitgud.io/prkc/hydrus-companion


Anonymous Board owner 09/28/2019 (Sat) 22:42:44 Id: 273f4f [Preview] No.138 del
>>125
Check the new 'tag migration' system under services->review services. I assume the tags are under 'my tags', right? Go to that services panel on review services and then hit 'tag migration'. That'll let you export all current tags to a Hydrus Tag Archive. You can do the same on the other end to import to your 'my tags'.

For the videos that failed, what's your FFMPEG version under help->about. For the Linux release of hydrus, it uses whatever your OS has. If it is old (I think current version on Windows is 4.1 or so), try updating it on the system level or putting a static built executable under hydrus's install_dir/bin. There should be a readme there explaining the situation as well.

Otherwise, do you have any of the error traceback 'notes' for the failed video imports? This might help figure out what went wrong. Click the little icon button on the importer that failed and find the import row that failed and then right-click->copy notes.


Anonymous 09/29/2019 (Sun) 10:59:43 Id: e9d7d9 [Preview] No.139 del
(6.47 KB 757x362 fringeBenefits.png)
(1.83 KB 263x99 gelbooru.png)
>>34
Yes I can confirm that Gelbooru requires you to set a cookie to access all the content. This is because they recently added an option on the website which hides loli by default. The setting doesn't require an account to toggle because it simply uses a cookie to check if you've enabled it. This means it's pretty easy to set in Hydrus.

Go Network -> Data -> Review session cookies

There should be a cookie for gelbooru.com. The setting you need to add uses name "fringeBenefits", and the value is "yup". That's all you need to do, Hydrus should then import all content.


Anonymous 10/01/2019 (Tue) 05:59:25 Id: a5a141 [Preview] No.140 del
So I have a weird question/idea about image collections

I have a lot of porn, a lot of amateur stuff.

Now, whats the best way to tag all this outside what you see, how do you identify a person assuming you don't have a name or pornstar name for them.

I'm thinking maybe some kind of either metatag or some kid of hidden identifier tag?


Anonymous 10/02/2019 (Wed) 00:22:01 Id: bf8e84 [Preview] No.143 del
The help guide and the discord make it sound like there's a big collection of pre-tagged images that other hydrus users share publicly. (not from booru sites, images from other users)
How do I access this? Does it have any good memes?


Anonymous 10/06/2019 (Sun) 02:57:17 [Preview] No.156 del
>>10
So, I was wondering if there was a way to order my images when they display in hydrus? So, like, before they were in files and had names so I could rename them and ensure that, for instance, panel 1 of a 3 panel comic always came first, and the 3 panels were always grouped together, but I had a large number of duplicates because of the lack of tag support so things that were both "memes" and "comics" would get copied bloating my drive(and there is some shortcut and sysmlink BS I've pulled to help with redundancy with larger collections of like images). Now in hydrus after importing some test images I recently grabbed I'm seeing that nothing has names, but it is out of order. How can I force certain elements to be grouped in an ordered subset when they are visible? For instance when I have a tag "dilbert" with a parent "comics" and I want all the dilbert comics to appear when I include the "dilbert" tag, I want all the comic strips to be grouped together with the panels in order, a consecutive ordered subset, folowed by the next comic strip, also in order, and so on. I expect I will also have an issue when importing full manga and comics. There are also some images from models I have that I don't have a good name for, and where I don't know the name of the model, but I would still like to be able to select them all and group them, and give them an order so similar images appear one after the next(If I have to manually order them this is not a huge deal, though I have enough doujins in particular it would be nice if I could import that based on the file name somehow so I don't need to manually re-order thousands and thousands of files).

Also, I have a slightly different question. I noticed there is some kind of special tag for system inbox and system archive, but I'm not sure what the difference is? are they just local collections, or something else, and if so can I make my own custom collection and sub collections?


Anonymous 10/06/2019 (Sun) 03:00:01 Id: c8154d [Preview] No.157 del
I can't find anything in the help on how to set a subject name for watchers. And it doesn't seem there's a way to change it once added. Am I missing something?


Anonymous 10/06/2019 (Sun) 09:09:52 Id: df5f0f [Preview] No.158 del
Has anyone been using an image viewer to interact with Hydrus? I know it has its' built-in image viewer but I've been thinking about something more minimalist, something without any menus, only a thin border that'll change the the shown image at every 10 minutes - this could be configurable - or by a shortcut. In other words, an image viewer that'll use hydrus as the backend. Maybe this could be added to Hydrus even but it seems like too much to ask from hydrus dev.


Anonymous Board owner 10/06/2019 (Sun) 22:23:05 Id: ea6894 [Preview] No.163 del
>>140
My biggest thing about tags is that they are for searching, not describing. If you are likely to search for 'amateur' or 'blonde hair' or 'favourites', that may be much more useful to you than person names. If you do not know the name, I recommend you leave that blank until you do. If you sync with the PTR and these are likely to be on some 'realbooru' or similar, another Anon will probably eventually add the right person: tags for you either manually or by parsing a booru they are on.

If these are for 'my tags', you can of course do anything you like, but in general, I have learned that trying to tag everything you see, or try to do anything like a 'perfect' job, is a rabbit hole with no end and very low productivity.

You could try a kind of 'artist request' tag for 'person' if you like, like 'unidentified person', as a 'my tags' tracker just so you can revisit these files at some point.


Anonymous Board owner 10/06/2019 (Sun) 22:26:58 Id: ea6894 [Preview] No.164 del
>>143
I do not know of any large public file repositories. I know some users are playing around with them in private, mostly 1-5 friends sharing kind-of-thing.

I used to run a simple file repo that had a couple hundred of my old imageboard files and some fine art, just as a demo for how the tech works, but I no longer run this. Perhaps that was what you had heard about?


Anonymous Board owner 10/06/2019 (Sun) 22:33:07 Id: ea6894 [Preview] No.165 del
>>156
Hydrus is not excellent at paged content. I hope to improve this in future, along with more general cbr/cbz support. For now, try adding 'page:1', 'page:2' and 'chapter:10', 'volume:5' and 'series:blame!' style tags. It takes a little effort to set up, but you can often regex-parse these from your original file import filenames. Then, in the client, set it to sort files by 'creator-title-series-volume-chapter-page'. This will line things up nice, but the workflow, particularly for editing things once in the client, is not as I'd like it.

inbox/archive is a hardcoded special 'tag' that works like your email inbox. New files start 'inbox', with the little envelope icon, and then if you decide you want to keep a file (e.g. it was not some garbage you downloaded by accident), you hit F7 to archive it. The system:inbox/archive system search predicates let you filter by these statuses, letting you set up a 'filter this new stuff' page or 'look at this good stuff' page. It is just a neat way to keep track of new and verified 'good' files. Check out the 'archive/delete' filter as well, as per here:

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


Anonymous Board owner 10/06/2019 (Sun) 22:35:34 Id: ea6894 [Preview] No.166 del
>>157
It is parsed from the website at the moment. It is not currently user-editable. If you edit the parser, you can change what gets added, but this is fairly advanced.

If you are making a thread parser and want to see how it works, check out the existing 4chan/8chan defaults to see how I did it. There's a special content type you attach to your content parser, and that value gets sent to the downloader object for its 'name'.


Anonymous Board owner 10/06/2019 (Sun) 22:37:08 Id: ea6894 [Preview] No.167 del
>>158
The Client API would let you do this. I know some users are working on booru-like Phone App wrappers for the client using this, but not this particular case of a different desktop viewer.


Anonymous 10/07/2019 (Mon) 00:59:59 Id: c8154d [Preview] No.170 del
>>166
You know what, I've been looking for the subscriptions function this entire time and just found it.


Anonymous 10/07/2019 (Mon) 01:26:09 Id: f60503 [Preview] No.171 del
Due an issue on my HDD, I've lost some files from my data folder. Specifically, 6 full folders, to the tune of 3600 files. I have a recovery program running, so I might be able to find them. But if not, I have the thumbnail folders still available, so I could reverse search to replace the files manually. If I did this, ended up with a slightly smaller version of the file, put this file into the correct folder with the correct name... What would happen? Would it break my database? Would it still be tagged with the previously saved tags?

tl;dr: If I replaced an image with another of the exact same name, but it didn't have the exact same properties, what would happen? Would the file still have the same tags? Or would it just bork everything?


Anonymous Board owner 10/08/2019 (Tue) 00:31:24 Id: ea6894 [Preview] No.179 del
>>171
Some different things would happen depending on various maintenance timers. The client does not constantly monitor its file store, and it only ever accesses a file when it needs to do something with it like showing it in the media viewer. It does not rescan a file when it loads it.

If a file were to change, the client will try to load it with its current understanding, so if you swapped one jpeg for another with a different resolution ratio, it would probably draw it ok but all stretched to the wrong size canvas. It wouldn't lose the tags for those files, as tags and so on apply to the db store's understanding of the file, which is still all correct for you. It might sperg out if it needed to generate a thumbnail or something else clever.

A problem would come if hydrus ever did file maintenance on those files. It would recognise the files had changed in terms of bytes and throw some errors, and probably remove the file and put it in a 'missing_and_invalid_files' directory in your install_dir/db directory. Maintenance can be triggered by things like thumbnail resizes gone bad, so I expect these weird files would get looked at eventually.

I recommend you make a copy of those thumbnail folders, just in case you want to try something clever with them later. Then get the client going with empty file folders and go database->maintain->file maintenance->review scheduled jobs and then add a 'check if file is present in file system' for a 'system:everything' search. Force it to do that job, and it'll clear itself of the missing records and export all those files' known urls to that 'missing_and_invalid_files' directory. There will be a 'all_urls.txt' masterfile, which you can open, copy-all, and then paste back into a 'urls' download page in the client. The client will then download all those correct files again. This will make for a more 'correct' db and get you nice files again. It would not work well if most of your files were not downloaded and do not have nice known urls.

If you wanted to have thumbnails in lieu of the files, it is probably safer just to import those thumbnail files, although they won't get the tags of course. There's the slightest chance that your png thumbnails will be the same file as the original png, but I can't say that with confidence.

Let me know how you get on here.


Anonymous 10/08/2019 (Tue) 12:26:27 Id: febc66 [Preview] No.185 del
Hydrus isn't saving any page changes. Whether I import files in a URL import page, close a page, or open a new page; closing the client and reopening it reverts any changes. The files I import are still in the database, but they don't appear in the page where they were imported. Any idea how I can fix this?


Anonymous 10/08/2019 (Tue) 15:17:15 [Preview] No.186 del
>>10
>>165
How do I do regex parsing on import?


Anonymous 10/09/2019 (Wed) 10:52:41 Id: 5b136f [Preview] No.192 del
>>167
Well, it's nice to hear at least Hydrus makes the idea possible, maybe I'll try my hand later after finishing studying Python. Using Hydrus as a backend while having a minimalist image viewer somewhere on the desktop seems like the dream.


Anonymous 10/09/2019 (Wed) 12:51:45 Id: 5b136f [Preview] No.193 del
What is the maximum limit of thumbnail size? What do you guys use it?


Anonymous Board owner 10/09/2019 (Wed) 22:55:16 Id: ea6894 [Preview] No.197 del
>>186
It only works for hard drive imports, and works on file paths. When you drag and drop some files onto the client and it opens up the window with the list of files to import, click the 'add tags based on files' button, which will launch a new, more complicated dialog. The 'advanced' tab here lets you match tags based on some regex rules.


Anonymous Board owner 10/09/2019 (Wed) 22:58:50 Id: ea6894 [Preview] No.198 del
>>192
Here is the API stuff, if you are interested:

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

There are links to several libraries, and a nice python interface that makes it easier. The API isn't 'complete' by any means, but I am slowly adding more to it.

I was just told about this module, which I will be adding to the page for next week:

https://gitgud.io/koto/hydrus-archive-delete

If you get into it, let me know if you run into any trouble.


Anonymous Board owner 10/10/2019 (Thu) 00:10:05 Id: ea6894 [Preview] No.199 del
>>193
There's no practical limit, I think it is something silly like 4096x4096. My main machine has a 4k screen, of which my hydrus window occupies about two thirds. 320x240 works well for me.


Anonymous Board owner 10/10/2019 (Thu) 00:13:04 Id: ea6894 [Preview] No.200 del
>>185
I can think of two possibilities:

1) Is your default gui session under options->gui pages 'last session', or something else? The 'last session' is special, and is saved regularly and on exit, but if you have it set to something else, you will need to re-save that session if you want the next boot-load to reflect those changes.

2) If your session is gigantic, it may be that hydrus is having trouble saving it, and because it is doing so on exit, it is unable to report that error properly. Does your session have millions of total thumbnails&urls in it? Or are you on a slightly lower-end machine, with say 4GB ram? If you look at your install_dir/db directory and look through the latest client.log, are there errors in there about your session?


Anonymous 10/10/2019 (Thu) 07:39:26 Id: 1b38dd [Preview] No.206 del
I have a question, it got asked in the 8ch before.

What software do you use for productivity/to-do list?


Anonymous 10/11/2019 (Fri) 11:32:47 Id: d3be39 [Preview] No.210 del
>>200
Well, whatever was wrong seems to have fixed itself. My default gui session was set to 'last session' and there weren't any errors in the client.log file. I have 16GB of RAM, but my web browser was eating up quite a bit of it. Maybe that was the issue? Anyway, thank you for the help!

>>206
Hydrus Dev says he uses ToDoList in the Hydrus documentation. https://hydrusnetwork.github.io/hydrus/help/getting_started_installing.html
The link in the docs goes to a 404 page, but the homepage http://www.abstractspoon.com works fine.


Anonymous Board owner 10/12/2019 (Sat) 21:02:14 Id: ea6894 [Preview] No.211 del
>>206
>>210
Yeah, that's the one. I'll fix the broken link, thanks.

ToDoList is tremendously customisable. I prefer to cut down almost all of the features to just time_due, recurrence, and job name. If you find there are too many fields to customise, don't be afraid to just hide them. I have many daily tasks and also birthday/dentist tracking and so on. Most of them re-occur every day/week/whatever. Even stuff like 'check this web page for new vidya releases this month that might be interesting'. Probably 200 jobs a day, it runs my whole life.


Anonymous Board owner 10/12/2019 (Sat) 21:04:20 Id: ea6894 [Preview] No.212 del
>>210
Hmm, thank you for the note about your session as well. I am not sure if that was it. Maybe memory shortage was stopping the session from packing and saving right, but I would usually expect you to see OS warnings at that point.

Please let me know if you see this again or learn anything new.


Anonymous 10/14/2019 (Mon) 22:42:14 Id: cc4cb8 [Preview] No.218 del
>>210
>>211
Thanks a lot, I'll definitely give it a try.


Anonymous 10/18/2019 (Fri) 05:14:03 Id: a8b074 [Preview] No.224 del
Are there any plans to include Hydrus in some package manager's repo? I'm talking about stuff like NixOS or Guix System, since they seem to be pretty interesting package managers to use, and in the case of Guix, I know you can run such packages on other distros without Guix.


Anonymous 10/19/2019 (Sat) 15:37:24 Id: 129814 [Preview] No.226 del
>>224
I'm not so well informed about packaging on Linux, but I think openSUSE's OBS could be the best method instead of expecting every distro to add to their repos.


Anonymous Board owner 10/19/2019 (Sat) 17:13:07 Id: df4582 [Preview] No.227 del
>>224
I don't know anything about package managers, so I don't maintain anything, but some users do. If you are enthusiastic about this stuff, please feel free to rebundle hydrus however you like for whatever platform.

Examples:

https://chocolatey.org/packages/hydrus-network
https://aur.archlinux.org/packages/hydrus/


Anonymous 10/27/2019 (Sun) 10:25:21 [Preview] No.244 del
Can I regex parse the urls I paste into the url download page as multiple urls with newlines? I'm currently generating the suffixes for the URLs that map to the images on NHentai for a given URL and image count, but when I download them into hydrus I then have to go in and manually enter all the pages. How can I automate this?


Anonymous 10/27/2019 (Sun) 10:47:10 [Preview] No.245 del
how can I open multiple photos in an external program? I like to use ifran view to read manga, but when I go to the next page I either reach the end of the folder or I get a random image...


Anonymous 10/28/2019 (Mon) 23:58:16 Id: 73a873 [Preview] No.247 del
Is there yet a way to default exclude a tag/namespace from searches that don't explicitly include them in the query?
Like if I were to blacklist "pepe" and I had an image tagged with both "pepe" and "meme," querying "meme" would not bring up that image because I explicitly didn't include the blacklisted term "pepe" attributed to it. I'm aware that I could get this functionality with "-pepe" in my search, but it's not exactly the functionality I'm seeking


Anonymous 10/30/2019 (Wed) 07:49:55 [Preview] No.251 del
>>247
You could add a "Show" tag to all whitelisted items. I actually uses a "visibility" namespace and I use the subtags "public","school","work","politics","family","personal", and "private". Unfortunately, this way if you want to show "school" and "work" you still have to '-' the other tags, but to prevent people from knowing I have edge lord "offensive" memes like PePe, NRA, and right wing memes, I just stick them in politics instead of public, and I do a similar thing with the "private" tag which is also a namespace for "sexy", "porn", "ecchi", and "hentai" Maybe that will help you if all you want to do is keep people from seeing your dank memes when working with colleagues when searching for a flowcart, etc? Still not perfect though and the having an | operator for queries would be really nice.

That said, my question is, is there an easy way to download galleries from sites like https://mangakakalot.com/
https://e-hentai.org/
https://nhentai.net/
while preserving all the tags? I actually wrote a macro to automatically use hotkeys to grab URLs to the clipboard and generate the image downloads to attempt(by parsing the URL to the gallery and generating 3 entries per image (the page number followed by the extension[jpg,png,gif]) but this is slow, and wastes a lot of time waiting for non existent images to timeout, as well as having sites like https://e-hentai.org/ using obfuscated URLs that make it impossible to tell how to generate the image from the URL pattern alone(and I'm not a web developer, so I don't know how to loop over the next pages on the site through script to grab the image links...).


Anonymous 10/31/2019 (Thu) 23:33:14 [Preview] No.253 del
>>10
How do I get Hydrus client to play back sound when playing back an mp4 file? According to the little audio symbol in the file there should be sound, and when I open it externally there is sound.


Anonymous 11/01/2019 (Fri) 16:48:57 [Preview] No.255 del
>>10
I have a bunch of manga stored into different folders by story, and along with the pictures each folder has a .json file with all the tags and data like title and page count. How can I import this into hydrus easily? Doing a few folders manually wouldn't be a big deal, but I have probably collected around 1500 or so folders of manga, maybe more. I don't want to have to manually add 1500 files and all the tags when the data is already there in json, and is probably parse-able in some way


Anonymous 11/02/2019 (Sat) 07:19:44 Id: 3ecf94 [Preview] No.256 del
>>253
Hydrus doesn't support sound playback yet

>>255
Hydrus isn't ideal for manga at the moment, you're better off waiting until manga support gets improved


Anonymous 11/02/2019 (Sat) 22:18:01 Id: 4cafea [Preview] No.257 del
Is there a way to use URL import for webms? When I try to use it, it imports as a zip


Anonymous Board owner 11/16/2019 (Sat) 18:54:50 Id: 62de2f [Preview] No.298 del
>>244
I am not totally sure what you want to do here.

There's a paste button beside the text input--if your clipboard has newline-separated urls, it should pull them all in one go. I know some users who have pasted like 10,000 urls all at once like this.

Have you tried this and not had it work? Did it not split by newline? Do you know if your text editor is using NL (newline) or TLCR (newline-carriage return)? I think it is supposed to be able to handle either, but maybe it is misfiring.


Anonymous Board owner 11/16/2019 (Sat) 18:57:28 Id: 62de2f [Preview] No.299 del
>>245
There is no way to do this at the moment, but I expect to add it at some point. Can you say how ifran view takes multiple files as a launch parameter? Is it something like:
program_exe file1.jpg file2.jpg file3.jpg


Different programs have different 'playlist' formats, so I'll have to extend the 'launch external program' code to have some way of describing that.


Anonymous Board owner 11/16/2019 (Sat) 19:00:21 Id: 62de2f [Preview] No.300 del
>>253
>>256
With some luck, I may be able to embed a legit mpv window into the media viewer before the year is done. This is a Qt thing. If we can get this going, there will be native smooth video and audio playback! Keep your fingers crossed.


Anonymous 11/17/2019 (Sun) 16:13:47 Id: 0b82b6 [Preview] No.318 del
(50.88 KB 612x700 Unbenannt.PNG)
(10.94 KB 415x213 Unbenannt2.PNG)
I have been on the hunt for over two hours now, and I feel it is now time to ask a professional: why is my regex, that works the way I want in an online simulator, not working when I actually use it in hydrus?
The attached picture is there to show what I am trying to do:
I want the yellow marked column in a month:xx and the red circled column in a day:yy namespace, but as you can see my shit is being fucked up here.
I basically want to know what two regexes I need to perform to get the month and the day filtered out, so that I can put them in two separate namespaces.
Sorry if this is a stupid question, but as it turn out I am too much of a brainlet for regex.
The second picture shows that I had theoretically figured it out, but to no avail. :(


Anonymous Board owner 11/18/2019 (Mon) 04:53:42 Id: 8b213f [Preview] No.324 del
>>318
Hey, yeah, regex is a pain in the ass at the best of times, and the way I apply some of it isn't totally standard.

Just a quick thing, if it happens to help, hydrus is in python rather than PHP. I'm not sure there will be any big differences, but maybe in some edge cases.

I think in my regex parsing there, I get every matching string, rather than just the first, as perhaps that online one is doing. If you want day and month with those paths, I'd probably go for:

month - (?<=\\\d{4}\-)\d+
day - (?<=\-\d\d\-)\d+


I am making sure the look-behind matches the right number of numbers (\d) in order to capture the second or third group in the yyyy-mm-dd set.


Anonymous 11/19/2019 (Tue) 19:24:44 [Preview] No.338 del
>>298
Ah, was just checking back to look for the answer, and I saw this reply to another post: >>197
I think I want to do this, but with URLs. To be clear, all the images do import properly, but the information about the title and page is lost which I'd hoped to parse into tags using regex from the URL as it is part of the path. Doesn't look like that's an option though? In that case I would like to suggest/request you please support that feature if you can? Would be really handy.


Anonymous 11/20/2019 (Wed) 14:49:33 Id: 4b9118 [Preview] No.344 del
is it just me?
I have been getting a missing cookie "password_hash" when trying to log into danbooru

---------------------------
Information
---------------------------
Network error: 401: The server's error text was too long to display. The first part follows, while a larger chunk has been written to the log.
<!doctype html>
<html>
<head>
<title> Sign in - Danbooru
</title>
<link rel="shortcut icon" href="/favicon.ico" type="image/x-icon">
<link rel="top" title="Danbooru" href="/">
<meta name="csrf-param" content="authenticity_token" />
<meta name="csrf
---------------------------
OK
---------------------------


Anonymous 11/20/2019 (Wed) 22:05:56 Id: a3b755 [Preview] No.346 del
>>344
I've been getting the same problem on danbooru. Not just you.


Anonymous 11/20/2019 (Wed) 23:55:22 [Preview] No.347 del
If I add a "hydrus tag repository" and point it at the PTR, will it sync all my current private tags, like person:<familly member name: last middle first>
? I've spent a lot of time tagging my personal familly photos and vacations with relevent tags, but I would prefer to not leak those to some public repo. I want to tag a bunch of images I grabbed from boorus and chans though with the PTR. I've read the help article, but exactly what happens with my local tags is still very confusing. It also talked about "committing" tags to the repo? What if when I'm setting tags 99% of the time it will be for personal use and I don't want to push them to a repo? It mentions that you can have multiple repos, but I can't find more information on this in the help.


Anonymous 11/21/2019 (Thu) 19:30:07 Id: 8a1e2f [Preview] No.356 del
When tagging I used to be able to enter a namespace and Hydrus would suggest me all tags of that namespace, which was really crucial to my "workflow".
Today I updated after skipping several versions and that isn't the case anymore.

Is there a way to re-enable this? Can't find anything under "tag suggestions" in the options.


Anonymous 11/22/2019 (Fri) 04:45:13 Id: f7613c [Preview] No.358 del
>>344
If you make a copy of the Danbooru login script and remove the two cookies that are checked for successful login then you should be fine for now. Works for me.


Anonymous 11/23/2019 (Sat) 18:59:19 Id: b5c3ad [Preview] No.365 del
Core parts of the 'old place' are (more-or-less) back online, so smaller boards have been given a deadline of 2nd December for the former Board Owners to verify / claim their boards, which will restore them to their pre-shutdown state; have you secured the /hydrus/ board yet?


Anonymous Board owner 11/24/2019 (Sun) 02:02:54 Id: 3e4345 [Preview] No.372 del
>>338
Hmm, yeah, you can parse tag info from URL string easily at the moment. You could probably wangle a custom parser to pull this, but it would be impractical for real use. Is there any chance the title/page info is available in a web page the file is in, like you would see in a gallery site like Hentai Foundry, or do you only have access to a raw jpg file URL? Normally, hydrus likes to grab that data from json or html.

Converting tags to ratings, or ratings to tags, or now tags from URLs, any sort of cross-metadata-type conversion, is not really supported right now. It is complicated and requires a bunch of new UI. I am not opposed to the idea, but I don't think I could fit it into normal weekly work.

I will soonish be adding the ability to easily export and import URLs using neighbouring .txt files. It is possible you could cobble something together here by exporting the files you care about here with their URLs in a txt file, using a script to convert the urls.txt to a neighbouring tags.txt file, and then importing those tag .txts and the files again and pulling the tags. Is that something you might be interested in, or is it a bit too inconvenient?


Anonymous Board owner 11/24/2019 (Sun) 02:04:04 Id: 3e4345 [Preview] No.373 del
>>344
>>358
>>346
Thanks lads. Yeah, I have had a couple of these reports now. I'll roll out an updated login script for everyone for 374. They just changed something, not sure what.


Anonymous Board owner 11/24/2019 (Sun) 02:11:12 Id: 3e4345 [Preview] No.374 del
>>347
No. When you add something like the PTR, it'll make a completely separate service that will appear in the manage tags dialog as a second tab. PTR tags will stay in one list, your 'my tags' will stay in another. Any time you choose to edit tags, or to parse them from something like a booru download, you will be choosing which tag 'domain' they will go to.

Normally when you search on search pages, you will see the tag domain is 'all known tags'. This is essentially a union of all your tag services, which for you at the moment is just 'my tags'. It'll be 'my tags' and the PTR if you add the PTR.

The only way the tags you have added to 'my tags' can be moved to the PTR is if you specifically choose to. You might copy them from the 'my tags' page and paste them to the PTR's, or you might do a larger migration using services->tag migration. If set up to sync with the PTR and touch nothing, no privacy violation will occur. You can't really make a 'mistake' here unless you are specifically reckless, you have to choose everything.

The 'commit' part is a special part of sending tags to the PTR. Rather than upload immediately, when you add them to the manage tags dialog, they are instead in a local-only 'pending' mode. This lets you edit them if you like before you commit by uploading them to the server. When you have pending tags, you will get a new 'pending (count)' menu in your main gui's menubar that lets you send them all up in one go when you are ready. This is the only way that tags can ever be uploaded to the public space--you control it completely.

If you do wish, you do not ever have to pend a tag to the PTR, and can use it exclusively in a read-only way.

If you want to get a feel for multiple tag services, try going into services->manage services and adding a new 'local tag service'. Call it 'clothing' or something and experiment with the new 'clothing' page you have anywhere you edit tags.

Let me know if there is anything still confusing. Thank you for your feedback, I will improve that help.


Anonymous Board owner 11/24/2019 (Sun) 02:13:33 Id: 3e4345 [Preview] No.375 del
>>356
Yeah, I took that out because many users had terrible CPU costs with it, especially when syncing with the PTR. You should be able to force it now by going 'series:*', with the explicit asterisk, let me know if it doesn't work for you.


Anonymous Board owner 11/24/2019 (Sun) 02:20:25 Id: 3e4345 [Preview] No.376 del
>>365
Thanks. Yep, we should be good, assuming it actually comes back ok. I sent an email a few weeks back, along with dozens of other other BOs, when they sent out a tweet asking for emails with shared secrets, but unfortunately those emails it seems were not actioned. Most of us did not get replies, wew. I guess they got eaten when one of the domains got killed.

Today on another bunker, there was a thread to figure out the mess, and /hydrus/ is now on an explicit list for migration. My BO username and pass don't work atm, let's hope they do in December!

I am not super enthused about the new terms of service or the situation in general, but we'll see how it goes. If most of us end up shitposting there again, I'll cheerfully run /hydrus/ from there too, but if we don't, or they can't stay up, I'll throw up a poll to figure out how people feel about different places for a permanent home, here on Endchan or elsewhere.


Anonymous 11/24/2019 (Sun) 06:20:19 [Preview] No.380 del
>>372
The only site I have kind of working, is https://nhentai.net/ I navigate to the last page and hit a key that runs a macro to copy the current URL from the clipboard to parse it and generate the page numbers. I then click on the floppy icon to paste this as a URL download. Unforunately, I then have to manaully enter "Page:1", "Page:2", etc, which is the most annoying part as it must be done for every page. However as the URls I amp parsing are formatted like this "https://i.nhentai.net/galleries/1516309/2.jpg" where the 2 at the end is the page number, I had hoped to regex pars that for each image as it downloads. I have another hotkey macro that lets me select the tag text for all the images in the gallery and parse it into the multi-line tags that hydrus understands. But currently, my macro that tags the page numbers needs to wait for all of the images to download, then I give my macro the number of pages, and it uses keyboard controls to add the page numbers by hitting "home" sleep 1000ms "f3" sleep 1000ms "page:x" sleep 1000ms "enter" sleep 1000ms "right" and then it repeats from the "f3' key. but, this always takes at least 4 seconds to run plus some time wile waiting for the window to activate. So if I have 600 images in a doujin it I have to leave my computer running and not touch it for like 40 minutes minimum where I can't do anything. This is better than manually adding hundreds of tags, but I would prefer to be able to parse at least the page from the urls I paste into the URL downloaded to a tag for just that file. Btw, if anyone is interested in the Auto hot key scripts I wrote to do this, here they are:
https://pastebin.com/DFEyckqu
https://pastebin.com/39xUWT5U
https://pastebin.com/3U4y13cq
https://pastebin.com/AHuTuXyQ
https://pastebin.com/4zNtSYVw
https://pastebin.com/Lrb6E97E


Anonymous 11/24/2019 (Sun) 19:17:50 Id: 0f0b8e [Preview] No.383 del
>>324
Thank you very much. :D


Anonymous 11/24/2019 (Sun) 20:06:07 Id: 8a1e2f [Preview] No.384 del
>>375
That works, thank you.

I would still appreciate it if an option to re-enable this as default behaviour could be included though.


Anonymous 11/25/2019 (Mon) 01:13:12 [Preview] No.385 del
>>10
So, I added the PTR to my hydrus and let it sync overnight, but it looks like non of my files have been tagged with those PTR tags? I thought the Syncing was to download all the tags and the hashes for the files that would be recieving those tags, and then they would be applied? But this doesn't seem to be how it works? So how do I get my files to update their tags from the PTR, and do I need to broadcast the image hash as a fetch to the PTR or something to get the tags back? If possible I would rather not leak hashes either, or be able to just look up a subset of my files and try to fetch tags that way if this is the case? Or is there just some setting I have missed enableing that would let my client used the data I have already spent all night syncing tag my files? Because I assumed that I was essentially syncing a dictionary/map of Dictionary<string, vector<int>> lookupTable. where string was the tag, and int was the hash? Or did I just sync tags/keys when DLing the PTR? And how do I see what tags I actually DLed?


Anonymous 11/27/2019 (Wed) 05:02:17 Id: 94f60c [Preview] No.386 del
Is there a straightforward use case to sync a database between two computers on LAN?

I use both a laptop and a desktop and I store my large media on a NAS, so I'd like to manage the same database no matter which computer I'm using. Network mounting the hydrus directory would probably be too slow (not yet tested), and I'm wary of running a full server per the warning above help/server.html.

Thank you for your work on Hydrus!


Anonymous Board owner 12/07/2019 (Sat) 21:33:01 Id: ee440e [Preview] No.411 del
>>385
Hey, I am sorry for late reply. Did this fix itself? It is possible that if you do a huge sync from nothing, you might not see tags until the next client reboot. You can check this by opening a new search page, changing the tag domain from 'all known tags' to 'public tag repo', and then doing a search for system:number of tags>0. Does anything show up at all?

The PTR in hydrus should just make tags appear, you don't have to do anything but wait for some syncing. If you check the PTR's page under services->review services->remote->tags->ptr, how far is the 'processing' stage done? Is it mostly finished? It should have a mappings count, and as soon as you get, say, 50 million mappings, you should see a bunch appearing on older popular files.

For some situations, tags will not line up with a file. A couple of hydrus users had run their files through image resizers/optimisers years ago, and since the file content here has changed uniquely, those files will not line up with the (hash, tag) pairs the PTR offers. But if you have any decent sample of original files from the big boorus, you should see some tags from the PTR.

Here's a picture of Samus:

https://safebooru.org/index.php?id=1919822&page=post&s=view

If you select and drag and drop that URL onto hydrus, it should import into a new page. The file has tags on the PTR--do you see them?


Anonymous 12/11/2019 (Wed) 01:48:31 Id: 9b4d83 [Preview] No.425 del
I dont know which of these boards is used more so im gonna post here too.

Im getting disk I/O errors because my hdd is presumably dying on me and I neglected to make any backups of my database. I move the entirety of the db over to a new drive successfully except for client.mappings, I get errors whenever I try to move or copy it. Am I fucked?

I have tried cloning it in sqlite, seems like its getting stuck at current_mappings_10.

The program also still works fine, no tags seem affected though I dont really know what client.mappings handles. The only reason I noticed something was wrong was because I suddenly couldn't process repository updates.


Anonymous Board owner 12/14/2019 (Sat) 22:54:48 Id: 06c914 [Preview] No.432 del
>>425
It stores the tags. You are not permanently fucked, but it might take a little work to get back up and running, and if you have a lot of 'my tags' (as opposed to just, say, the public tag repository), it may be trickier.

It may be simpler to do this one-on-one by email (hydrus.admin@gmail.com) or discord DMs (https://discord.gg/3H8UTpb), but we can also do it here.

My best first suggestion is to try to copy as much of client.mappings.db to the nice drive. If your OS is doing some of the move but sperging out and cancelling due to a bad hard drive sector, maybe a program like FreeFileSync will let you copy most of the file and still leave the truncated half-copy in the nice drive after the error. I am not sure if FFS will do that, but I think some copying program or command line call will be able to do the front half. You may have more luck cloning this half-copied db, and if that works you should be able to just boot and reset your PTR processing cache and be back up and running.

If you absolutely cannot copy any of client.mappings.db, we'll have to see if we can possibly create a barebones one and fool the client into thinking that is the correct one.


Anonymous 12/15/2019 (Sun) 05:05:10 Id: 9b4d83 [Preview] No.433 del
>>432
I dont seem to have lost any tags actually. I copied the entire DB other than client.mappings to a new client on the fresh ssd and as far as I can tell every tag is still here, My Tags and the repository.

The only real issue I can identify is that I have about 145,000 tag data for the ptr that I cant upload or forget


Anonymous 12/15/2019 (Sun) 22:00:44 Id: 06bf66 [Preview] No.436 del
Do you guys have a lot of weird, autistic, aesthetic debates with yourselves about how things should be tagged? Like rn I'm debating if oc's in the style of a series should be given the series tag despite not actually depicting any characters from said series.


Anonymous 12/17/2019 (Tue) 03:17:46 Id: cdd882 [Preview] No.438 del
How do you completely uninstall on macOS?


Anonymous 12/28/2019 (Sat) 22:34:32 Id: 4d32a0 [Preview] No.448 del
>>10
Is there any specific place to request/suggest plugins? I would like, for example, to ask for plugins for websites like motherless, BDSMLR, or even for an update on xhamster's download script.


Anonymous 12/28/2019 (Sat) 22:34:52 Id: 4d32a0 [Preview] No.449 del
>>10
Is there any specific place to request/suggest plugins? I would like, for example, to ask for plugins for websites like motherless, BDSMLR, or even for an update on xhamster's download script.


Anonymous Board owner 12/29/2019 (Sun) 18:56:41 Id: 1ef48d [Preview] No.451 del
>>433
Great, I am glad you are basically working again. If you used a 'fresh' client.mappings.db, you will run into trouble sooner or later when you add new tags or process the PTR. The tags you see are probably being nicely loaded from the fast caches on client.caches.db, but the actual tags (and the storage areas where they should be) are missing from the real-deal client.mappings.db. This may need you to hit services->review services and completely reset your PTR's processing, to reformat client.mappings.db and fill it up again properly.

Also, if you have local 'my tags', they may now only exist in the semi-ephemeral client.caches.db. Make sure you have a good backup of this current 'working' db before you do anything crazy, as they may be wiped by a big 'reset/recalc' action.

The 145k tags is for the 'pending (145,000)' menu, right? This again is because it thinks there are pending tags, but the actual ones have disappeared, so it is out of sync with the db. We can reset it, but this may cause an error. Hit help->debug->data actions->clear db service info cache and then restart the client, which will force it to re-count that number from raw db tables. I expect the true number is 0, so the pending menu will disappear on restart.

Let me know how you get on. If you have a lot of local 'my tags', make sure you have a backup of client.caches.db at least--we may be able to do some SQLite magic to re-populate the missing tables in the bare client.mappings.db.


Anonymous Board owner 12/29/2019 (Sun) 19:01:06 Id: 1ef48d [Preview] No.452 del
>>436
Yeah, it is a constant battle. You can go crazy falling down this rabbit hole, and it is easy to become unproductive. My very strong tl;dr recommendation in tagging is to follow two rules:

1) Don't try to be perfect.
2) Only add tags for things you search for.

Tags are for searching, not describing. If you aren't searching for 'series' tags, don't get lost in them.


Anonymous Board owner 12/29/2019 (Sun) 19:05:55 Id: 1ef48d [Preview] No.453 del
>>438
Delete the App from Applications, and delete the folder at /Users/You/Library/Hydrus. That library folder is all my code writes to permanently. If you want to be very thorough, clear out your OS temp folder, as there may be some old temp files hanging around if you have had crashes.

I am not an expert at macOS. Perhaps there are also some system shortcuts or history or something that is added at the OS level. If you know macOS does this and you want to completely wipe all memory of hydrus, you'll have to search how to do this.


Anonymous Board owner 12/29/2019 (Sun) 19:07:50 Id: 1ef48d [Preview] No.454 del
>>448
Best place is probably the Discord, where a bunch of downloader creators hang out. There are several channels just for discussing it.

https://discord.gg/3H8UTpb

I posted about BDSMLR there a little while ago based on a request from another user, but I do not know the current status of any downloader for it.


Anonymous 01/31/2020 (Fri) 03:11:29 Id: 345f8b [Preview] No.522 del
Is there a way to have collections be sorted randomly, but have files in the collections be sorted by page? In other words, is there a way to have a separate sort for files in a collection then you do for the collections themselves and uncollected files?

I want to be able to browse comics I have randomly with the rest of my non-comic files, but still have the pages in those comics be in the correct order when I get to them. I've tried everything I can think of, but I can't find a way to do this.


sorting inside collections? Anonymous 01/31/2020 (Fri) 03:22:27 Id: 345f8b [Preview] No.523 del
Is there a way to have a separate sort for files in a collection then you do for collections themselves and uncollected files?

I want the comics I have to be sorted randomly among my non-comic files, so that I can browse them like I browse my normal files, but still have the pages in the comics appear in the correct order, so that the comics make sense when I get to them. I've tried every combination of sorting and collecting that I can think of, but I can't find a way to do this.


Anonymous Board owner 02/17/2020 (Mon) 23:04:40 Id: e8c215 [Preview] No.525 del
Hey, just catching up after Endchan downtime.

>>522
There is no way to do this at the moment. I am ultimately unhappy with how collections work in many ways and want to clean up and rewrite handling of multi-page content, both with explicit cbr-style file support and a virtualised 'single object, multi-file comic' cbr format to handle paged comics as single thumbs rather than auto-sorted files.

I apologise for the inconvenience in the meantime. The next time I am in that code, I will see if adding a simple double-sort system as in >>523 is easy to add, but if not, this may have to wait for the bigger rewrite.


Anonymous 02/20/2020 (Thu) 17:13:11 Id: 242d48 [Preview] No.532 del
>>525
>to handle paged comics as single thumbs rather than auto-sorted files
when/if you do this, you should give priority for files tagged as "doujinshi cover", "cover", "medium:cover", "cover page", "doujin cover", etc, as the thumbnail.


Anonymous 02/23/2020 (Sun) 15:34:04 Id: 345f8b [Preview] No.535 del
>>525
It's fine. In the mean time, I'll just hold off on importing my comics and series until I can. If I imported them now, the individual files would just get in the way, since "random" is the primary way I browse.

>>532
I agree, but if that were to be done, then instead of it being hard-coded, there should be an option to set which tags get thumbnail priority, and maybe just have the tags you listed be the defaults. This ensures flexibility in the case that there are other groups of files where one kind makes obvious sense as the thumbnail.


Anonymous 02/23/2020 (Sun) 19:45:17 Id: 345f8b [Preview] No.536 del
When adding tags manually, there's an option that auto-replaces old siblings with new siblings when you enter the old sibling, but it can be disabled, and I do. How do you disable that for tags that are automatically added through the gallery downloader. I checked the tag import options, but I didn't see anything about siblings there.

I'm importing files from Sankaku Channel, and I want the tags that are imported into Hydrus to be the tags that are there exactly as they are on Sankaku Channel. For example, I have 1boy as the old sibling of "1 male" in "my tags" but when I import a file from Sankaku Channel, "1boy" is just replaced by "1 male" entirely. I don't want that to happen. I want the underlying tag that's imported to still be 1boy, even though it will displayed as "1 male" since it's the old sibling.


Anonymous Board owner 02/24/2020 (Mon) 01:39:46 Id: 3d2bd5 [Preview] No.540 del
>>532
>>535
Thanks, I agree. I will have to be careful about how we do this stuff. As we have found with collections, page: tags are not comprehensive.

For a long time, I have thought that virtual comic support in the client, and related tasks like 'please tag these twenty files with page:x, starting with 17 and adding 1 each time' really needs thumbnail re-ordering with mouse drag and drop.

Once we have that, it will be much easier to solve niggling file sort problems where a comic has 'main page' vs 'front page' vs 'copyright page' vs 'last page' in just a couple of mouse drags.


Anonymous Board owner 02/24/2020 (Mon) 01:41:04 Id: 3d2bd5 [Preview] No.541 del
>>536
I regret that there is not yet. This work got delayed in Q4 last year because of Qt, but I hope to be back to it in the coming months. I want better sibling filters and sibling controls for all services so you have power over this.


Anonymous 02/24/2020 (Mon) 20:32:30 Id: 294240 [Preview] No.544 del
Hi, I'm having a problem: apparently database issues, so I've been following all the steps in the broken database txt, however... the master.db file (6.something gb) has been doing the "PRAGMA integrity_check;" for over 24 hours now. The drive is a normal 3.5 inch magnetic usb3 one, it's not abnormally slow, and the cpu is a brand new Ryzen 5 3600. It's not magically stuck because the task manager shows the drive being with an active time of 99%, and there's nothing else using it, not right now.

plz halp.


Anonymous 02/25/2020 (Tue) 19:32:34 Id: 294240 [Preview] No.545 del
Endless check has finished, took well over 24 hours but less than 48 (not sure how much exactly).

No errors. Now on to the next one and let's hope for a faster job lol


Anonymous 02/28/2020 (Fri) 16:51:36 Id: 345f8b [Preview] No.551 del
When a sibling relationship is created or petitioned to the PTR, the changes show up for you, even if it gets rejected by janitors to be shown to everyone. Is there a way to see which ones are actually just showing up for you (because that matters sometimes), and a way to be able to remove them, or makes new ones without having to suggest that they go to everyone else too?


Anonymous Board owner 03/01/2020 (Sun) 22:22:09 Id: 2a3204 [Preview] No.554 del
>>544
>>545
This sounds a very long time. Normally SQLite is ok at this stuff, sometimes as slow as 250KB/s, but rarely that bad. It could be due to abnormally high drive latency (which could be due to it being mechanical and over a usb, and/or it being very very fragged), or your machine being very tight on free memory (so your OS's disk cache cannot populate, or perhaps the drive itself is having trouble.

Were you able to run a program like CrystalDiskInfo? Is the drive itself healthy?

Run time normally scales with file size, so if your client.mappings.db is 20-odd gigs, it could be several days to run. This sounds ridiculous to me, so I would suggest you cancel the job and figure out what is wrong with the connection to the drive first. It would likely be easier to copy the file to your system drive with sqlite3.exe, especially if it is an ssd, and run the integrity check there.

Please let me know how you get on.


Anonymous Board owner 03/01/2020 (Sun) 22:26:56 Id: 2a3204 [Preview] No.555 del
>>551
Unfortunately, there is not. In its development, the biggest problems with the tag repo code have been in dealing with its growth, and admin and user controls took a back seat behind the back seat.

This is now a bigger deal, and the PTR janitorial team have given me some suggestions as well. I do not know if I can make it as nice as you want, but when I next do a big iteration on tag repo content updates, I expect to add better client knowledge of clientside changes, the ability to make changes locally and not upload them, and likely some sort of anon 'petitioner id' so you can see your petitions' results and exchange anon messages with janitors.



Top | Return | Catalog | Post a reply