I wrote this in the q&a thread but it's basically the reason why i can't get myself to fully entrust hydrus for all my stuff, since for now i'm keeping the db small to test things and i'm already having problems with how to use it, i'd have no idea what would happen if i had already put everything in there I can't stand the way hydrus handles collections, if in explorer i had a general folder for creators or series, hydrus insists that i get a collection for every pairing of the namespace i chose to collect with In explorer i had to choose in what folder to put an image, giving up seeing it if i decided to browse another folder related to it, unless i abused shortcuts, which are maybe ok for a thumbnail view in explorer, but get skipped if you want to browse with your imageviewer or use something like irfanview thumbnails to load them But it was kinda easier to just decide to open my image folder and browse a bit, even if the folders didn't have files related to them because they were elsewhere Also subfolders still helped keeping things in order, in a creator folder i could have subfolders or shortcuts for his other works, keeping different types of images separated, but in hydrus if i search for that creator tag i'm just gonna get everything at once and i'll have to sort and collect every time i do a search Another minor gripe i have would be the inability to group namespaces in the left window (which would be a partial solution to my collections problem, as i could live with just browsing and picking something with the text list) These are the things that make me think about giving up on hydrus and just go back to folders, but i'm still trying because i really want to use hydrus tagging capabilities

thumbnail pages/infinite scrolling I've brought this up many times, but it's still much faster for me to go on a booru and type in a query and get results back instantly. system:limit doesnt really help.

>>9013 It sounds like you just need a better tagging solution, but maybe I'm misreading. >but in hydrus if i search for that creator tag i'm just gonna get everything at once and i'll have to sort and collect every time i do a search Why not search "artist:[name]" "title:[name]" or "collection:[name]" >Another minor gripe i have would be the inability to group namespaces in the left window Not sure I follow. Being able to search a page by two namespaces works just fine. Does the a-z grouped by namespace not do what you want?

>>9014 not sure what you mean, I have 139 tabs of thumbnails open right now, and something around 700k images open with 1 tab being 90k images big I honestly can't imagine a query that you would want that would bring up more possibilities then this. I can tell you this, at 40K+ images is where the program starts to slow down and lock up, 200k is where it starts to get annoying to use. really need to get around to processing the imports. as for instant results, it depends on how big the tag pool you have is (I have the db and thumbs on an 960 nvme) On a fresh hydrus, I can search something like futanari (a tag pool of 84k images or there around) and limit 1000 faster then a booru would load 20 results. granted, if you are going off a hdd, sure, its going to take a while, especially compared to a booru that likely has the entirety of its db and thumbnails in ram, most use images/new ones on an ssd, and the rest on several hdds, not sure if there is a whole hell of alot you could do outside of getting faster storage for at least the database portion of hydrus. >>9013 personally for me, I downloaded so many artists and so many images I have over 2 million, for me I gave up manually sorting images a long time ago as it was a losing game where I had to constantly decide what was more important because I sure as hell wasn't having images in multiple spots. what your issue sounds like is you download an artist, that's good, but then this artist also makes a comic, so when you search the artist you get all their singles and then their comic. you should be able to have an artist you should also be able to sort by another name, lets say 'set' this would be a catch all for all images that are in a set, be it comic or someone decides to release an image compilation of things related. from there you also name the set something specific. this way you can search artist name, get all their shit you can search artist name and sets and only get the image sets they did from there you can select one of the images and if you did the tags right, load up everything from that specific set or you can search artist name, sets then set sets to exclude, and you would only see the single images the artist put out. It sounds like the issue you have is more in how you tag/import the images unless im completely missing something.

>>9017 Yeah, i surely need to find a tagging solution that fits my needs in how i want to browse the db, but i have to make sure it's not overly complicated that in the future i can't be arsed to follow it anymore and everything gets messy again For the creator tag search i meant it in a way that getting all results at once would mix images of "different types", while in explorer i'd probably have a subfolder for his pixiv, his official works, artbooks, hmanga or cgs he may have made etc They'd be sorted the same way every time i opened the folder, while in hydrus i'd have to make tags and everytime i do that search i'd have to sort and collect Another example would be searching for a series and getting fanart, official art, screncaps from the anime, pages of the manga etc all at once, and everytime i'd have to sort them correctly, gets tedious when doing multiple searches in a row As for grouping namespace i meant for the left window where tags are to work as a cascade menu, instead of an endless list of all the tags In that way, i could for example click the series namespace and unveil the list, but as of now i have instead to scroll down every time to reach them This gets extremely tedious because of another reason, the fact that whenever you do a search (as in insert\exclude a new tag), the tag window refreshes and starts at the top again, meaning that, for example, browsing series tags in sequence would mean to scroll down each time to search the next one, while in explorer it was just a matter of enter folder, exit folder, press down, enter next folder etc The "solution" is, instead of clicking the tag to search it in the same page, to open it in a new selection, so that if you wanted you could tab back to the previous page and the tag list wouldn't have moved While this would be fine (equivalent of ctrl+enter in explorer) it brings other problems, as i mentioned above opening a new selection means resetting the sort\collect to the default, and you'll also be unable to add further tags to filter that selection >>9020 You're correct when saying that when i want to search for a specific creator, if i tagged things correctly to keep type of images separated i can sort and collect by them just fine (altho as i wrote above it's annoying to sort\collect every search while in explorer the subfolders would be already there) But my biggest problem lies when i don't already know what i want to search, for example when i'm just thinking "lets browse my creator, series, artbook, whatever folder" In explorer, easy enough, i'd open that general folder and get a list of all the creators, series, etc, but of course there were limitations, for example if i wanted to exclude from the view creators that did only fanart or exclude series folders that had only anime screencaps or manga pages i couldn't If i wanted that kind of thing it'd mean to make a collection folder elsewhere for every search i liked and then shortcut everything in there: one for anime screencaps, one for dakimakura pics, one for official art, etc, but in hydrus it's way easier as it's just a matter of them having the tag, which is probably already present since i download tagged pics Now in hydrus, if i think "lets browse my creator, series, artbook, whatever folder" the situation brings different results, let's say i just search for series:anything, and then sort and collect by series to mimic the opening of a general folder in explorer What this does, instead of giving me a just ONE collection for every series namespace, is to give me surplus collections for every image that has 2 or more series tags, the equivalent of creating extra folders for just a few pictures, cluttering what would be a clean view If there's an azur lane collection, next to it there's gonna be an azur lane&fgo collection, when i would instead like for that picture to be present in both azur lane and fgo single collections For artbooks it's a bit uglier because i'd expect those to be a bit cleaner, but an image that for example belongs to 3 artbooks is gonna make a collection by itself

>>9023 Personally this is how I currently do it I do all my sorting in ratings my top level is a filter pony funny art favorite real/drawn safe/suggestive/explicit tag further each one acts a filter of sorts and is the most basic way I can make a quick sort safe/suggestive/explicit acts as a hard filter here, and it will be what I search for when I do a stage 2 filter this keeps what I need to do very simple, very short, and gets images sorted by the widest strokes. from here it allows me to very easily go into more and more detail without going autistic with each image in tags, I would suggest something along these lines acting as a hard filter. thing is, once I rate everything, I can VERY easily prime them to tag everything that comes up in a search of the rating. when I came up with this system I found tagging cumbersome but rating to be far less so. now you keep going back to explorer is easier, and yea, I can see that for some cases, personally I don't want my db like you apparently want yours. If I think about it, try making a rating called 'keystone' and have 2 of them, 1 called creator, 1 called series, and depending on how low you want to go, character. only select 1 image for each of these, one that you feel represents what the creator character or series is, that way you have 3 things to search, and from there, a tag to search which will bring up everything from either series, character, or creator. your issue honestly sounds more like one that comes from a place of you can't think of how to tag/rate your images in a way that satisfies you. personally I give 0 fucks about this, my favorite images will be favorited, and ill be able to pull random sets of favorite + explicit images when I want to get the job done, ill be able to pull art up and look at some things I want to be able to create, ill be able to real + suggestive or real + art and have a set of images to draw from, when I feel like I want to look at a specific series again, i'll just search the series and pull up whatever comes up. but honestly I have a db that's over 2 million, there was no way to manage that shit with folders from the beginning, but I know where you are coming from, I had a bunch of artist sets that I didn't want to mix and match, till I realised I had so fucking much that I would just pick a few go to's and never look at ones that may or maynot be ones I liked more but never got around to parsing.

I'd like server to be fixed for OS X (thumbnail generation is broken).

-No audio for videos. It's 2018, how is this still a thing? -Sub-par tools for image comparison. Keeping the zoom level and position when switching between images is almost necessary for a program like this. -In addition, the viewer shits itself when zooming in. -Very lackluster metadata support. Having a mediainfo summary when viewing a video would be great. -Terrible Linux stability. -No import-time duplicate checking. Checking for duplicates when importing new files instead of checking in a massive batch would make the workflow much more streamlined. -It's not possible to show subgroups of multiple different groups. When browsing all your artists, having something like a "show max 10 images of each namespaced tag (ie. artist)" would be great. -When drag-and-drop exporting, you can't dynamically manipulate the filenames. Automatically naming the file with a tag namespaced as "filename:<something>" (or other of your choosing) would solve this. -Booru tag dumps seem to be dead. The tags on my files are constantly becoming more and more outdated, and rechecking tags for 100 000+ files by manually checking every page would create massive strain on both the booru and my internet connection.

>What stops you from using hydrus more? the fact that days only have 24 hours, and years only have 365 days

Home internet speed.

The feeling of being locked in to using one application forever. I like tags, but I'm not ready to ditch the file hierarchy. If I have a shitload of files on an external drive I want to be able to access them in some reasonable fashion in generic file managers too.

>>9061 The further you go with images, the less and less they are manageable through folders. It took I think a year after I first started fucking around with hydrus to say 'fuck it' and go all in. I had folders that were several hundred thousand images big, and trying to any form of sorting was impossible due to every program working against me and the only one that didn't, crashed constantly. for me, if hydrus dies, and gets crashy, I will be able to export the images with some semblance of sorting as in drawn/real safe/suggestive/explicit and that will likely get expanded on with more hard checks where images are one of the ratings. I would think of it that way rather then the folders that hydrus sorts images into, before the program full dies, you will have plenty of warning for an exodus, and hydrus facilitates it fairly well.

>>9061 >>9083 Even if the project stop being updated and most downloader break, the archiving side won't stop working. It's like complaining you've been using honeyview for a decade to look at your illustrations.

>>9085 one of the things I was using before hydrus was an image sorting program that moved files to folders based on single key presses, it worked out pretty good, real porn/real safe/drawn porn/drawn safe however the program had massive issues with crashing due to how long ago it was last updated and some of its dependancies fucked itself. in the end, this was how I was told about hydrus, but the whole hydrus decides how everything is stored make me very hesitant, but after I came to terms just how fucking unmanageable everything was, I did a small 5k image import to test it out, then it got more and more, seeing how it would handle shit, till I just said fuck it, and imported everything, which took a solid month of importing because I did it in a unknown to me at the time, retarded way. point being, if hydrus ever goes tits up, have an exit plan but hope for the best.

The system:limit feature only works on everything. which is a big pain in the ass not picking up the ratings if i use 'sort by'. i imagine this is going to be a real problem when i get tons and tons of pictures.

The UI documentation is not "complete" so that might lead to issues.

>>9093 system limit 600 futanari only brings up 600 images with the tag futanari system limit 600 system rating for favorite 1/1 brings up 600 of my images I labeled as favorites. random 600 then sort by favorite the favorited ones show up at the bottom, unrated at the top able to give an example of what you are talking about as I may be misunderstanding what you want to do.

I get random system hangs after around half an hour of it working on subscriptions. I have monitored system usage and nothing strikes out when it happens. no noteworthy hdd activity, no noteworthy ram usage, no noteworthy cpu usage on any thread and no unusal temps. It just freezes everything apart from the mouse and it usually takes around two minutes until my desktop reacts again to my input. Not even swap gets used so I have no fucking clue why it does that. I guess hp laptopism strikes again.

>>9111 Keep in mind, unless something changed, the program starts to lag once you pass 40k images or so, and even if you remove the tab, the program still stays that way. I have around 600k open right now, the program is more or less useable at least as far as downloadsers go, and I can parse something around 400~ images inbetween hangs if I can parse from thumbnails, and around 20 if I need a better look at given images, considering shit hangs for 5-15 seconds, not exactly the funnest time.

>>9103 Well an example is like what you already gave. I try system:limit 100-1000 or whatever number and try to sort by favorites but it only picks up a few I actually rated as favorite, then drops off into random pictures. If I use system:everything then all my favorites show up. Same behavior no matter the filtering, system:limit just doesnt work. Using Hydrus 307

>>9143 Updated to 310 but same behavior. I should note that it's ratings that are broken. If I use system:limit and punch in a tag then it brings them all up fine. but without ratings there's no real order to it. I could tag with 'favorite' but thats just redundant to what ive already been tagging with favorite ratings.

>>9143 >>9114 >>9111 I don't know if this is "working as intended' but all stuttering I had cleared itself when I would routinely go to "undo -> closed pages -> clear all" I would start to notice this behavior after multiple queries of multiple 10k+ files, so it became part of my routine.

>>9144 Honestly it sounds like you want limit images and system favorite but you are only ever doing one at a time. you know you can put as many filters as you want into it? at least as far as I have looked as many as you want.

The only thing that I want from Hydrus right now is subscriptions to Reddit, since I get a lot of images from specific subs.

tl;dr but I agree that duplicate sorting is incredibly tedious right now. I realize the problem is that an image could be detected as a potential duplicate for two different images not within the similarity threshold to each other, but the "show some pairs" function works fine so it would be really nice if the dupe sorter could bring up groups like that to scroll through, or at least work through one group like that in pairs rather than swapping between different incomplete similarity judgments like it does now. I'm usually using a 1366x768 screen but I still would like the option to scroll through all duplicates that would otherwise appear using the "show some" button, since I can't easily do the same dupe operations from just the normal window without launching the dupe manager popup. >>9216 I'm mostly waiting on the new download manager where we can make custom user scripts for various sites, because not only could I then use it for subreddits and "scrape through" to linked imgur stuff, I could scrape links and comments and hopefully also scrape through to video posts people make from other sites. Actually, maybe what Hydrus really could use is a link scraper so if there's unscrapeable content it dumps it into a txt file that it keeps in its db so I can go back later and grab the things it couldn't get….and/or run another custom userscript download on a list of URLs from a given site.

>>9020 >not sure what you mean, I have 139 tabs of thumbnails open right now, and something around 700k images open with 1 tab being 90k images big How much CPU and RAM is Hydrus using when you do that?

lack of a vi(m)-mode

the PTR is a major double edged sword. adding specific tags are reletively painless but if its an image hydrus didnt pick up in autotagging and I have to tag the character myself? fucking forget it. the program just locks up entirely trying to load all of the possibilities for character: before letting me finish what I was typing

>>9219 I thought that too at first, however the only time I would ever find it useful is when an asshole decides to bomb a thread with the same image, just 1 pixel is changed I suggest in cases like that, you set up a rating, something along the lines of dup process mass, or something, search the image, then right click and find all its dups. as for the way it handles dups now, Its honestly far more useful then you think, especially in the case of near identical images, its not easy to remember 3 images, then parse through the 3 to find the one that is the best, is it the 10mb one that is largest, is it the 3mb one that is half its size, or is it the 500kb image that is the same as the 3. in my case when this came up in show me some pairs, the 10mb one was a blown up and blurred version of the 500kb image, with the 3mb one giveing the most clerity. granted on that small of a screen dup detection would be a pain in the ass no matter what till some changes are made in how its handled. as for the scraper, look at the log and see if that's what you want, I know it catches links in there that it didn't scrape from 4chan >>9264 I have gone through a bit, its currently 78 tabs and 630k images, but it uses 12-15gb when I do this, cpu is anywhere from 0-6% im on a ryzen 1700 though so 6% I think is just maxing 1 thread.

I M A G I N E: >tab containing a navpanel that loads in the json of a 4chan/8chan/any chan software that uses json >can load in ads too as normal to keep hiroshimoot and others happy >point is that you skip most of the net code/bloat/fud/timewaste of making a real browser by strictly pulling json and ad elements >barebones default look and you can build your own client-side css theme >any really nice themes that get posted here or made by the homos in Discord get added to a simple repo you can pull from inside of the client, maybe a few get added to default install if they're lightweight >muh dark themes >simple posting form in bottom panel, always visible, pulls captcha element from 4chan and 8ch, you can scroll through in your main window and never have to click to bring up a post form >post form is flat one line by default with the name/options fields inline, you can click-drag the top of the panel up with mouse or change a default setting in options to make visible text area bigger, but by default it covers very little screen real estate >either just ask devs/mods of imageboards to accept your custom user agent because hydrus is good, or spoof a new-install chrome useragent or something, can't be too hard, Opera used to offer full-featured custom user spoofing before they had to ditch it for PR >both view panel and post panel are bordered on the right by the 4chan X style thread watcher, not like a hydrus thread watcher, this one is for "starred" or favorited threads and just watches them in the background, lets you know when there are updates (posts rather than files) >watched threads also have a dropdown at the end opposite the close button, options like open a hydrus thread watcher in new tab, new tab at top level of tabs >you can set presets in options to always open new hydrus thread watchers in given tab-subtab structured areas based on site and board, so I no longer have to navigate to wait>4chan>/wg/ and make a subtab there >you can set specific subtab-located "topic watchers" so instead of making a new thread in wait>4chan>/wg/ for each new cat thread and reopening the tag import options and retyping (or choosing from favorites) download:4chan, board:wg [with media:wallpaper subtag], animal:cat, etc, and reticking that import filenames box each time - I can instead just go to the catalog view of my hydrus not-browser (I think 4chan may provide a catalog json directly, not sure though), search for the new cat thread, star/favorite it right from the catalog, and send it to my pre-defined "/wg/ cat watcher" in wait>4chan>/wg/ subtab, or send it directly from the thread icon on the catalog, and it will insert itself into the list and download all of the files and apply those same tags. >just like the simple downloader works where you can paste in as many URLs as you want, but for thread watcher. Then just have a popup where you can set a policy to auto-delete the threads on 404 or archive, or keep them around but stop checking them so they can be cleaned out manually. >going a bit further, an internal pseudo-RSS/filtering function which scrapes loaded catalogs, OPs, or threads for specific text or even md5s to grab for auto-add to thread queue on right panel (4chan X style thread watcher) and/or send directly into "topic watchers" (maybe archthreads would be a cooler and more fitting name for them) >right watcher panel of fake browser is always the same however many tabs of fake browser you open >recheck rate of all of these various things like the thread watchers and the autofilters can be configured to avoid getting b& and to reduce load on hydrus >fake browser tabs give (1) and (!) functionality for new posts in their specific thread and also update more often than right-panel watcher threads do, which is also configurable It might be beyond the capabilities of python itself, or maybe need certain features and chunks written in C++ with python wrappers, but man. I might only open an actual browser once every few months if it got to this point. And then if it added IRC and Youtube tabs… In fact, hydrus_dev, if you're reading this…hydrOS when? remember no bully for my flawed tagging and tab structure system, I'm learning in baby steps

>>9360 That would be cool, but I don't need emacs 2: booru edition. I'd rather get an extension framework so that the community could add features, since hydrus_dev likes working solo. Dev could keep working on the main part of the program, like the new downloader system, and people could make stuff to extend it if they want, like browsing 4chan. Hydrus shouldn't have the extra bloat of an entire browser by default, especially since half the point of it is that it doesn't connect anywhere without you doing it manually. Also >youtube sound when?>>9360

>>9010 I think I would use it more if it were more usable from the command line. By this I mean basic stuff like adding a thread to be watched or a new search query without needing to be at the machine (i.e. when ssh'd in). To make that possible I suppose a Hydrus daemon would need to exist, which would be very nice. >don't need to keep a window open all the time >don't even need to have a graphical session running >download subscriptions/watched threads in the background >starting Hydrus regularly just pops open a GUI for the daemon

>>9364 If you don't want eMacs: usable GUI edition you must be impaired or something

I'm afraid that it'll enhance my already excessive hoarding addiction even more.

>>9470 vim 2 when

I am seriously so frustrated with Hydrus- I basically hit a wall of not wanting to deal with it anymore, and I sincerely felt like just abandoning ship and going back to Windows folders. My use of the program is very superficial- I still download images manually, even- but here is my attempt at a comprehensive list of my problems with Hydrus, in no particular order: 1.Lack of tag trees Consider my situation that made me basically give up on Hydrus just now, of my arbitrary tag of "string clothing:apron (tied behind neck, not shouldering burden)". This is a really layered tag to communicate to myself: [string clothing] [string clothing:tied behind neck {note:distinct from the parallel straps of a backpack or conventional bra}] [string clothing:tied behind neck; {sub-tag(s): "apron"}] [apron] [apron {note:arbitrary tag meaning clothing tied from behind rather than using straps* or sleeves**}] *insert arbitrary legend here **insert arbitrary legend here But as you can see, Hydrus offers none of this layering, so instead the tag is a disaster of purposely truncating my own terms to avoid it just being an infinite amount of text scrolling horizontally in a single tag. As well, namespaces mean little to me when following the namespace is countless other tags that should be sub-tags, so I have to arbitrarily choose one tag of the multiple I'm forcing into a single tag to set as the namespace. The infrastructure isn't there for tags and their one-layer parent tags to communicate to me what means what when I'm tagging anything. Just now I was about to create a parent tag to my disaster tag dense with information I shared above, but I'd started wording it like "both:apron…" and then I just gave up. 2.Shitty search In another program, "Search Everything" by voidtools (they often drop the "search" part of the name) I can search "dic oc mon" and it will show me my "dicks made of chocolate ejaculating money" folder. In Hydrus, I search "dicks choco" and it returns no results, because I have to type every character correct in sequence for it to yield results. This creates this complete waste of time game of my having to distinguish which sequences of characters are most distinct, easiest to type, shorted, and still communicate the tag to me so that I can create and search for tags without infinite results flooding the results when searching. This was really a bad experience for me, considering I was using Search Everything to search my folders before converting to Hydrus. I was told there is a wildcard search but I don't understand how to use it, and if I have to type a star (shift+8) to represent truncated terms that would be hell. 3. Import not respecting import order (and the import window lying about it) HOLY SHIT this fucked me so hard for WEEKS after I got into the program. This is a really simple issue to explain. You import files, the import window respects the order you imported them in, which is nice… however, the Hydrus client itself will almost certainly be discrepant from your import order you're fucking seeing in the fucking import tab. I have imported dozens upon dozens of artist galleries, hentai, and didn't tag them right away, so now the order is fucked and can't be trusted. If you close the import tab without collecting the images you will have to import them again to with 100% certainty view them in the correct order, without having to go externally to contest your Hydrus against the original source to confirm the order. What a crock of shit. 4. No auto-paging infrastructure I should be able to import a 200-page doujin and hit a toggle to have Hydrus automatically tag everything "page:1" - "page:200". Hydrus does not do this, however. Wastes my life. Let alone that a few versions ago (Windows user here) Hydrus started briefly showing a horizontal scroll bar at the bottom of the "manage tags" window when tagging from the "Hydrus client media viewer". This brief redundancy eats my inputs for over a second, since my computer is rather poor. So to evade this I have to tag from the Hydrus client thumbnail view, which requires two "enter" pressed to confirm tags. I realize this redundant appearing-disappearing scroll bar wasn't intended but it makes an awful lack of feature even worse.

>>9481 5. No radical "creator upload order" order option that undermines tags and namespaces This is a big one for me, and is related to my third complaint. I remember manually (my own fault) downloading semahiro's entire pixiv gallery, and saving it from newest to oldest. This fucked me up, because whenever new semahiro was shared, the "sort by time imported" dropdown basically became worthless since it was completely backwards until the point I started adding new ones. I was torn between what to do here for a while, and I ended up manually (my own fault…) re-downloading semahiro's entire pixiv gallery again, but uploading it in reverse order. And I was paranoid about accidentally fucking it up, so I double-checked it, which was what enlightened me of my third complaint. Had the option exist to just fundamentally tell Hydrus "this is the creator upload order" without just creating a retarded unique namespace and manually tagging every image with a "page" namespace one by one, this whole problem wouldn't have been an issue. And I don't think Hydrus can even deal with a broad namespace to arbitrarily sort like this anyway, since it isn't layered, so it should just break any sub collections within the artist namespace… Let alone that hypothetically if I miss an artist upload, or layer decide to import one I'd passed over before, what am I gonna do if everything is tagged [namespace to set upload order] and [page:x]? I'll just insert it beween say "page:2" and "page:3" as "page:2.1"? This is a failure. 6. Tag siblings append the amended tag name to the end of the old tag when searched, basically making it invisible since the "manage tags" window doesn't have the horizontal space to show it, and the new tag has a lower number of images tagged under it, making the results redundant (the same tag appearing twice under two different names), and the amended tag being lower despite it being your current superior choice for searching this tag This is kind of a clusterfuck to explain. For a real example from me, see the following tags: clothing:paizuri beg (underboob)" (73) (will display as paizuri beg:underboob) paizuri beg:underboob (2) And also it's a tagging disaster anyway from lack of tag trees, as always. I hope that's everything that bothers me with this program, but I guess I can throw out there some minimal complaints that don't break inhibit my usage to my knowledge: -the "darkmode" theme is ugly and clashes too hard with the default window appearance otherwise in my Windows 7. I would prefer instead if I could just set an arbitrary (perhaps partially transparent) background image instead, while still having the freedom to simply choose a color. Also a "restore default" or however button specific to this options page would make playing with themes basically a free move. -I don't remember how I made a custom namespace visible above image thumbnails similar to the default "title" namespace, but it was a disaster, place make this less autistic. Out of the box Hydrus' only defaults here are "series/creator/title/volume/chapter", and none of these sound appropriate for arbitrarily collecting images. I created an arbitrary "pool" namespace in the place of the above defaults, but out of the box I used "title". I was tired of my informal collections obfuscating official titles of for example hentai, so I named arbitrary collections "title:(set)". That just involved so many characters, I eventually manually converted them all to a "pool" namespace. This was not a fun experience for me and I delayed it because it didn't seem obvious to do (and it wasn't). -please add a way to preserve filenames by default; I have tons of videos from porn sites downloaded using my tablet (rather than youtube-dl), and tons of tumblr images, images from 4chan (unix timestamp filename) etc. These all communicate to me when a file is of worse integrity than the original, or in the case of a unix timestamp filename, if I can feed it to an archive to view all reposts of it. It just sucks that Hydrus eats filenames. Currently I've imported every image/ video from my old folders into Hydrus, without tags, but I kept the originals intact so I could confirm the filenames and if the info is worth preserving. -please remove the forced lowercase tags: this kills imgur filenames, makes some artist names illegible, makes hentai/ manga titles look like illiterate shit, I don't know why this is even a "feature". I ruins my day every time.

>>9481 >>9482 Now this last request might but impossible (or against the philosophy of what Hydrus opts to be), but I am obsessed with archiving, and often save webpages for offline viewing, even using Firefox's shift+f2 > "screenshot –fullpage" (or right click screenshot if the output is a png) command to save threads. I also save a text document of the webpage URL, and a screenshot of my entire desktop at the time of viewing, with my mouse hovered over the clock at the bottom right for convenient viewing of the timestamp. Anyway, Hydrus doesn't offer the infrastructure to add this to for example when I download a ddl link to some content. I wish it did, but wasn't visible in nomal viewing, since I will virtually never want to see a screenshot of webpages and shit alongside my porn; I wish it was a sub sort of thing. It's a similar request to my wishing Hydrus offered tag trees. For now I end up for example recognizing imgur filenames on 4chan, importing the image to my Hydrus, then creating a unique tag to link to the 4chan post, tagging the image something like "imgur:hyrus cant preserve filenames"… but I tag it the forced-lowercase imgur filename anyway, in case somehow the 4chan post get lost to time. This isn't immediately related to my saving offline webpages n shit, but I just realized I could do the same by creating a unique tag linking to the Windows explorer path to the archive folder. But obviously if I change the path at all down the road, every single such path tag would break. I wish Hydrus streamlined this somehow, by offering more than one layer of tagging. But this is a bit much compared to something like tag trees. I didn't read the thread at all, so many of these might have been mentioned already and articulated way less autistically but here is my input anyway, for what it's worth. My use of Hydrus is really superficial, so I'm mostly in disbelief I have so much seemingly (in my eyes) simply everyday problems stopping me from thinking it "just werks". Also the latest release of Hydrus makes it fail to respect import order on the level of you dragging images into the client- it instead forces a name sort. So after downloading a gallery, I now have to use Irfanview to batch rename the images so the import window stacks them in my arbtrary order (then fucks it up anyway when feeding them to the client from there…). Also I would like to be able to arbitrarily inject images between other images in the import window (also would like this for the client itself, as I proposed above with the "No radical "creator upload order" order option" thing). Also when I first installed Hydrus, it wasn't obvious to me that toggling "exclude previously deleted files" from the import window wasn't enough to set that toggle permanently… I ended up realizing an unfortunate amount of time down the line that on import it was still skipping previously deleted files, and I had to go to the hard options version of that dialog window to set the shit permanently. Maybe this behavior was changed by now, since this was a few months ago. Hydrus for me is a lowest common denominator image-tagging software, unfortunately. I feel I have really basic complaints with it, and I feel I discovered the program so late… I am really surprised it's not delivering in even my extremely superficial use of the program- I don't even rip artist galleries (I'd only use it for pixiv), since I tried once but it started from the newest first, and like I said somewhere above, I didn't see a way to broadly sort it all by "creator upload order"… so I cancelled it, deleted what it downloaded to far, and started from the artist's gallery from the beginning… But this obviously doesn't fix the problem, since I can't inject images arbitrarily between the order I import it in n shit… this "creator upload order" sort being something I could set would allow me to basically dump a puzzle on a table and conmplete it using the cover (the artist's gallery) as the guideline for which order to set it in, you know… So I am grateful for Hydrus, since it is infinitely better than just copying images across Windows folders, but it does not satisfy me yet, nor anywhere near… but it lowest common denominator tags images, so I use it, and remain grateful for it, still. I have no idea what the character limit for 8chan is, so I suppose I just have to split it in two until it fits then split it up further from there.

I started reading my three posts above until I lost interest (pretty useless habit) and the only two other complaints I can think of are collections being treated the same as images(files in Hydrus), and thumbnails being easier to change the size of, and me being able to set custom thumbnail sizes and swap between custom sizes freely, like in Windows (though they just give you a few to choose from in windows- you can't choose how big each option is). Actually I have a few more complaints. Compared to JPEGview, Hydrus is much better at scaling images for me, however Hydrus is shit tier at everything else. When my mouse is hovering over a spot, making the viewer go to 100% zoom doesn't zoom into where the mouse is hovering; it zooms into the top of the image. This is essentially death, because scrolling is a laggy mess of the image ghosting behind, and it's much slower than JPEGview doing the same thing flawlessly. However I do appreciate being able to drag images out of the frame of my monitor- I've never had that freedom before, so it makes viewing manga pages really nice. Tbh I was moving around in Hydrus to job my memory of things to criticize, but the tag dropdown got stuck again and I had to restart the client. In case that's impossibly not reported as a bug yet, here's me reporting it… About the "collections being treated the same as images" thing, I mean currently my go-to arbitrary collection method is to use my arbitrary namespace of "pool:". But for a long time I'd actually flood each collection with tags, for example, "pool:semahiro, string bikini, etc. etc." So tagging for me is unfortunate since before I discarded this to instead just mash my keyboard to create a distinct collection tag under this namespace, I actually just flooded collections that will never gain content again with tags. So every time I go to tag anything I have to deal with an infinite amount of basically false tags that I would never use to for tagging purposes. This problem could have been undermined completely had collections by default not had names- they remained distinct even without a namespace, shit like that. And these collections should have their own search- It should basically be a mirrored Hydrus, exclusively for collections of images. And each collection can be tagged as if it were an image. So I could search "string bikini" in my images Hydrus search, and it'll yield individual images tagged that way. Or I can search "string bikini" in my collections Hydrus search, and I'll receive a collection tagged that way. I guess it makes for poorer tagging, since you could just search for the individual image instead, always, but my problem is just that I have infinite tags like "pool:semahiro, askhsdkjahsddjkdks" just for the sake of collecting sets of semahiro images. And before I started spamming in the name just to make the tag, I would tag it like "pool:semahiro, string bikini, looking behind, beach…" It just flooded my search results with things I would never tag with again. And before I started spamming in the names there was a long downtime of my just refusing to collect images since flooding the collection name with those keywords just took too much time and wasn't worth it. So now those are out of order due to my above complaints. It all sucks. Hydrus doesn't deliver what I need, and I don't think I need that much, but I'm grateful I can tag with it anyway.

>>9481 4) It is here. Just poorly advertized. Before hitting "Import now" button, use "add tags based on filename" and switch to "advanced". There you can add "page:x" tags by typing "page:#" into "regexes" input and page into "#namespace". Or you can use a regex to add number based on filename. Admittedly, it's harder than I wish it would be, but it is there. Same thing about storing filenames.

>>9481 Ehh… While I understand your frustration, I think you kinda missed what hydrus is about. It's a media-management software, sure, but it was created with *booru principles in mind (I guess). 1. That's categorizing stuff in even greater detail. Honestly, to me there's no difference between the tag trees system you suggest and the one that hydrus is using. Basically I would just tag this photo as "string clothing, apron, tied behind neck". Of course it leaves some ambiguity. There are two girls in an image, one is wearing only an apron and has blonde hair, the other one has black hair and wears an apron on top of normal clothing. Now in hydrus it would look like: "2 females, naked apron, blonde hair, black hair". And you would create something like: "[female 1: [apron: naked apron], blonde hair]; [female 2: apron, black hair]". Less ambiguity? Sure, but in the end it all blends together. Say you searched for "naked apron", you get the same image as I do, the only difference is how we described it (I guess, didn't give it much thought). 2. I guess you could use wildcards. But I do agree that the search aspect could be improved (please add tag1 OR tag2 dev :c). But since almost every aspect of hydrus is being developed we just have to give it some time. 3. Well it could have been save on import. You could have either inferred the page number from the filenames. IIRC hydrus also has the ability to number stuff while it's imported though it's been a ling time since I stored paged stuff in hydrus. 5. Hard not to agree on this one. Since the hydrus only keeps the dates when something was imported it really can be a hassle to see stuff ordered chronologically. From what I can recall you can set the default import settings, so setting something like "filename:$filename" as the default should be possible. The forced lowercase tag was implemented so there won't be any discussions as to how things are spelled. Sometimes it breaks stuff I guess. Too some people it comes as relief cause we don't have to care about capitalization.

>>9481 1: Define parent/child tag relationships and/or use namespaced tags. Auto-apply appropriate parent tags as you apply their children. Search for the intersection of tag:subtag and subtag:subtag2. Problem solved. 2: It seems you grew attached to a particular implementation of dumbed down regex, this is probably a good thing for dev to look into implementing, but for now, why not just use regex? 3: Use import timestamp to order them by, there's nothing to "respect". I'm not sure what you think this is, were you importing under 1k files total or something? 5: This is indeed a failure, but seems like user error to me. 6: Tag siblings is bad in Hydrus no matter how many people shill it, use parent/child relationships on tags instead wherever possible, only don't fuck it up 7: Custom namespaces are piss easy are you high my man 8: The lowercase autism is a carryover from Linux filename/tag autism I think, Linux doctrine holds that caps don't exist on a system level (sometimes) because it leads to confusion (because a lot of the guys who work on Linux are hacks, BSD is superior in every way btw)

>>9493 Wait, why are tag siblings bad? I need some enlightenment I recently started making use of them since there often are discrepancies between tags among the boorus, and i can't say i'd rather use parents on artists names (would look like an image has multple artists when it's just one), or character names and copyright names

>>9481 3. I hate this so much, lots of times i find myself importing a bunch of images where among them there are some sequences meant to go in order 1-2-3-4 or even just pairs And despite my best efforts to have them in order while importing (so that at least the import time is going to reflect and respect the correct image order), it always finds a way to fuck it up by putting random images in-between those sequences My only options then are to import them separately or to tag them "sequence:##, order:#", but i find both of these to be horrible solutions First one fucks up the "true" date in which i have got them, since because i have to upload them separately from the rest they get put at the end or at the start of what could be a month worth of images Second method is first a pain having to create a tag and order for every sequence of 1-2+ images you have, and second you're then forced to use the tags to the sorting fucking it up, for example if you search for creator-sequence-order (with secondary as newest), all the sequenced images are gonna be sorted after all the non-sequenced, despite the fact that they could be the newest ones I don't see why it has such an hard time to remember that they were already uploaded in the correct order, there weren't any pictures in between so why does hydrus thinks otherwise?

>>9493 Just got it why 5) (no creator upload order) is bad. Imagine pulling one picture from 2 boorus. That means 2 creator times. Which one is to be selected? Because selecting one will break ordering of the second one.

DOWNLOAD BY TWITTER ACCOUNT Way too many jap artists either post exclusively on twatter or post shit there that they don't on pixiv (whether uncensored or just not there). It's annoying to copy each tweet URL into Hydrus. Also, a twitter login thing would probably be necessary because of private accounts.

>>9555 Seconded. Scraping by hand is a pain, and I can't find a no-api-key scraper.

The thing I'd want most is being able to disable(/enable) PTR tags on an image-by-image basis, so I could only use my local tags for certain images that I want organized completely my way (ie doujins and serialized work), while still using the PTR for everything else to avoid having to tag too much. There's workarounds but this would still be nice. >>9305 You can fix that in settings, either by increasing time delay before the autocomplete fires or by disabling autocomplete entirely.

>>9481 >1.Lack of tag trees You might like Daminion (pic related) or Adobe Lightroom if you can stomach the eye rape interface.

>>9683 One of the issue is that some accounts are protected. Example: https://twitter.com/uzusio04

>>9493 For #2, dumb-search to smart regex translator? Anyone? Maybe use https://github.com/VerbalExpressions/PythonVerbalExpressions ?

>>9481 >>9482 >>9483 Most of this garbage is user error / lack of knowlege. Since others have explained some of your other gripes pretty well, I'll explain custom namespaces. It's SUPER EASY if you took more than a minute to look at the options. Just go to file > options > tag summaries. From there, just select what part of the thumbnail you want to change and you can add whatever namespaces your little heart desires. You can also change their colors so the dark theme won't look as bad (pic related). You can also go to file > options > sort/collect to make custom namespace sorting options for your shitty porn collection, and file > options > colours will let you make your namespaces look pretty.

https://www.ebay.com/itm/HP-DL380p-G8-16x-SFF-Intel-XEON-E5-2600-v2-Serverschmiede-Server-Konfigurator/222417019581?hash=item33c91662bd:g:CdgAAOSw11ZZeL8A https://www.ebay.com/itm/Cisco-UCS-C240-M3-16x-SFF-32GB-4x-8GB-2x-Xeon-E5-2660-2-20-GHz-Raid-Server/222438386802?hash=item33ca5c6c72:g:D0gAAOSwuxFYxqFx https://www.ebay.com/itm/HP-Proliant-DL380p-G8-Gen8-Server-16x-2-5-SFF-XEON-E5-2600-v1-v2-P420i-Raid-CTO/222751766786?hash=item33dd0a3902:g:TPAAAOSwZq5bqiIR:rk:5:pf:0 https://www.ebay.com/itm/Cisco-UCS-C240-M3-16x-SFF-8GB-2x-4GB-2x-Xeon-E5-2660-2-20-GHz-Raid-Server/222438386145?hash=item33ca5c69e1:g:OYkAAOSwB-1YxqG6:rk:28:pf:0 https://www.ebay.com/itm/Dell-PowerEdge-R720-16x-SFF-XEON-E5-2600-Serverschmiede-Server-Konfigurator/322381837178?hash=item4b0f74737a:g:1zIAAOSwH9lZeL2y:rk:35:pf:0 https://www.ebay.com/itm/Cisco-UCS-C240-M3-16x-SFF-32GB-8x-4GB-2x-Xeon-E5-2660-2-20-GHz-Raid-Server/222438386453?hash=item33ca5c6b15:g:ftwAAOSwsW9YxqGX:rk:46:pf:0