00:51 < zoq> In case you have 7 minutes to spend: https://www.youtube.com/watch?v=tsG95Y-C14k

01:49 < sreenik> That video was a great one! Nice thing to start my day with :)

03:05 < rcurtin> that was highly entertaining :)

03:06 < rcurtin> it reminded me of my favorite technical talk of all time: https://www.destroyallsoftware.com/talks/wat

10:06 < janek_> hey, I'm having a bunch of issues installing mlpack on macOS, would you guys be willing to help? should be easy enough

10:50 < zoq> janek_: If you can post some error message, we can see what we can do.

11:15 < Viserion> After I fork the repository and do some changes how could I effect my changes locally?

11:16 < janek_> zoq: please see this for how it went with trying a pip installation (it's brief, I hope) - https://github.com/mlpack/mlpack/issues/1795

11:18 < janek_> and with building from source, when I do `cmake ../` - -- Configuring incomplete, errors occurred! See also "/Users/janek/Developer/mlpack-3.0.4/build/CMakeFiles/CMakeOutput.log".

11:18 < janek_> and then in the log there's nothing about an error, only one deprecation warning

11:19 < janek_> but if I try to go to the next step (make) - I get make: *** No targets specified and no makefile found. Stop.

11:20 < zoq> The python bindings are also available via pip for manylinux platforms, not sure mac OS is included. But the build from source works just fine.

11:20 < zoq> can you post the direct output for cmake?

11:20 < zoq> probably some dependency is missing

11:21 < zoq> btw. nice terminal style

11:23 < janek_> thanks! it's mostly from here https://gist.github.com/kevin-smets/8568070. iTerm is a terminal for mac, but zsh is crossplatform, so I'm sure you can use some of it

11:24 < janek_> I'll double check my dependencies, try again and come back with cmake output if it doesn't work

11:25 < Viserion> After I fork the repository and do some changes how could I effect my changes locally?

11:30 < zoq> Viserion: What do you mean with effect?

11:30 < zoq> janek_: Using iTerm and zsh as well :)

11:31 < janek_> ah ok, nice ;)

11:32 < Viserion> I mean I changed few functions so would I have to again execute make command?

11:33 < zoq> Viserion: Yeah, make should only build the changes.

11:33 < Viserion> Yes when I do it I get error: In file included from /home/viserion/MLPACk/mlpack-master/src/mlpack/methods/cf/cf_model.hpp:183:0, from /home/viserion/MLPACk/mlpack-master/src/mlpack/methods/cf/cf_main.cpp:18: /home/viserion/MLPACk/mlpack-master/src/mlpack/methods/cf/cf_model_impl.hpp:29:7: error: prototype for ‘void* mlpack::cf::GetValueVisitor::operator()(mlpack::cf::CFType<DecompositionPolicy, NormalizationTyp

11:34 < zoq> janek_: Recently switch the font to Fira Code, I think you use monaco?

11:35 < zoq> Viserion: The error got cut off.

11:35 < janek_> yeah pretty sure I use monaco, saw the Fira Code repo and wanted to try it though :) can I see how it looks in yours?

11:36 < Viserion> class ‘mlpack::cf::GetValueVisitor’ void* GetValueVisitor::operator()(CFType<DecompositionPolicy, NormalizationType>* c) const....remaing part

11:39 < zoq> Viserion: hm, any chance you can push the code to your fork? Difficult to point something out, you can also change smaller parts and see what changes breaks the code.

11:44 < zoq> janek_: http://kurg.org/media/terminal.png

11:46 < janek_> oh sweeet. would love to make my vim look like this. is it nvim+plugins?

11:47 < janek_> installation worked, btw. it was the dependencies. So I'm at fault here for not checking them, I got mislead by the fact that I was warned about missing one

11:47 < janek_> so then I installed it and I assumed I would be warned about missing others, if applicable, but I wasn't

11:48 < janek_> anyway, thanks a lot, I'll hang out here for a bit longer if you don't mind and try to get it to run with python before I leave ;)

11:48 < janek_> how many contributors do you guys have btw?

11:59 < zoq> janek_: Right nvim + plugins

12:02 < zoq> janek_: https://github.com/mlpack/mlpack/graphs/contributors should give you a pretty good impression, I think Ryan mentioned we have contributors from almost every continent, only Antarctica is missing.

12:02 < zoq> So if you know someone from Antarctica, let us know :)

12:03 < janek_> haha I'll keep an eye out

12:03 < zoq> Ahh, here is the mail: http://knife.lugatgt.org/pipermail/mlpack/2019-March/003977.html

12:03 < janek_> would it be easy for you to share what plugins you have? I haven't gottent into that yet but would like to

12:09 < zoq> I think it would be easier if I push my zsh and vim config, I'll see if I have a chance to do this later today.

12:18 < zoq> rcurtin: Agreed, great talk :)

12:25 < Hemal> Are there any graph based algorithms implemented in mlpack?

12:38 < janek_> zoq : thanks, that would be very cool. but no stress about that, I'll do my own diggin at some point and I'm sure there's staff out there

12:45 < janek_> I would really appreciate some help with this last step, i.e. getting python keybindings to work

12:45 < janek_> 1. I can verify that mlpack is installed by writing mlpack_knn in terminal

12:46 < janek_> 2. I did export LD_LIBRARY_PATH=/usr/local/lib/ and when I echo $LD_LIBRARY_PATH it does indeed point there

12:46 < janek_> 3. There are indeed mlpack files there, such as mlpack_knn

12:48 < janek_> 4. but if I try `import mlpack` or `from mlpack import knn` in python, I get an import error

12:48 < janek_> I'm following this http://www.mlpack.org/docs/mlpack-3.0.4/doxygen/python_quickstart.html and there is nothing more there

13:15 < rcurtin> janek_: saw your bug report, actually the pip package name is 'mlpack3': https://pypi.org/project/mlpack3/

13:15 < rcurtin> somebody had already claimed the name mlpack and we have not been able to recover it yet even though the issue has been open for many months at this point

13:16 < rcurtin> I'll respond to the issue you opened shortly, but I think 'pip install mlpack3' should work

13:21 < rcurtin> oh wait, you said OS X. nevermind... I think we only have them available for manylinux

13:22 < rcurtin> I don't know how to package them for OS X

13:24 < rrs_> hey Ryan, thanks for responding! I also found the package on PyPI, that's why I was confused it doesn't work. I was not aware installing packages through pip can be platform-specific

13:25 < rcurtin> rrs_: yeah, I was surprised by that too, but at the end of the day the mlpack bindings do depend on the platform-specific mlpack libraries

13:25 < rcurtin> I should probably open an issue on Github about OS X versions of the Python bindings

13:25 < rcurtin> (and Windows too)

13:25 < rrs_> I'll edit my issue to be clear that it's macOS, if it isn't not

13:25 < rrs_> now*

13:25 < rcurtin> anyway, I have to run for now, but I'll be back in a little while and respond to the issue and also try and help with the build

13:26 < rcurtin> it may have already said that and I may have overlooked it :)

13:26 < rrs_> I got logged out and may have missed some messages because of that, but did anyone respond to getting python bidnings to work once it's insalled?

13:26 < rrs_> ok, thanks so much for writing!

13:32 < rrs_> I'll check back here from time to time, hope to talk to you later :)

13:39 < sreenik> rrs_: You can check the logs if you think you've missed something http://www.mlpack.org/irc/

13:44 < rrs_> thanks, yeah. sorry for missing that, I realized after I wrote :)

14:04 < rrs_> i dug into the issues a bit more and this comment from Ryan solved it for me:

14:04 < rrs_> https://github.com/mlpack/mlpack/issues/1760#issuecomment-469839621

14:05 < rrs_> but it only works for python 2, in 3.7 I get an error

14:05 < rrs_> is mlpack compatible with 3.7 or just 3.6?

14:06 < rcurtin> it should work for 3.7 too, but the key would be whether CMake configured the build to use python2 or python3

14:06 < rcurtin> you can configure with -DPYTHON_EXECUTABLE=/path/to/preferred/python

14:06 < rcurtin> i.e. -DPYTHON_EXECUTABLE=/usr/bin/python3 for instance should work and build for python3

14:07 < rrs_> understand. is there a way of checking it now, or should I reinstall?

14:07 < rrs_> or does the fact itself that it works with 2 and not 3 mean that it's this?

14:07 < rrs_> https://paste.ee/p/YOZM9 this is the error I get in 3.7, in case that's relevant

14:15 < rcurtin> you can take a look at the output from CMake if you have it, or you can look through CMakeCache.txt for PYTHON_EXECUTABLE and see what it is

14:15 < rcurtin> but I'd bet that it's configured for Python 2 and not Python 3 and that's the source of the problem

14:16 < rrs_> ok started reinstalling just this moment anyway ;d

14:30 < rcurtin> rrs_: thanks for the comment on pypa/warehouse#5365

14:30 < rcurtin> I was disappointed to see, last time I looked at that repository, that it is primarily sadness and desperation and, when I looked at it last, nobody's package had been successfully claimed at all

14:40 < rrs_> no prob. and ouch, that doesn't sound good :/

15:07 < rrs_> yay, it works! thanks a lot for the help rcurtin, zoq

15:12 < rcurtin> sure, glad you got it working :)

17:18 < Toshal> zoq lozhnikov ShikharJ rcurtin:Hi sir, I am planning to implement StackGAN, ConditionalGan and if possible LAPGAN this summer. What are your thoughts about it? Also, are they implementable in the given time?

17:27 < atulim> @zoq @rcurtin is there a support for sigmoid kernal? If not, wouldn't it be a nice idea to include it.

17:29 < Viserion> @zoq. here's my changed repo: https://github.com/Zupiter/mlpack pls have a look at it that why it is giving the error

19:23 < Viserion> i am not able to remove the bug...after changing functions for building support for normalization when executing make command for building.

19:23 < Viserion> Error: In file included from /home/viserion/MLPACK/src/mlpack/methods/cf/cf_model.hpp:183:0, from /home/viserion/MLPACK/src/mlpack/methods/cf/cf_main.cpp:18, from /home/viserion/MLPACK/build/src/mlpack/bindings/python/generate_pyx_cf.cpp:34: /home/viserion/MLPACK/src/mlpack/methods/cf/cf_model_impl.hpp:29:7: error: prototype for ‘void* mlpack::cf::GetValueVisitor::operator()(mlpack::cf::CFType<Dec

19:24 < Viserion> ‘void* mlpack::cf::GetValueVisitor::operator()(mlpack::cf::CFType<DecompositionPolicy, NormalizationType>*) const’ does not match any in class ‘mlpack::cf::GetValueVisitor’ void* GetValueVisitor::operator()(CFType<DecompositionPolicy, NormalizationType

19:24 < Viserion> Github repo: https://github.com/Zupiter/mlpack

19:56 < pranit> will the projects based in mlpack bestrictly based in cpp or we have projects in python as well?

21:51 < rcurtin> pranit: mlpack is a C++ library...

22:20 < Suryo> Guys, I hope that there isn't a word or page limit for the GSoC proposals?

22:34 < Viserion> s

22:35 < Viserion> In file included from /home/viserion/MLPACK/build/src/mlpack/bindings/python/generate_pyx_cf.cpp:34:0: /home/viserion/MLPACK/src/mlpack/methods/cf/cf_main.cpp: In function ‘void AssembleFactorizerType(const string&, const string&, arma::mat&, size_t)’: /home/viserion/MLPACK/src/mlpack/methods/cf/cf_main.cpp:399:95: error: no matching function for call to ‘PerformAction(arma::mat&, const size_t&, const size_t&, const double&)â€

22:35 < Viserion> PerformAction<NMFPolicy, CombinedNormalization>(dataset, rank, maxIterations, minResidue);

22:36 < Viserion> Why did this error pop up and not before adding Normalization template

23:52 < zoq> Viserion: Did you fix the first issue? It looks like in your header file you used a void pointer but not in the implementation file.

23:55 < zoq> Toshal: I don't think it's feasible to impelement/test and document all three over the summer. So my recommendation is to focus on one or two. Maybe you can think of a second one that builds on the first one or on something we already have.