--- Log opened Wed Aug 03 00:00:49 2016 | ||
-!- sanuj [~sanuj@117.203.8.128] has joined #shogun | 07:51 | |
sanuj | wiking, `#cmakedefine HAVE_JINJA2 @JINJA2_FOUND@` is not setting HAVE_JINJA2 | 07:53 |
---|---|---|
sanuj | i'm using | 07:53 |
sanuj | #define HAVE_JINJA2 @JINJA2_FOUND@ | 07:53 |
sanuj | http://public.kitware.com/pipermail/cmake/2011-March/043464.html | 07:53 |
sanuj | even #define is not able to set it | 08:37 |
@wiking | sanuj, i explained in that PR | 08:40 |
@wiking | that you should SET | 08:40 |
@wiking | a variable | 08:40 |
@wiking | IF(JINJA2_FOUND) | 08:41 |
@wiking | set(HAVE_JINJA2 1) | 08:41 |
@wiking | ENDIF() | 08:41 |
sanuj | wiking, okay thanks | 08:42 |
sanuj | !! | 08:42 |
@wiking | i mean | 08:42 |
@wiking | there are like | 08:42 |
@wiking | 50 examples | 08:42 |
@wiking | in that CMakeLists.txt file | 08:42 |
@wiking | that does the very same thing | 08:42 |
@wiking | it's ok if you dont know | 08:42 |
@wiking | but it's not that ok that you dont follow something that is in the code itself | 08:43 |
@wiking | these are my 2 cents | 08:43 |
sanuj | wiking, i didn't know | 08:44 |
@wiking | yaeh that's fine | 08:44 |
sanuj | so there is | 08:44 |
sanuj | IF (CURL_FOUND) | 08:44 |
sanuj | SET(HAVE_CURL 1) | 08:44 |
@wiking | again the problem is there are like million of these particular examples | 08:44 |
@wiking | in that file | 08:44 |
sanuj | and then in config.h.in | 08:44 |
sanuj | #cmakedefine HAVE_CURL 1 | 08:44 |
@wiking | yeah | 08:44 |
@wiking | like that | 08:44 |
@wiking | exactly | 08:44 |
sanuj | wiking, cool | 08:45 |
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has joined #shogun | 09:40 | |
-!- sanuj [~sanuj@117.203.8.128] has quit [Ping timeout: 258 seconds] | 10:49 | |
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] | 11:51 | |
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has joined #shogun | 11:54 | |
@wiking | c4goldsw, how's going? | 12:23 |
c4goldsw | wiking I'm good, how are you? | 12:24 |
@wiking | good good | 12:24 |
@wiking | working on scijs :) | 12:24 |
c4goldsw | scijs? | 12:24 |
@wiking | i.e. creating shogun a javascript interface | 12:24 |
c4goldsw | Oh cool! | 12:24 |
@wiking | and be able to use javascript for you data science tasks | 12:24 |
@wiking | :) | 12:24 |
@wiking | i thought it's bs | 12:25 |
@wiking | but then when i twitted yesterday abou tit | 12:25 |
@wiking | people went crazy about it :) | 12:25 |
@wiking | so that's good | 12:25 |
@wiking | :) | 12:25 |
c4goldsw | Where would it run from? | 12:25 |
@wiking | nodejs | 12:25 |
c4goldsw | Ah, ok. | 12:25 |
@wiking | nono no way i wanna do ML in browser | 12:25 |
@wiking | :)))) | 12:25 |
c4goldsw | Haha, good. | 12:25 |
@wiking | it's just that instead of R or python | 12:26 |
@wiking | why not JS? | 12:26 |
c4goldsw | Fair enough - that would provide a good entry point for a lot of web dev people who want to muddle around. | 12:26 |
c4goldsw | Just redid the LARS clean up patch (after Heikos comments). I'm about to send in a PR for a templated LDA. | 12:27 |
c4goldsw | Well, not templated, but it can support different floating point types. | 12:27 |
@wiking | that's great! | 12:28 |
c4goldsw | wiking Do you do Kaggle competions? Just curious. I've started getting into it myself. | 12:28 |
@wiking | hehe not really | 12:28 |
@wiking | no time | 12:28 |
@wiking | and quite frankly | 12:29 |
@wiking | most of the solutions there are very bruteforce ones | 12:29 |
@wiking | i rarely saw an elegant solution | 12:29 |
c4goldsw | Interesting - I'll keep that in mind. | 12:30 |
-!- sanuj [~sanuj@117.203.8.128] has joined #shogun | 12:30 | |
@wiking | i mean usually i see | 12:30 |
@wiking | people generating | 12:30 |
@wiking | KAZIILIONS of features | 12:30 |
@wiking | and then do usually extreme boosting | 12:30 |
@wiking | :) | 12:30 |
sanuj | wiking, for my tags jinja PR, shall i add a test in python_modular to check if things work. (it works on my local) | 12:32 |
@wiking | is it possible to add a meta example for testing this? | 12:36 |
sanuj | wiking, do meta examples test python etc? | 12:36 |
sanuj | i think they only do cpp integration testing currently | 12:37 |
@wiking | fuck yeah | 12:37 |
@wiking | :( | 12:37 |
@wiking | yeah then lets have a fixture for python | 12:37 |
sanuj | wiking, okay! | 12:38 |
@wiking | thnx | 12:49 |
sanuj | lisitsyn, there? | 13:32 |
lisitsyn | sanuj: yeap | 13:32 |
sanuj | lisitsyn, shogun-base.i when generated looks like https://gist.github.com/sanuj/a85986fb35052de761ac251496e90e90 | 13:33 |
sanuj | lisitsyn, we still have not decided what to do with shogun templates | 13:33 |
-!- sanuj [~sanuj@117.203.8.128] has quit [Ping timeout: 276 seconds] | 14:21 | |
-!- sanuj [~sanuj@117.203.8.128] has joined #shogun | 14:25 | |
sanuj | wiking, what happens to data that is returned in python_modular integration tests? | 15:01 |
sanuj | is it necessary to return something? | 15:01 |
@wiking | what do yo umean? | 15:01 |
sanuj | wiking, every modular test has a function that returns something | 15:01 |
sanuj | is it saved somewhere like it is done for the new meta integration tests?? | 15:02 |
@wiking | the data that is being used for python_modular | 15:02 |
@wiking | or any modular tests | 15:02 |
@wiking | are in the data submodule | 15:02 |
@wiking | just like in case of the meta examples | 15:02 |
sanuj | okay, let me see | 15:03 |
sanuj | wiking, so it's in data/testsuite/python2-tests | 15:25 |
sanuj | how is the data generated for the very first time | 15:25 |
sanuj | or for my tags modular test, shall i just test that it compiles? | 15:26 |
@wiking | what do you mean generated for the first time? | 15:27 |
@wiking | i mean i'm not so sure if i can follow what you mean here | 15:27 |
@wiking | for python_modular tests the data residing in data submodule is being used | 15:27 |
@wiking | see for example | 15:27 |
@wiking | ~/shogun/examples/undocumented/python_modular/classifier_multiclassliblinear_modular.py | 15:28 |
sanuj | wiking, what i mean is, suppose i just added a test called tags_params_modular.py which has a function which returns some data | 15:29 |
@wiking | doesn't need to return data | 15:29 |
@wiking | it just should run an exit 0 | 15:29 |
sanuj | okay | 15:29 |
@wiking | i mean in tags cars | 15:29 |
@wiking | i mean in tags case | 15:29 |
@wiking | you just wanna create an object | 15:29 |
@wiking | where you do .sets() and .gets() | 15:29 |
@wiking | check that what you set is the same that you get | 15:30 |
sanuj | wiking, yep, so no worrying about data then | 15:30 |
@wiking | and texit | 15:30 |
@wiking | *exit | 15:30 |
@wiking | you just want to test the tags part | 15:30 |
@wiking | and for that you dont really need that | 15:30 |
@wiking | *data | 15:30 |
@wiking | just to be able to use the tags backend from modular interface | 15:30 |
sanuj | wiking, so it will only fail if it doesn't compile? | 15:32 |
sanuj | if i check for equality in the test, and if it's not equal | 15:32 |
sanuj | then how to make it fail? | 15:32 |
@wiking | no | 15:40 |
@wiking | it will fail | 15:40 |
@wiking | if there's a problem with the generated interface | 15:40 |
-!- sanuj [~sanuj@117.203.8.128] has quit [Ping timeout: 244 seconds] | 15:41 | |
-!- c4goldsw [5da420e6@gateway/web/cgi-irc/kiwiirc.com/ip.93.164.32.230] has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client] | 16:04 | |
-!- sanuj [~sanuj@117.203.8.128] has joined #shogun | 18:37 | |
-!- lambday [6a3311c4@gateway/web/freenode/ip.106.51.17.196] has quit [Ping timeout: 250 seconds] | 18:44 | |
-!- sanuj [~sanuj@117.203.8.128] has quit [Remote host closed the connection] | 21:07 | |
--- Log closed Thu Aug 04 00:00:51 2016 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!