IRC logs of #shogun for Wednesday, 2016-11-23

« Tuesday, 2016-11-22 Index Thursday, 2016-11-24 »
--- Log opened Wed Nov 23 00:00:27 2016
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Quit: Leaving.]00:00
-!- HeikoS1 [~heiko@host-92-0-178-129.as43234.net] has joined #shogun00:00
-!- HeikoS1 [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 244 seconds]00:15
@sukeyPull Request #3534 "LinalgRefactor - Cholesky"  synchronized by OXPHOS - https://github.com/shogun-toolbox/shogun/pull/353404:42
@sukeyPull Request #3534 "LinalgRefactor - Cholesky - CPU only" - https://github.com/shogun-toolbox/shogun/pull/353405:42
@sukeyPull Request #3562 "LinalgRefactor - Matrix elementwise product"  opened by OXPHOS - https://github.com/shogun-toolbox/shogun/pull/356206:24
-!- praisethemoon [~praisethe@41.226.248.123] has joined #shogun08:46
CaBahey people15:57
praisethemoonCaBa, o/16:19
-!- HeikoS [~heiko@217.138.5.14] has joined #shogun16:57
-!- mode/#shogun [+o HeikoS] by ChanServ16:57
-!- praisethemoon [~praisethe@41.226.248.123] has quit [Quit: Leaving]17:26
-!- HeikoS [~heiko@217.138.5.14] has quit [Ping timeout: 268 seconds]17:32
-!- HeikoS [~heiko@217.138.5.14] has joined #shogun17:59
-!- mode/#shogun [+o HeikoS] by ChanServ17:59
-!- HeikoS [~heiko@217.138.5.14] has quit [Ping timeout: 244 seconds]18:36
CaBawiking: happen to be around?21:24
@wikingyes21:30
CaBawiking: given a CParamterCombination object, can i iterate through the tree and access the parameters?21:37
CaBawiking: or doesn't it offer a public interface for that?21:38
@wikingmmm21:38
@wikinggood questuin21:38
@wiking*question21:38
@wikingif not then we should expose it21:38
@wikingwhat does doxygen say?21:39
@wikinghttp://shogun.ml/api/latest/classshogun_1_1CParameterCombination.html21:40
CaBawiking: well there is a get_parameter(const char *name, CSGObject *parent) function - i'm not sure if it is for that...21:40
@wikingdoesn't look like you gonna get21:43
@wikingi mean that you can iterate through with that api21:43
CaBawiking: with that get_parameter() function?21:44
@wikingyes21:44
@wikingit'll give you a param21:44
@wikingwith a given name and parent21:44
@wikingbut this way21:44
@wikingyou would need to know21:45
@wikingthe name and parent21:45
@wikingi guess if you have the reference21:45
CaBawhat's the parent?21:45
@wikingparent of the given param21:45
@wikingso manually21:45
@wikingif you have the reference on the ROOT item in the tree21:45
CaBaso how do i get one step down from the root using this function?21:45
@wikingthen you could say21:45
@wikingget_parameter("name of param", ROOT);21:45
@wikingand then like this21:45
@wikingone by one21:45
@wikingyou can iterate through the whole tree21:46
CaBaso i do21:46
@wikingi'm wondering whether this is the only way21:46
CaBaroot->get_paramter("name", root)21:46
CaBa?21:46
@wikingi would hope that there's21:46
CaBai pass root to itself?21:46
@wikinga better function for this21:46
@wikingbut i never really looked into this part of shogun unfortunately21:46
@wikingdidn't dare to ;)21:46
@wikingyeah21:46
@wikinghttp://shogun.ml/api/latest/ParameterCombination_8cpp_source.html#l0025021:46
@wikingas you can see21:47
@wikingthis is the implementation21:47
CaBawiking: #255 will never evaluate true for the root node21:47
CaBawiking: hm, maybe it doesn't have to...21:48
@wiking255?21:48
CaBaline21:48
CaBafrom your lin21:48
CaBalink21:48
@wikingif (m_param)21:48
@wiking?21:48
CaBayes21:48
CaBam_param is a Parameter* member, which should be a nulptr for root i think21:49
@wikingbtw just for the record21:49
@wikinghttp://shogun.ml/api/latest/ParameterCombination_8cpp_source.html#l0030721:49
@wikingthis will print you the tree21:49
@wikingif yo udont wanna print21:49
@wikingthere's the logic21:49
CaBawiking: yes, that's what i want to do from the outside21:49
@wikinghow to actually traverse the tree21:49
@wikingok so there's the code21:49
@wikingright?21:49
CaBawiking: no, whats in that function does not work through public interfaces21:50
@wikingok21:50
@wikingso create a PR21:50
@wikingthat uses that logic21:50
@wikingand does21:50
@wikingiterator()21:50
CaBa>.<21:50
@wikingand basically you implement an iterator for that whole thing21:50
@wikingpretty simple21:50
@wikingyou have there the code yourself how to iterate through the tree21:50
@wikingbut instead of printint21:50
@wikingprinting21:50
@wikingyou return always the next element21:51
@wikingit's almost a copy paste shit21:51
@wikingand you are done21:51
CaBano, extremely frustrating if that whole logic of that data structure is a mystery to you...21:51
@wikingyeah but21:51
@wikingit's not21:51
@wikingthere's the src21:51
@wikingthis is why it's opensource21:51
@wikingyou can see what's happening under the hood21:51
@wikingand add new functionality21:51
@wikingthat might be interesting for other people as well21:51
@wikingand in this case21:52
@wikingthat request that you have21:52
@wikingis super valid21:52
@wikingas well as that the solution for it is super easy21:52
@wikingas you just copy paste shit21:52
CaBai know it's probably unfair to complain, but the level of documentation is really not such that one can easily understand the code by reading it21:52
@wikingare you serious?21:53
@wikingi have not21:53
@wikingever looked into this src21:53
@wikingjust now21:53
@wikingand in 2 minutes i gave you an answer21:53
@wikinghow you could do it21:53
@wikingso please21:53
@wikingbut21:53
@wikingif you think that it's not adequate enough21:53
@wikingplease go ahead21:53
@wikingand send in PRs21:53
@wikingfor having better documentation21:53
@wikingwe are an open source project21:53
@wikingwith 0 income21:53
@wikingall of us does this out of passion21:54
@wikingthis is as much we could do free21:54
@wikingif you think that there's some more need to be fixed21:54
@wikingplease go ahead21:54
@wikingwe know that we are not perfect21:54
@wikingand never will be21:54
@wikingbut we are trying to make things better as much as we can21:54
CaBai know, that's why i raised the fairness disclaimer. i totally sympathise with your situation, really do...21:54
@wikingso ok21:54
@wikingthen if you want21:54
@wikingpatch it21:54
@wikingand send it21:55
@wiking:)21:55
@wikingother than that21:55
@wikingunfortunately personally21:55
@wikingi have other priorities21:55
@wikingin this project21:55
@wikingso i wont be able to help you with it21:55
@wikingbut again21:56
@wikingthe solution21:56
@wikingis pretty simple21:57
CaBawiking: i know. i also didn't expect you to. i just wanted to know if i was overlooking an interfae that's there. totally aware that if there wasn't any, nobody was going to make me one..21:57
@wikingmmm based on the21:57
@wikinginterface of that class21:57
@wikingit seems there's no iterator for it21:57
@wikingbut21:57
@wikingi could imagine that there's somewhere else21:58
@wikingbut there's just a small chance for it21:58
@wiking:)21:58
CaBawiking: wirred bug costed me half my day today... i patched something in shogun to take an std::function and used it from my program code... shogun i build with gcc-6 (for openmp) and my own code i build with clang...23:17
CaBawiking: that didn't go well at all... for whatever reason the std::function object didn't make it to the shogun world. only after i started building also my own code with gcc-6 it worked fine.23:18
CaBawiking: c++11 foo still seems to be weak at some points between compilers / libc implementations...23:18
CaBaeven with the latest and greatest xcode and gcc-6...23:19
CaBamaybe just the former...23:19
--- Log closed Thu Nov 24 00:00:29 2016
« Tuesday, 2016-11-22 Index Thursday, 2016-11-24 »

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!