/hydrus/ - Hydrus Network

>>6398

I'm still not sure gelbooru's working correctly; the tag, "a1" gets 798 urls when gel has a tag count of ~3000. I know the tag counts displayed on the site aren't the true count, but that still seems too small. Another case: "agenasu" has a count of 233, 180 thumbnails show up when you search it, but my subscription only finds 96 urls. I've clicked about half of the thumbnails and they're all valid in my browser, too. Any idea?

I'm getting a huge amount of errors after the update.

"Daemon Save Dirty Objects encountered an exception:"

DBException
UnicodeEncodeError: 'ascii' codec can't encode characters in position 0-1: ordinal not in range(128)
Traceback (most recent call last):
  File "include\HydrusThreading.py", line 195, in run
    self._callable( self._controller )
  File "include\ClientDaemons.py", line 247, in DAEMONSaveDirtyObjects
    controller.SaveDirtyObjects()
  File "include\ClientController.py", line 1047, in SaveDirtyObjects
    self.WriteSynchronous( 'serialisable', self.network_engine.bandwidth_manager )
  File "include\HydrusController.py", line 413, in WriteSynchronous
    return self._Write( action, HC.LOW_PRIORITY, True, *args, **kwargs )
  File "include\HydrusController.py", line 121, in _Write
    result = self.db.Write( action, priority, synchronous, *args, **kwargs )
  File "include\HydrusDB.py", line 877, in Write
    if synchronous: return job.GetResult()
  File "include\HydrusData.py", line 1739, in GetResult
    raise e
DBException: UnicodeEncodeError: 'ascii' codec can't encode characters in position 0-1: ordinal not in range(128)
Database Traceback (most recent call last):
  File "include\HydrusDB.py", line 525, in _ProcessJob
    result = self._Write( action, *args, **kwargs )
  File "include\ClientDB.py", line 10302, in _Write
    elif action == 'serialisable': result = self._SetJSONDump( *args, **kwargs )
  File "include\ClientDB.py", line 8172, in _SetJSONDump
    ( dump_type, version, serialisable_info ) = obj.GetSerialisableTuple()
  File "include\HydrusSerialisable.py", line 136, in GetSerialisableTuple
    return ( self.SERIALISABLE_TYPE, self.SERIALISABLE_VERSION, self._GetSerialisableInfo() )
  File "include\ClientNetworking.py", line 1128, in _GetSerialisableInfo
    all_serialisable_trackers = [ ( network_context.GetSerialisableTuple(), tracker.GetSerialisableTuple() ) for ( network_context, tracker ) in self._network_contexts_to_bandwidth_trackers.items() if not network_context.IsEphemeral() ]
  File "include\HydrusSerialisable.py", line 136, in GetSerialisableTuple
    return ( self.SERIALISABLE_TYPE, self.SERIALISABLE_VERSION, self._GetSerialisableInfo() )
  File "include\ClientNetworking.py", line 1460, in _GetSerialisableInfo
    serialisable_context_data = self.context_data.encode( 'hex' )
  File "encodings\hex_codec.py", line 24, in hex_encode
UnicodeEncodeError: 'ascii' codec can't encode characters in position 0-1: ordinal not in range(128)


Database Traceback (most recent call last):
  File "include\HydrusDB.py", line 525, in _ProcessJob
    result = self._Write( action, *args, **kwargs )
  File "include\ClientDB.py", line 10302, in _Write
    elif action == 'serialisable': result = self._SetJSONDump( *args, **kwargs )
  File "include\ClientDB.py", line 8172, in _SetJSONDump
    ( dump_type, version, serialisable_info ) = obj.GetSerialisableTuple()
  File "include\HydrusSerialisable.py", line 136, in GetSerialisableTuple
    return ( self.SERIALISABLE_TYPE, self.SERIALISABLE_VERSION, self._GetSerialisableInfo() )
  File "include\ClientNetworking.py", line 1128, in _GetSerialisableInfo
    all_serialisable_trackers = [ ( network_context.GetSerialisableTuple(), tracker.GetSerialisableTuple() ) for ( network_context, tracker ) in self._network_contexts_to_bandwidth_trackers.items() if not network_context.IsEphemeral() ]
  File "include\HydrusSerialisable.py", line 136, in GetSerialisableTuple
    return ( self.SERIALISABLE_TYPE, self.SERIALISABLE_VERSION, self._GetSerialisableInfo() )
  File "include\ClientNetworking.py", line 1460, in _GetSerialisableInfo
    serialisable_context_data = self.context_data.encode( 'hex' )
  File "encodings\hex_codec.py", line 24, in hex_encode
UnicodeEncodeError: 'ascii' codec can't encode characters in position 0-1: ordinal not in range(128)

>>6403

>>6404

I also got these when I first updated and ran subscriptions as normal, but after restarting the client a few times I do not encounter them.

Everything works fine here, feels good to be able to turn on all my subscriptions again.

>>6405

>>6404

>>6403

Same, restarting doesn't help, all my subs are fucked.

>>6407

Should add that restarting my client doesn't fix it for me.

>>6403

If you have this problem, please pause your subscriptions through services->pause, then go into review bandwidth, and delete any history for your subs.

I think maybe your subs have unicode names? Could you post some examples if so?

If you do have unicode sub names, renaming them in manage subscriptions should fix this problem for now.

I will have a proper fix for this next week.

>>6409

Most of my subs have japanese names with unicode letters, and I have 500+ subs, so I guess I'm fucked until next wednesday.

>you can't have more than 128 page

Warn me, but please don't stopping to loading de facto deleting these pages.

I can't recover them now, using last session doesn't do it, stopping the same way..

I should have saved the session before updating I guess, but I really didn't expect that innovation.

It wasn't unstable, just slow to start (bunch of 3/4 illustration pages.)

I had to revert to a week old update and revert my version to cut it in half.

>>6411

That behavior was part of the previous update, which I have now read.

I hope I can still somehow restore it.

I think the best way would have been to simply refuse to open any new page until the session is under 127 page at least.

One of my pixiv subscriptions will only download some ~90 or so files (~70mb) before stopping. The only way to get it to continue as I want it to is to delete the bandwidth usage history. Is this intended?

>>6401

did some testing with the gallery downloader page. it looks like it's detecting that there are less pages than there really are?

a1: found 1148 urls over 29 pages, exists 55 pages

agenasu: found 96 urls over 3 pages, exists 5 pages

aki99: found 168 urls over 4 pages, exists 8 pages

apostle: found urls over 4 pages, exists 5 pages

Ok, got something weird

a dialogue box will open in the upper left hand corner, then close itself

one time it happened, it locked me out of the program entirely because apparently it wanted something but was behind the client but I couldn't do anything to the client otherwise and just now I saw it again but it just closed itself on its own. I don't know what it is because it doesn't stay open long enough or in the foreground long enough for me to see what it is.

I updated all my Hentai Foundry subscriptions and left their new file tabs open, and closed Hydrus. Today I opened Hydrus again, but now some of the tabs have no files in them. No idea what happened there, the tabs don't seem to have saved properly. But only a few of them.

I added a HF subscription to this artist, but Hydrus only finds one of the images…

http://www.hentai-foundry.com/pictures/user/Mosq

I'm hitting a lock up that seems to be triggered by subscriptions downloading new images while the client is minimised.

When I maximise the client again an extra window pops up in the top left, and I think I can make out the text 'similar' but it closes too quickly. I'll try to get a screenshot.

That window disappears and I get the following exception. At this point I can't interact with the client at all, I had to copy this from the console I started the client from.

2017/07/30 17:21:13: Uncaught exception:
2017/07/30 17:21:13: PyAssertionError
C++ assertion "IsRunning()" failed at ..\..\src\common\evtloopcmn.cpp(83) in wxEventLoopBase::Exit(): Use ScheduleExit() on not running loop
  File "F:\hn\art-266\include\ClientGUIPopupMessages.py", line 943, in TIMEREvent
    self.GetParent().DoOK()
  File "F:\hn\art-266\include\ClientGUITopLevelWindows.py", line 512, in DoOK
    self.EndModal( wx.ID_OK )
  File "C:\Python27\lib\site-packages\wx-3.0-msw\wx\_windows.py", line 809, in EndModal
    return _windows_.Dialog_EndModal(*args, **kwargs)

Upgraded from v264a, dunno if it happened on 265.

I just want to point this out: https://wxpython.org/news/wxpython-4.0.0b1-release/index.html. Months ago I had mentioned a Python 3-compatible wxPython release (Project Phoenix) to you and you had expressed interest in taking a look at a potential port to Python 3, once a release was available: this seems to be the case now: https://pypi.python.org/pypi/wxPython/4.0.0b1.

Do you have any plans on taking a look at this and assessing how much effort it would be? Python 2 won't be around (at least officially supported) much longer, and general usage is steadily decreasing. Many aspects of Python 3 are also much faster by now.

Thanks as always for your continued efforts!

>>6435

very helpful with the hook, at least on threads that I want to keep up with no matter what, otherwise I think i'm better served by dumping everything into a notepad ++ file and copy paste for the time being.

I'm not even going to pretend to know how programing works, so ignore this if its not a feasible part of the steps to get it nested.

I would take a single thread watcher page that does not display images, but has all the threads listed and keeps them after they 404, but displays they 404, then through right clicking the link you could open all the images that thread imported in a new tab. This is functionality I believe would be there even after you do everything you want to do, If there is any way to get some kind of function skeleton of a multi thread thread watcher page without bending over backwards to do it, that would be great.

on a side note, got the 128 page message, is there a way to open a page after you get it or does it just hard block? I had the need of going over the limit for a moment a few hours ago because the program got locked up because of the messages, i missed an import folder import, and had to use the file search, but when I wanted to ctrl+A and open all files in new tab, I couldn't till something was closed, I have a lot of closeable shit open so its not a big issue for me, but it still an issue nonetheless.

>>6435

> The switch to Python 3 will be the same deal, but on a slightly longer timeframe. Next year, probably, but we'll see.

Sounds great. I hope at that time you might also consider being more active on GitHub, especially regarding contributions,

>>6438

I have relatively recently got a new computer, so I don't have every program I use installed, one being photoshop, or I would mock something up for you to consider. but because I think you don't want a cluster fuck of 100~ watched threads images all in one tab, take this as an idea

the left side is blank but a placeholder for all the shit when you get the threads.

thread watcher itself should act on its own weather the tab is open or not, kind of like how duplicates still retains its place in searching,

Where images would usually be is the thread watcher mass threads

Now I don't know what is possible to do in the program without hitting a limit, take everything I suggest as either a simple dont need to double click function, or double clicking on the thread pops a window up that presents these functions.

[Thread url][user text][Images found][Images imported][Already in DB][Failed][Deleted][Green/yellow/red circle indicating can check, something's wrong and 404][Check interval][Time to next check][Check now]

the thread url would be a click to copy the address incase you want to see it in browser

user text is just a note, as some threads could be up for days or months depending on the board

The images are self explanatory

The indicator would just be an easy way to see what the status is

check interval and time are self explanatory with a check now button.

and it just goes like that all the way down to how ever many threads you want to keep up with.

now on the side that is a place holder at the top just under the tabs are 3 buttons the leftmost is blanked out the middle one is open in new tab and the right is see images

New tab would do just that, open the selected images in a new tab, nothing different from currently selecting images and opening them in a new tab

the right one would open the images in the same tab however and when seeing the images, the 3 buttons would still be there, but with the right most blanked out, and the left being go back the middle would stay unchanged.

at least thats my idea, you could also scrape the subject of the threads and put them in the user text field, as seeing 4chanx is able to see the subject as its own thing there has to be a hook there somewhere.

>>6438

Sooner or later you'll have to add CloudFlare detection/opening a browser window to solve it for the scapers.

CloudFlare is getting miore and more shitty.

>>6398

I could watch that girl (at the end of the video) double-wielding those swords all day ! :-)