--- Log opened Sun Apr 06 00:00:10 2014 | ||
-!- diver-in-sky [uid28350@gateway/web/irccloud.com/x-okkuwpmicheahbzm] has joined #shogun | 00:30 | |
-!- kislay [~kislay@117.239.94.112] has joined #shogun | 00:57 | |
lisitsyn1 | diver-in-sky: welcome home ;) | 01:36 |
---|---|---|
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-bnghsadaioaqovax] has joined #shogun | 09:32 | |
-!- kprah [67157e4c@gateway/web/freenode/ip.103.21.126.76] has joined #shogun | 10:09 | |
-!- newlido [~quassel@41.218.172.126] has joined #shogun | 11:06 | |
-!- kprah [67157e4c@gateway/web/freenode/ip.103.21.126.76] has quit [Quit: Page closed] | 11:13 | |
-!- kislay [~kislay@117.239.94.112] has quit [Read error: Connection reset by peer] | 11:40 | |
-!- kprah [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 12:00 | |
-!- kislay [~kislay@117.239.94.112] has joined #shogun | 12:02 | |
-!- kprah [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Quit: Page closed] | 12:14 | |
-!- thoralf [~thoralf@91-65-137-212-dynip.superkabel.de] has joined #shogun | 12:38 | |
thoralf | Heyho. | 12:38 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 12:39 | |
shogun-notifier- | shogun: Thoralf Klein :develop * 2f61caf / src/shogun/base/init.cpp: https://github.com/shogun-toolbox/shogun/commit/2f61caf066674b27beb8cfca33adcd6187489a06 | 12:39 |
shogun-notifier- | shogun: Cleaning up protobuf library in exit_shogun(). | 12:39 |
shogun-notifier- | shogun: tklein23 :develop * 11c25d3 / src/shogun/base/init.cpp: https://github.com/shogun-toolbox/shogun/commit/11c25d37eaa73a4b0a9210a6d70da4685cf3bc40 | 12:39 |
shogun-notifier- | shogun: Merge pull request #2107 from tklein23/fix_protobuf_leak | 12:39 |
shogun-notifier- | shogun: | 12:39 |
shogun-notifier- | shogun: Shutting down protobuf library in exit_shogun() to fix memory leak. | 12:39 |
-!- newlido [~quassel@41.218.172.126] has quit [Remote host closed the connection] | 12:50 | |
-!- newlido [~quassel@41.218.172.126] has joined #shogun | 12:53 | |
-!- kprah [67157e4c@gateway/web/freenode/ip.103.21.126.76] has joined #shogun | 12:56 | |
thoralf | kprah: Have a look how we're reading files in the unit tests or examples. | 13:13 |
thoralf | kprah: I think there are plenty of examples. | 13:13 |
thoralf | Do you want your file to be part of the unit tests, i.e. you're committing this file to the repository? | 13:14 |
kprah | yeah. | 13:14 |
kprah | This file was generated in pythn | 13:14 |
kprah | as you see in the issue | 13:14 |
kprah | my fix works locally. | 13:15 |
thoralf | Does the created file exist in the repository already or how should it be created before running your test? | 13:15 |
kprah | thoralf: generating such a matrix from c++ gives no errors. | 13:16 |
kprah | thoralf: I mean it works perfectly. But while reading file generated by this way, | 13:16 |
kprah | thoralf: we get an extra feature. So my fix takes care of both the cases locally. | 13:17 |
kprah | thoralf: By the way, yeah it should be generated before test | 13:17 |
kprah | any particular examples? | 13:19 |
kprah | thoralf? | 13:19 |
thoralf | Where does the error occur? While saving or while loading? | 13:29 |
-!- iglesiasg [~iglesiasg@82.74.224.167] has joined #shogun | 13:29 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 13:30 | |
thoralf | Hey iglesiasg | 13:34 |
@iglesiasg | thoralf, sup | 13:34 |
thoralf | everything sup ;) | 13:35 |
@iglesiasg | haha that's good then :) | 13:36 |
kprah | sorry. | 13:37 |
kprah | Was away for a few minutes | 13:37 |
kprah | Error occurs while loading by ctest after make install step | 13:38 |
thoralf | kprah: No, I mean the error you're working on. | 13:38 |
thoralf | The serialization error. | 13:38 |
thoralf | This error happens while saving or loading? | 13:38 |
kprah | loading | 13:39 |
thoralf | Well, then commit this example file to shogun-data? | 13:39 |
thoralf | So you can load it from inside the test. | 13:39 |
thoralf | Good weather outside. Need to charge my batteries in the sun. See you! | 13:39 |
kprah | ok. What should be the filename then? | 13:40 |
kprah | thoralf? | 13:40 |
thoralf | Don't know. test_input_for_bug_1991.csv? | 13:40 |
thoralf | Away now. | 13:40 |
kprah | ok. Thank you very much. Bye! | 13:41 |
kprah | hello, iglesiasg | 13:41 |
@iglesiasg | hello kprah | 13:41 |
kprah | I was discussing a doubt with thoralf just now. | 13:42 |
kprah | iglesiasg: can you help me? | 13:42 |
@iglesiasg | kprah, let me check the logs | 13:42 |
kprah | iglesiasg: most of it happened in the private chat | 13:43 |
kprah | iglesiasg: can I explain? | 13:43 |
@iglesiasg | kprah, sure. Go ahead please | 13:43 |
kprah | iglesiasg: I am working on issue 1991. | 13:43 |
@iglesiasg | kprah, ok, what have you done so far? | 13:44 |
kprah | I found the bug and fixed it locally. But I am facing a strange problem while loading the test_20k.csv | 13:44 |
kprah | iglesiasg: can you see it? | 13:44 |
-!- thoralf [~thoralf@91-65-137-212-dynip.superkabel.de] has quit [Ping timeout: 255 seconds] | 13:45 | |
@iglesiasg | kprah, I am reading the issue yes, if that's what you mean | 13:45 |
@iglesiasg | kprah, I see you opened a pull request as well | 13:46 |
kprah | iglesiasg: yeah. | 13:46 |
@iglesiasg | kislay, although I don't think we want those csv files in the repository | 13:46 |
@iglesiasg | kprah, oh sorry, I talked to the wrong person :) | 13:47 |
kislay | iglesiasg, :D | 13:47 |
kprah | but the unit tests fail when ctest runs. I checked it in travis logs. The reason is it is unable to load that csv file. | 13:47 |
kprah | iglesiasg: any insights on how to load this file? | 13:47 |
@iglesiasg | kprah, what are the dependencies of CSVFile? Maybe there is something not installed in travis and that's why the error | 13:48 |
kprah | CSVFile depends on three classes, CLineReader, CDelimiterTokenizer and CParser | 13:49 |
kprah | iglesiasg: Well in the unit test I need that csv file generated by np.random.rand(20000, 70) | 13:49 |
kprah | iglesiasg: The fact is when I run the unit test via ./shogun-unit-test | 13:50 |
-!- HeikoS [~heiko@2.248.35.30] has joined #shogun | 13:50 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 13:50 | |
kprah | it runs fine | 13:50 |
@iglesiasg | kprah, I meant dependencies as installed libraries, not other classes | 13:50 |
kprah | It is a core class of shogun only | 13:51 |
kprah | doesn't depend on other libraries | 13:51 |
@iglesiasg | kprah, all right | 13:51 |
@iglesiasg | kprah, have you found out what the exact line where the error arises is | 13:52 |
kprah | iglesiasg: | 13:53 |
kprah | yeah | 13:53 |
kprah | http://pastebin.com/1w20Ubpg | 13:53 |
kprah | iglesiasg: here while loading test_20k.csv, ctest fails. | 13:54 |
kprah | probably because ctest runs from a different directory | 13:54 |
@iglesiasg | kprah, could it be related to the path to the file? | 13:54 |
kprah | Yeah, | 13:55 |
kprah | I thought of using absolute pathnames but that is not a good practice. | 13:55 |
@iglesiasg | kprah, why don't you just generate random data within the test? | 13:55 |
kprah | iglesiasg: That works fine. There is a unit test already present. See CSVFile_unittest.cc -> matrix_int32 | 13:56 |
kprah | iglesiasg: The bug was found days ago by one of the guests who reported it on IRC | 13:57 |
@iglesiasg | kprah, I meant to generate random data in the test, save it into a temporary file, do the test and then remove the file | 13:57 |
kprah | iglesiasg: Yeah. There is a test for that already. It is the test that I just mentioned above | 13:58 |
kprah | and it passes well. | 13:59 |
kprah | it only doesn't pass for the ones generated from numpy , as described in the issue by viktor (wiking) | 13:59 |
kprah | so I thought maybe loading the file directly would be fine | 14:00 |
@iglesiasg | kprah, I don't think the problem is related to creating the file from numpy.. | 14:01 |
kprah | iglesiasg: Well I used the file given by this, and confirmed that there is indeed an extra feature. | 14:03 |
kprah | iglesiasg: But if you do it by the one generated from libshogun | 14:03 |
kprah | it just works fine. | 14:04 |
kprah | Can I post the screenshots? | 14:04 |
kprah | Well actually ..... You are right. THe bug was in CSVFile class itself | 14:05 |
@iglesiasg | kprah, why don't you just try to see if the problem is because of the path to the file? | 14:05 |
kprah | where, in the unit test? | 14:05 |
kprah | iglesiasg? | 14:06 |
@iglesiasg | kprah, run locally the tests exactly as travis does, from the same directory, using the same commands. You should be able to see this in the travis logs | 14:06 |
kprah | Yeah. I did it locally too. | 14:06 |
kprah | it failed while loading the file. | 14:06 |
kprah | but when I make a separate program | 14:06 |
kprah | using int main() and do all my things there it works fine (as it should) | 14:06 |
kprah | iglesiasg? | 14:07 |
@iglesiasg | kprah, I cannot tell you more apart from what I already said :) | 14:08 |
kprah | So the main difficulty that I am facing right now is while loading this file in my unit test.I fixed the bug otherwise, and confirmed it by making a separate program. | 14:08 |
kprah | iglesiasg? | 14:08 |
-!- HeikoS [~heiko@2.248.35.30] has quit [Quit: Leaving.] | 14:30 | |
-!- iglesiasg [~iglesiasg@82.74.224.167] has quit [Quit: Leaving] | 14:35 | |
-!- kprah [67157e4c@gateway/web/freenode/ip.103.21.126.76] has left #shogun [] | 14:57 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 15:39 | |
-!- kislay [~kislay@117.239.94.112] has quit [Read error: Connection reset by peer] | 16:24 | |
-!- sonne|work [~sonnenbu@91-64-78-194-dynip.superkabel.de] has quit [Ping timeout: 255 seconds] | 16:27 | |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has joined #shogun | 16:28 | |
-!- HeikoS [~heiko@90-227-238-158-no15.business.telia.com] has joined #shogun | 16:46 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:46 | |
-!- newlido [~quassel@41.218.172.126] has quit [Read error: Operation timed out] | 16:46 | |
lisitsyn1 | HeikoS: hey | 16:53 |
@HeikoS | lisitsyn1: hej! | 16:54 |
lisitsyn1 | HeikoS: soeren submitted slots before me so we should be ok ;) | 16:54 |
@HeikoS | greeting from a Stockholm's hip coffee place :) | 16:54 |
@HeikoS | ok great well done | 16:54 |
lisitsyn1 | HeikoS: I envy you | 16:54 |
lisitsyn1 | HeikoS: so it sieben to acht | 16:56 |
lisitsyn1 | it is* | 16:56 |
@HeikoS | Its quite nice here, fernando lived in this town for a while here | 16:56 |
@HeikoS | yeah | 16:56 |
@HeikoS | lisitsyn1: thats good right? | 16:56 |
lisitsyn1 | HeikoS: totally | 16:56 |
@HeikoS | and once we have the slots we submit who is accepted etc? | 16:56 |
lisitsyn1 | HeikoS: yes we just mark them | 16:56 |
lisitsyn1 | and that's it | 16:56 |
lisitsyn1 | HeikoS: I'm currently setting up my blog :D | 16:58 |
@HeikoS | lisitsyn1: cool | 16:59 |
@HeikoS | what about? | 16:59 |
lisitsyn1 | well you know | 16:59 |
lisitsyn1 | :D | 16:59 |
lisitsyn1 | HeikoS: mostly some geeky details of programming some stuff | 17:00 |
@HeikoS | cool | 17:00 |
@HeikoS | let me know once its up | 17:00 |
@HeikoS | lisitsyn1: where is the draft for the modular framework stuff? | 17:00 |
lisitsyn1 | HeikoS: non yet | 17:01 |
lisitsyn1 | there is a few unresolved issues | 17:01 |
lisitsyn1 | I need to be sure it is possible to do everything I wanted | 17:01 |
lisitsyn1 | especially swig things | 17:01 |
@HeikoS | well the draft is for discussion right? doesnt need to be perfect :) | 17:02 |
lisitsyn1 | ok ok I was lazy ;) | 17:02 |
@HeikoS | haha | 17:04 |
@HeikoS | lisitsyn1: where is your blog? | 17:04 |
lisitsyn1 | HeikoS: not yet ready | 17:09 |
lisitsyn1 | I spent quite a while setting up theme etc | 17:09 |
lisitsyn1 | and I need to finish two posts I had in mind | 17:10 |
@HeikoS | I see | 17:14 |
@HeikoS | what are you using? | 17:14 |
lisitsyn1 | HeikoS: octopress | 17:15 |
-!- newlido [~quassel@41.237.211.127] has joined #shogun | 17:45 | |
-!- newlido_ [~quassel@41.237.211.127] has joined #shogun | 17:48 | |
-!- HeikoS [~heiko@90-227-238-158-no15.business.telia.com] has quit [Quit: Leaving.] | 18:02 | |
-!- thoralf [~thoralf@91.65.138.123] has joined #shogun | 18:54 | |
thoralf | Hey. | 18:54 |
-!- thoralf [~thoralf@91.65.138.123] has quit [Read error: Operation timed out] | 19:13 | |
-!- kislay [~kislay@210.212.61.185] has joined #shogun | 19:26 | |
-!- kislay [~kislay@210.212.61.185] has quit [Read error: Connection reset by peer] | 19:30 | |
-!- khalednasr [~k.nasr92@41.69.180.184] has joined #shogun | 19:35 | |
-!- thoralf [~thoralf@91-65-138-123-dynip.superkabel.de] has joined #shogun | 19:37 | |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has quit [Ping timeout: 240 seconds] | 19:42 | |
-!- sonne|work [~sonnenbu@91-64-78-194-dynip.superkabel.de] has joined #shogun | 19:43 | |
-!- kislay [~kislay@14.139.41.150] has joined #shogun | 19:51 | |
-!- kislay [~kislay@14.139.41.150] has quit [Read error: Connection reset by peer] | 19:52 | |
-!- kislay [~kislay@14.139.41.150] has joined #shogun | 19:55 | |
-!- kislay [~kislay@14.139.41.150] has quit [Read error: Connection reset by peer] | 19:56 | |
-!- khalednasr [~k.nasr92@41.69.180.184] has quit [Quit: Leaving] | 20:14 | |
-!- rho_ [~quassel@217.55.243.29] has joined #shogun | 22:07 | |
-!- kprah [67157e4c@gateway/web/freenode/ip.103.21.126.76] has joined #shogun | 22:26 | |
thoralf | Hey kprah | 22:29 |
kprah | hey | 22:29 |
kprah | can we get the shogun root directory name from SGIO? | 22:30 |
kprah | or some other file, thoralf? | 22:30 |
kprah | You are from? | 22:32 |
thoralf | kprah: If you have a minimal file, then it should be possible to generate it automatically from the unit test. So no need for getting the path. | 22:32 |
thoralf | kprah: Do you know what the *minimal* parameters for this problem are? | 22:32 |
kprah | thoralf: I think just a csv file with some data. | 22:33 |
kprah | thoralf: Well if you see the already present unit test in tests/unit/io, there is a test already which passes. | 22:34 |
thoralf | The macro __FILE__ should be expanded to the name of the unit-test. | 22:35 |
thoralf | __FILE__ ".testdata" could be used as the filename for your data. | 22:35 |
thoralf | But still, I insist that you create some *minimal* example. | 22:36 |
thoralf | 12 MB example of such file is not acceptable. | 22:36 |
kprah | ok ..... I guess I'll have to make the whole library again , since python modular is turned off :D | 22:37 |
kprah | thoralf: can I get back to you after some minutes? | 22:37 |
thoralf | Wait. | 22:37 |
thoralf | No need to rebuild it. | 22:37 |
kprah | Then, how do I turn python modular on? | 22:38 |
kprah | thoralf? | 22:38 |
thoralf | No, try to analyze the problem. | 22:39 |
thoralf | You should be able to do it without running python again. | 22:39 |
thoralf | Just look to the file - what caused the problem? | 22:39 |
thoralf | If you run the (old, unfixed) code on the same file, how many features/vectors do you get? | 22:40 |
kprah | 20001 | 22:40 |
thoralf | Okay. And if you remove the trailing newline, how many will you get then? | 22:41 |
thoralf | Only the newline. All data stays the same. | 22:41 |
kprah | ok let me try that | 22:41 |
kprah | There is no trailing newline here I guess. I am opening it in gedit. | 22:42 |
-!- sonne|work [~sonnenbu@91-64-78-194-dynip.superkabel.de] has quit [Ping timeout: 240 seconds] | 22:42 | |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has joined #shogun | 22:43 | |
kprah | thoralf, but for any given file how do we know what are the insides beforehand? | 22:43 |
thoralf | Please answer my question first. | 22:44 |
thoralf | If you remove the trailing newline (there is one) will the output is correct then? | 22:44 |
kprah | How do I do it, manually or by code? I am unable to see a newline after the last line. | 22:46 |
kprah | thoralf? | 22:46 |
thoralf | Just edit the code with your editor. | 22:47 |
thoralf | Sorry, the data file. | 22:47 |
-!- Netsplit *.net <-> *.split quits: lisitsyn1 | 22:49 | |
-!- sonne|osx [~sonne@f053038174.adsl.alicedsl.de] has joined #shogun | 22:49 | |
kprah | yeah, same | 22:54 |
kprah | thoralf, 20001 | 22:54 |
thoralf | Okay, and what if you create a file that has exactly one line "0.12345" and that's all? | 22:55 |
thoralf | No more lines, no more features. | 22:55 |
kprah | ok let me see. | 22:55 |
-!- lisitsyn [~lisitsyn@109-226-124-25.clients.tlt.100megabit.ru] has joined #shogun | 22:57 | |
kprah | features 1 | 22:57 |
kprah | vectors 1 | 22:57 |
kprah | this I created manually | 22:57 |
kprah | thoralf | 22:57 |
thoralf | Okay, one second. | 22:58 |
kprah | okay | 23:01 |
kprah | any thoughts? | 23:10 |
kprah | rhoralf: sorry but can you post any advice on the mailing list, its getting late (about 3am). | 23:11 |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has quit [Ping timeout: 255 seconds] | 23:12 | |
kprah | Thanks. Bye. | 23:12 |
-!- kprah [67157e4c@gateway/web/freenode/ip.103.21.126.76] has quit [Quit: Page closed] | 23:12 | |
-!- sonne|work [~sonnenbu@91-64-78-194-dynip.superkabel.de] has joined #shogun | 23:12 | |
-!- newlido_ [~quassel@41.237.211.127] has quit [Remote host closed the connection] | 23:13 | |
-!- newlido [~quassel@41.237.211.127] has quit [Read error: Connection reset by peer] | 23:14 | |
lisitsyn | thoralf: | 23:23 |
lisitsyn | hey | 23:23 |
thoralf | Hey lisitsyn | 23:23 |
-!- khalednasr [~k.nasr92@41.69.190.77] has joined #shogun | 23:31 | |
--- Log closed Mon Apr 07 00:00:11 2014 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!