[ / / / / / / / / / ] [ dir / cute / egy / fur / kind / kpop / miku / waifuist / wooo ]

/hydrus/ - Hydrus Network

Bug reports, feature requests, and other discussion for the hydrus network.

Catalog

Name
Email
Subject
Comment *
File
* = required field[▶ Show post options & limits]
Confused? See the FAQ.
Embed
(replaces files and can be used instead)
Options
Password (For file and post deletion.)

Allowed file types:jpg, jpeg, gif, png, webm, mp4, swf, pdf
Max filesize is 12 MB.
Max image dimensions are 10000 x 10000.
You may upload 5 per post.


New user? Start here ---> http://hydrusnetwork.github.io/hydrus/

Current to-do list has: 714 items

Current big job: finishing off duplicate search/filtering workflow


YouTube embed. Click thumbnail to play.

e2d10a No.3994

windows

zip: https://github.com/hydrusnetwork/hydrus/releases/download/v224/Hydrus.Network.224.-.Windows.-.Extract.only.zip

exe: https://github.com/hydrusnetwork/hydrus/releases/download/v224/Hydrus.Network.224.-.Windows.-.Installer.exe

os x

app: https://github.com/hydrusnetwork/hydrus/releases/download/v224/Hydrus.Network.224.-.OS.X.-.App.dmg

tar.gz: https://github.com/hydrusnetwork/hydrus/releases/download/v224/Hydrus.Network.224.-.OS.X.-.Extract.only.tar.gz

linux

tar.gz: https://github.com/hydrusnetwork/hydrus/releases/download/v224/Hydrus.Network.224.-.Linux.-.Executable.tar.gz

source

tar.gz: https://github.com/hydrusnetwork/hydrus/archive/v224.tar.gz

I had a good week, although I was a little short on time. I fixed the tumblr downloader and managed to rewrite the image rendering pipeline to work a lot faster and support more quality options.

image rendering

So I've improved how static (i.e. non-animated) images are loaded and rendered in the client. There is less redundancy and a faster and more advanced library does all resizing.

Furthermore, the new zoom quality options for each static image mime (jpg/png/static gifs) under options->media are now obeyed. If you have a slow computer, you can set 'nearest neighbour' to speed things up, or you can set higher quality algorithms like '8x8 Lanczos' to make more beautiful zoomed-in images. If you have a graphics card that supports OpenCL (usually anything by AMD/Nvidia), then resizing will be hardware accelerated as well!

I've also added an important 'load images with PIL' option to the same 'media' page. Turning it off will revert to an old mode that can make images load about twice as fast, but some users reported crashes on certain specific images. Try it if you want and let me know how you get on.

full list

- rewrote the static image rendering and caching pipeline – images are now resized on the fly, and only the master image is cached

- all image rendering and zooming is faster and consumes less memory

- image rendering now obeys the zoom quality options in the 'media' options page!

- static image zooming will take advantage of OpenCL (video card acceleration) wherever available

- added an option to the 'media' panel to allow fast but potentially unstable opencv image loading

- fixed the tumblr gallery downloader (tumblr put an extra character in their API response, wew)

- the debug code profiler now prints more information

- cleaned and updated reducing lag help page to match new profiler

- 'recent' suggested tags are now sorted by recency

- the thread watcher now checks url history as well as md5 match to compensate for cloudflare optimisation making for unreliable api and causing dupe downloads

- misc fixes

next week

I also worked a little more the new html parser this week. As I expect it to be used extensively in the impending downloader engine overhaul, I've not been certain whether I should plan for that and make a it more comprehensive and complicated or whether I should focus on suggested tags parsing for now and upgrade it later. I've decided to do the latter, not getting bogged down trying to make it 'perfect', and just get some actual running code out. So, I'll be doing that, I think, and some small other stuff.

23db22 No.3995

File: ff414a43a349c65⋯.jpg (57.76 KB, 894x894, 1:1, ff414a43a349c655fd568d5eb5….jpg)

The changes to image rendering are pretty noticeable. Some of my >20 MB images now load faster and look nicer. So props for that!


a17c51 No.3997

the import option - tags page seems to be broken. On trying to add a new explicit tag:

AttributeError

'list' object has no attribute 'add'

File "include\ClientGUICommon.py", line 1985, in EventKeyDown

ListBoxTags.EventKeyDown( self, event )

File "include\ClientGUICommon.py", line 1394, in EventKeyDown

self._Activate()

File "include\ClientGUICommon.py", line 1869, in _Activate

self._callable( predicates )

File "include\ClientGUIACDropdown.py", line 275, in BroadcastChoices

self._BroadcastChoices( predicates )

File "include\ClientGUIACDropdown.py", line 1107, in _BroadcastChoices

self._chosen_tag_callable( tags )

File "include\ClientGUIDialogs.py", line 2317, in EnterTags

self._tags.EnterTags( tags )

File "include\ClientGUICommon.py", line 2391, in EnterTags

self._tags.add( tag )


7d0c1a No.3998

>>3997

i just came here to report this - it's not just you


7d0c1a No.3999

My e621 subscription now pulls "150 new links discovered", but then fails, and spits

SSLError

('The read operation timed out',)

Traceback (most recent call last):

File "include\ClientImporting.py", line 2493, in Sync

self._SyncQuery( job_key )

File "include\ClientImporting.py", line 2312, in _SyncQuery

( page_of_urls, definitely_no_more_pages ) = gallery.GetPage( self._query, page_index )

File "include\ClientDownloading.py", line 757, in GetPage

data = self._FetchData( gallery_url )

File "include\ClientDownloading.py", line 727, in _FetchData

return HydrusGlobals.client_controller.DoHTTP( HC.GET, url, request_headers = request_headers, report_hooks = report_hooks, temp_path = temp_path )

File "include\ClientController.py", line 380, in DoHTTP

def DoHTTP( self, *args, kwargs ): return self._http.Request( *args, kwargs )

File "include\ClientNetworking.py", line 300, in Request

( response, size_of_response, response_headers, cookies ) = self._DoRequest( method, location, path, query, request_headers, body, follow_redirects = follow_redirects, report_hooks = report_hooks, temp_path = temp_path )

File "include\ClientNetworking.py", line 249, in _DoRequest

( parsed_response, redirect_info, size_of_response, response_headers, cookies ) = connection.Request( method, path_and_query, request_headers, body, report_hooks = report_hooks, temp_path = temp_path )

File "include\ClientNetworking.py", line 667, in Request

response = self._GetResponse( method_string, path_and_query, request_headers, body )

File "include\ClientNetworking.py", line 372, in _GetResponse

return self._connection.getresponse()

File "httplib.py", line 1136, in getresponse

File "httplib.py", line 453, in begin

File "httplib.py", line 409, in _read_status

File "socket.py", line 480, in readline

File "ssl.py", line 756, in recv

File "ssl.py", line 643, in read

SSLError: ('The read operation timed out',)


1e9b27 No.4000

>>3999

if its recent, I was just having issues with their site.


7d0c1a No.4005

>>4000

at the time of the error, i could browse their site fine…

that said, it is now working for me again, so this is no longer an issue


037454 No.4009

>>3997

> pulls "150 new links discovered", but then fails, and spits

yep, same thing happened to me Q_Q, what ever will i do now..


161109 No.4010

>>4009

This problem occured to me,too.

So I replaced line 2322 in "include/ClientGUICommon.py"

        self._tags = list( tags )

with

        self._tags = set( tags )
.

My hydrus is working well for now.


161109 No.4011

Oh sorry,I meant to reply to >>3997.


74f53b No.4020

Does hydrus take forever to shut down for anyone else? Seems to stick waiting for gui / daemons to shut down for about 30 seconds. Previous releases only took ~5 seconds


e2d10a No.4024

>>3995

Great!

>>3997

>>3998

>>4010

>>4011

Thank you for this report. I changed how tags are stored in some cases to fix something else, but then it broke this. I will fix this for next week.

>>3999

>>4000

>>4005

>>4009

Unfortunately, my networking engine is pretty shaky and does not recover from errors very well. I will add a catch for that ssl error and can add more if you see anything else new, but the ultimate fix here is to wait until I can write a better system that will retry on these sorts of fails. Please hang in there.

>>4020

If you want, please try closing the client with both 'db' and 'pubsib' profile modes set (under the help->debug menu). This will spam a huge amount of profile stuff to your install_dir/db/client.log with timing information. Email me that and I'll see if anything there is taking too long.


17eb77 No.4029

Hydrus became stupidly slow to start up (event from SSD). Not sure if it's this particular updates as I skipped a few. Where do I start digging?


a17c51 No.4031

File: c591d5846b9bd40⋯.png (1.18 KB, 211x88, 211:88, 2016-09-27 00_23_04-hydrus….png)

I upgraded to windows 10 recently, and sometimes my import dialogue boxes end up looking like this. they still function correctly, it's just hard to read


74f53b No.4032

>>4031

It does it on win 7 and 8 too.


e2d10a No.4060

File: a23834a6447886a⋯.jpg (502.96 KB, 1319x1664, 1319:1664, a23834a6447886a00036743d1a….jpg)

>>4029

Now I think of it, the program doesn't have good profiling support for boot yet. What happens if you hit help->debug->load whole db into disk cache and then shut it down and load it up again? It is real quick, or still delayed? How many tabs do you have open, and approx how many files do they have?

What settings do you have under file->options->speed and memory->disk cache?

>>4031

>>4032

Thank you for this report. I've seen this myself. I'm going to have another look at it again this week.




[Return][Go to top][Catalog][Post a Reply]
Delete Post [ ]
[]
[ / / / / / / / / / ] [ dir / cute / egy / fur / kind / kpop / miku / waifuist / wooo ]