--- Log opened Fri Feb 01 00:00:52 2019 | ||
-!- durovo1 [~durovo@7b.78.2ea9.ip4.static.sl-reverse.com] has joined #shogun | 00:01 | |
-!- durovo [~durovo@79.73.2ea9.ip4.static.sl-reverse.com] has quit [Remote host closed the connection] | 00:01 | |
-!- besser82 [~besser82@fedora/besser82] has quit [Quit: Freedom, Friends, Features, First [fedoraproject.org]] | 00:46 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 00:48 | |
-!- mode/#shogun [+o besser82] by ChanServ | 00:48 | |
-shogun-buildbot:#shogun- Build nightly_default #204 is complete: Success [build successful] - http://buildbot.shogun-toolbox.org:8080/#builders/17/builds/204 | 09:11 | |
-!- braceletboy [67157d53@gateway/web/freenode/ip.103.21.125.83] has joined #shogun | 10:19 | |
-!- besser82 [~besser82@fedora/besser82] has quit [Quit: Freedom, Friends, Features, First [fedoraproject.org]] | 10:24 | |
-!- besser82 [~besser82@fedora/besser82] has joined #shogun | 10:25 | |
-!- mode/#shogun [+o besser82] by ChanServ | 10:25 | |
-!- braceletboy35 [67157d53@gateway/web/cgi-irc/kiwiirc.com/ip.103.21.125.83] has joined #shogun | 10:48 | |
-!- braceletboy35 [67157d53@gateway/web/cgi-irc/kiwiirc.com/ip.103.21.125.83] has left #shogun [] | 10:50 | |
-!- braceletboy [67157d53@gateway/web/freenode/ip.103.21.125.83] has left #shogun [] | 10:50 | |
-!- braceletboy35 [67157d53@gateway/web/cgi-irc/kiwiirc.com/ip.103.21.125.83] has joined #shogun | 10:52 | |
-!- braceletboy35 is now known as braceletboy | 10:52 | |
braceletboy | wuwei[m] are you around? | 11:00 |
---|---|---|
wuwei[m] | braceletboy: hi | 11:00 |
braceletboy | What's the difference between the get_real_matrix and get() functions? | 11:01 |
braceletboy | **gets | 11:01 |
wuwei[m] | get is a template, so in swig we have several specialization | 11:02 |
wuwei[m] | see shogun.i | 11:02 |
braceletboy | So, is the get_real_matrix a part of the old API right? | 11:04 |
braceletboy | for getter functions | 11:05 |
wuwei[m] | no get_real_matrix is actually part of the new tag api | 11:05 |
wuwei[m] | in swig, we cant have template functions | 11:05 |
braceletboy | But the tags readme that you referenced in the issue talks only about the gets function. So, I got confused | 11:06 |
wuwei[m] | see https://github.com/shogun-toolbox/shogun/blob/1e369eafda5c6535c7c87ed12070b679f16deb74/src/interfaces/swig/shogun.i#L234 | 11:07 |
wuwei[m] | get_real_matrix is the same as get | 11:08 |
braceletboy | Okay. Thanks alot :) | 11:08 |
braceletboy | ! | 11:08 |
-!- gf712 [9052084e@gateway/web/freenode/ip.144.82.8.78] has joined #shogun | 11:26 | |
-!- Lefteris [836fb90d@gateway/web/freenode/ip.131.111.185.13] has left #shogun [] | 12:31 | |
-!- Lefteris [836fb90d@gateway/web/freenode/ip.131.111.185.13] has joined #shogun | 12:31 | |
-!- Lefteris [836fb90d@gateway/web/freenode/ip.131.111.185.13] has quit [Quit: Page closed] | 13:53 | |
-!- gf712 [9052084e@gateway/web/freenode/ip.144.82.8.78] has quit [Ping timeout: 256 seconds] | 13:58 | |
-!- gf712 [9052084e@gateway/web/freenode/ip.144.82.8.78] has joined #shogun | 14:16 | |
-!- Lefteris [836fb90d@gateway/web/freenode/ip.131.111.185.13] has joined #shogun | 14:36 | |
-!- Lefteris [836fb90d@gateway/web/freenode/ip.131.111.185.13] has quit [Quit: Page closed] | 15:25 | |
-!- Lefteris [836fb90d@gateway/web/freenode/ip.131.111.185.13] has joined #shogun | 15:25 | |
-!- Lefteris [836fb90d@gateway/web/freenode/ip.131.111.185.13] has quit [Client Quit] | 15:28 | |
-!- Lefteriss [~lefteris@2a05:b400:100:100:999d:d632:891b:81e2] has joined #shogun | 15:29 | |
-!- Lefteriss [~lefteris@2a05:b400:100:100:999d:d632:891b:81e2] has quit [Client Quit] | 15:30 | |
-!- Lefteris [~lefteris@2a05:b400:100:100:999d:d632:891b:81e2] has joined #shogun | 15:30 | |
-!- HeikoS [5aae0410@gateway/web/cgi-irc/kiwiirc.com/ip.90.174.4.16] has joined #shogun | 16:13 | |
gf712 | HeikoS: hi, how are you? | 16:16 |
HeikoS | gf712 hey all good | 16:16 |
HeikoS | watching some nerdy deep mind videos :) | 16:16 |
gf712 | I have a weird python error happening with shogun | 16:16 |
gf712 | haha | 16:16 |
gf712 | the Starcraft one? | 16:16 |
HeikoS | yes | 16:17 |
HeikoS | used to be a big fan | 16:17 |
HeikoS | whats the error? | 16:17 |
gf712 | never tried.. so can their AI play it properly? | 16:17 |
gf712 | I am trying to catch a shogun exception inside swig | 16:18 |
gf712 | but when I catch an exception and return a value, rather than throw that exception I get this error | 16:18 |
gf712 | returned a result with an error set | 16:18 |
gf712 | basically I am trying to check if a parameter is a sgobject | 16:19 |
HeikoS | they consistently beat the worlds better players | 16:19 |
gf712 | so I call $self->get(name); to check if it returns something that is not nullptr | 16:19 |
HeikoS | ok | 16:20 |
gf712 | but because get raises an exception I need to catch it | 16:20 |
HeikoS | shogun c++ exception within .i file swig c++ code you mean? | 16:20 |
gf712 | yes | 16:20 |
HeikoS | let me understand | 16:20 |
HeikoS | what would be the error msg of the get? | 16:20 |
gf712 | https://pastebin.com/zua2eDqc | 16:21 |
gf712 | this is inside the extend block of SGObject in SGBase.i | 16:21 |
gf712 | and I get an error like this https://pastebin.com/XNLtiJ18 | 16:22 |
gf712 | seems to be related to some change in Python3.5 in the way exceptions are handled from the little I found online on this subject.. | 16:23 |
HeikoS | uh | 16:23 |
HeikoS | no idea | 16:23 |
HeikoS | but maybe what you are doing can be done differently | 16:23 |
HeikoS | I have some questions | 16:24 |
HeikoS | why would a successful .get mean that something is an object parameter? | 16:24 |
gf712 | the issue is that I found at least another block of code that catches shogun exceptions and it probably wouldn't work either | 16:24 |
HeikoS | yeah, maybe this could be done without exceptions | 16:25 |
gf712 | because it is then either CDistance or CKernel | 16:25 |
gf712 | right? | 16:25 |
HeikoS | but what if you do | 16:25 |
HeikoS | svm.parameter_is_sg_base("C")? | 16:26 |
-!- bbc_ [67f98750@gateway/web/freenode/ip.103.249.135.80] has joined #shogun | 16:26 | |
HeikoS | I actually have an idea | 16:26 |
HeikoS | why not add a global templated ethod | 16:26 |
HeikoS | bool is_sg_base<T>() | 16:26 |
HeikoS | and then do a | 16:26 |
HeikoS | ah wait | 16:27 |
HeikoS | no :) | 16:27 |
gf712 | yes, I thought that first | 16:27 |
gf712 | but everything is any | 16:27 |
gf712 | at least I think that's where you were going? | 16:28 |
HeikoS | yes | 16:28 |
HeikoS | so | 16:28 |
HeikoS | svm.parameter_is_sg_base("C")? | 16:28 |
HeikoS | this would return a true as well or? | 16:28 |
gf712 | No, that should return false | 16:28 |
gf712 | but instead I get this weird error | 16:28 |
HeikoS | $self->get(name); | 16:29 |
HeikoS | how does this know what the template type is? | 16:29 |
HeikoS | which get method does it call? | 16:30 |
gf712 | but isn't there only one get(const std::string&) ? | 16:30 |
gf712 | it uses the definition at the end of SGObject.cpp | 16:31 |
HeikoS | ah this one | 16:32 |
HeikoS | ok | 16:32 |
HeikoS | got you now | 16:32 |
HeikoS | and this is supposed to handle all base classes | 16:32 |
HeikoS | you could just put the dispatching into a helpe that is called has_sg_object or something and then use that from the get | 16:33 |
gf712 | yes, I could do that. I was going to create a std::no_throw_t overload, but I wasn't sure how to write it | 16:34 |
HeikoS | that is probably neater :) | 16:34 |
HeikoS | but I don't know how it would work | 16:34 |
HeikoS | and the error seems weird | 16:35 |
HeikoS | probably swig stuff | 16:35 |
gf712 | yes the error is very weird | 16:35 |
gf712 | I don't think it's swig though | 16:35 |
gf712 | I think it's python | 16:35 |
HeikoS | ah even better :D | 16:35 |
gf712 | I will try it out in R and see if it gets confused too | 16:35 |
HeikoS | R might not work | 16:36 |
HeikoS | maybe use octave | 16:36 |
HeikoS | R has issues | 16:36 |
gf712 | ah yes I can see | 16:38 |
HeikoS | did you try google this? | 16:39 |
HeikoS | catching exceptions from %extend? | 16:40 |
gf712 | yes, a bit | 16:40 |
gf712 | but I coudnt find any examples of it | 16:40 |
gf712 | I just found the %exception stuff | 16:41 |
gf712 | I also tried to compile without it to see what happens | 16:41 |
gf712 | and I get the same error | 16:41 |
gf712 | I messed around with the shogunPYTHON_wrap.cxx | 16:41 |
gf712 | added lots of messages | 16:41 |
gf712 | and the data is there | 16:42 |
gf712 | but python detects that there should be an exception | 16:42 |
gf712 | but a value is returned | 16:42 |
gf712 | and then gets confused | 16:42 |
HeikoS | lol | 16:43 |
gf712 | there must be some internal error state in python that can't be avoided even with a catch | 16:43 |
HeikoS | funny | 16:43 |
HeikoS | yeah | 16:43 |
gf712 | so I am thinking this might not work as expected https://github.com/shogun-toolbox/shogun/blob/develop/src/interfaces/swig/SGBase.i#L496 | 16:44 |
gf712 | but I haven't tried compiling shogun without hdf5 support to test it | 16:45 |
gf712 | this is also all on my Mac | 16:45 |
gf712 | on python 3.6 | 16:46 |
HeikoS | no idea about this old code :) | 16:52 |
HeikoS | would need to find a test/build that uses it | 16:52 |
-!- bbc_ [67f98750@gateway/web/freenode/ip.103.249.135.80] has quit [Ping timeout: 256 seconds] | 16:53 | |
gf712 | hmm I'll see what octave says | 16:56 |
gf712 | HeikoS do you know yet what day the ws will be? | 16:57 |
gf712 | I would probably only be able to make it to the first weekend you suggested | 16:57 |
HeikoS | just talked to viktor earlier today and I think we were going for the second | 16:58 |
HeikoS | but if you cannot make that | 16:58 |
HeikoS | pls send an email to viktor now | 16:58 |
HeikoS | since already a few cannot make it so we can easily shift dates | 16:58 |
HeikoS | as it didnt matter too much i think | 16:58 |
gf712 | OK | 16:58 |
gf712 | btw are you coming to London next month too? | 16:58 |
gf712 | Lefteris btw Viktor is coming to the ATI on the 19-20 Feb if you are free and want to join? | 16:59 |
Lefteris | Sure! | 17:02 |
Lefteris | Send an email when you know the details (time, etc.) and I will come | 17:03 |
HeikoS | Lefteris do you also want to join the ws btw? | 17:03 |
HeikoS | workshop | 17:04 |
gf712 | Lefteris yup, I'll book a room and send you the details | 17:04 |
Lefteris | That one in Spain? Yes, I replied to Victor. Both dates are fine for me. I am just waiting for the selected dates to book tickets, etc | 17:05 |
HeikoS | ok cool | 17:05 |
HeikoS | we were thinking about the second date but now gf712 may only make the first | 17:05 |
Lefteris | ok, no problem for me | 17:05 |
wuwei[m] | HeikoS: hey | 17:12 |
HeikoS | yo | 17:12 |
wuwei[m] | auto combined_features = wrap(features("CombinedFeatures")); | 17:12 |
wuwei[m] | this doesn't work :) | 17:13 |
wuwei[m] | shall we add some factory for features? | 17:13 |
HeikoS | yes need that | 17:14 |
HeikoS | good point | 17:14 |
HeikoS | but type should be CFeatures still | 17:14 |
wuwei[m] | yeah sure | 17:14 |
Lefteris | HeikoS I pushed the changes for the octave typemap, let me know if there is anything else is needed | 17:16 |
Lefteris | HeikoS I pushed the changes for the octave typemap, let me know if anything else is needed | 17:16 |
Lefteris | damn there is no edit | 17:16 |
Lefteris | :P | 17:16 |
-!- bbc_ [67f98750@gateway/web/freenode/ip.103.249.135.80] has joined #shogun | 17:17 | |
bbc_ | Hello i am new open source and shogun i want to work on issue #4259 it involves changing code in CKernel I am not able to locate Ckernel can someone help | 17:19 |
HeikoS | bbc_ Kernel.h :D | 17:20 |
HeikoS | wuwei[m] check your email | 17:22 |
wuwei[m] | awesome! | 17:23 |
HeikoS | I will write to the lists and stuff later | 17:23 |
bbc_ | Thankyou for the help @HeikoS and one more question should i change the code as shown on the issue page or does it involves to change anything else also , sorry for this type of questions i am new to open source | 17:25 |
HeikoS | bbc_ check our wiki, there is a page for "getting involved" that should answer all your questions | 17:26 |
HeikoS | it also tells you how to make changes and send them to us | 17:26 |
bbc_ | Thanks @HeikoS i am currently going through getting involved page | 17:27 |
HeikoS | bbc_ cool, good luck and let us know if you need any help | 17:28 |
bbc_ | ok | 17:28 |
gf712 | HeikoS just tried octave | 17:31 |
gf712 | so it shows an error | 17:32 |
gf712 | but there is variable assignment | 17:32 |
HeikoS | mabe it is some swig generated code then? | 17:32 |
HeikoS | that causes the target langs to hikcup | 17:32 |
gf712 | hmm not sure what is happening | 17:34 |
gf712 | I'll ask in swig what they think | 17:35 |
HeikoS | good idea! | 17:44 |
-!- HeikoS [5aae0410@gateway/web/cgi-irc/kiwiirc.com/ip.90.174.4.16] has quit [Ping timeout: 272 seconds] | 17:58 | |
-shogun-buildbot:#shogun- Build deb3 - interfaces #412 is complete: Failure [failed configure (failure)] - http://buildbot.shogun-toolbox.org:8080/#builders/37/builds/412 | 18:02 | |
-!- gf712 [9052084e@gateway/web/freenode/ip.144.82.8.78] has quit [Ping timeout: 256 seconds] | 18:06 | |
-!- HeikoS [5067884f@gateway/web/cgi-irc/kiwiirc.com/ip.80.103.136.79] has joined #shogun | 18:30 | |
-!- Lefteris [~lefteris@2a05:b400:100:100:999d:d632:891b:81e2] has quit [Quit: Leaving] | 18:40 | |
-!- HeikoS [5067884f@gateway/web/cgi-irc/kiwiirc.com/ip.80.103.136.79] has quit [Ping timeout: 250 seconds] | 18:41 | |
-!- gf712 [560fc4e5@gateway/web/freenode/ip.86.15.196.229] has joined #shogun | 19:04 | |
-!- gf712 [560fc4e5@gateway/web/freenode/ip.86.15.196.229] has quit [Ping timeout: 256 seconds] | 19:40 | |
-!- bbc_ [67f98750@gateway/web/freenode/ip.103.249.135.80] has quit [Quit: Page closed] | 20:55 | |
-!- braceletboy [67157d53@gateway/web/cgi-irc/kiwiirc.com/ip.103.21.125.83] has quit [Remote host closed the connection] | 23:04 | |
--- Log closed Sat Feb 02 00:00:53 2019 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!