Code in C++, use in Python: tutorial needed [sf#40] about pagmo

[darioizzo]

Originally posted by mmarcusx on 2013-11-25 15:44:27.

I have already written a tutorial on that:

https://sourceforge.net/apps/mediawiki/pagmo/index.php?title=Developing_guide

it could be improved though...

[darioizzo]

Originally posted by darioizzo on 2013-11-25 15:55:19.

The tutorial you refer to is for developers. It shows how a developer can put in PaGMO a new problem / algorithm that is to be shipped in the next release.

Since v 1.1.5 we allow the option to install header files. The tutorial I am asking should show how a normal user (not touching/compiling the pagmo code) can still use libpagmo and PyGMO to code its own problems in c++ and have them visible in Python. You can see the attached files for an example.

[darioizzo]

Originally posted by darioizzo on 2013-11-28 16:51:08.

And here is a new updated version of the scripts that are more robust to porting

[darioizzo]

The files have been moved to github gist (https://gist.github.com/darioizzo/9215291)

[neduard]

Hello, I've started writing a tutorial at https://github.com/neduard/pagmo/wiki/PaGMO-C---Tutorial

However, I have a couple of questions:

1. What should be the format? Currently it's written as a github Markdown wiki page in line with the project's other wiki pages. Is this acceptable?
2. I am unable to load my_pygmo_probs into Python:
   Using the compile command from compile.script, I get:

$ g++ -shared -o my_pygmo_probs.so my_pygmo_probs.o prob_A.o  -std=c++11 -lboost_python-mt -pthread -DNDEBUG -lpython2.7 -lboost_serialization-mt -lpagmo
/usr/lib/gcc/x86_64-pc-linux-gnu/4.8.4/../../../../x86_64-pc-linux-gnu/bin/ld: cannot find -lboost_python-mt
collect2: error: ld returned 1 exit status

I can compile successfully using -lboost_python-2.7-mt flag instead.
However, when I go into Python and type import my_pygmo_probs I get:

>>> import my_pygmo_probs
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
RuntimeError: extension class wrapper for base class pagmo::problem::base has not been created yet

I found https://mail.python.org/pipermail/cplusplus-sig/2009-February/014244.html but changing the position of the -lboost_python-2.7-mt flag didn't seem to help.

I then fiddled with the code in my_pygmo_probs.cpp:
Changing class_<Problem,bases<problem::base> > to class_<Problem> seems to remove the issue allowing to import the module in Python (and allowing me to execute methods such as a_method()), but I am unsure of what is the impact of removing bases<problem::base>.

Any ideas on what I am missing?
My box setup:

Funtoo Linux (Gentoo Based)
boost package: dev-libs/boost-1.56.0-r1000:0/1.56.0  USE="icu nls python threads -c++11 -context -debug -doc -mpi -static-libs -tools" ABI_X86="(64) -32 (-x32)" PYTHON_ABIS="2.7 3.3 -2.6 -3.1 -3.2 -3.4 (-3.5)
Python version 2.7.9

[darioizzo]

Maybe you need to import PyGMO before importing my_pygmo_probs?

[neduard]

@darioizzo Thanks! That did the trick!

I finished writing the rest of the tutorial.
Should I leave it at its current address or should I move it somewhere else so that it can be reviewed and commented upon?

[darioizzo]

The tutorial is now available in the wiki at https://github.com/esa/pagmo/wiki/%5BTUTORIAL%5D-Code-a-problem-in-Cpp,-use-it-in-Python