File	Date	Author	Commit
example	2015-11-09	Thomas Natschlaeger	[5ee42e] Fixed linking problem on Windows
include	2015-11-08	Thomas Natschlaeger	[6b2fd1] - Some fixes for new version of numpy.i
test	2015-11-09	Thomas Natschlaeger	[5ee42e] Fixed linking problem on Windows
.gitignore	2014-02-09	Thomas Natschlaeger	[b26312] 1) some bugfixes under linux
CHANGELOG	2015-11-09	Thomas Natschlaeger	[5ee42e] Fixed linking problem on Windows
CMakeLists.txt	2015-11-09	Thomas Natschlaeger	[5ee42e] Fixed linking problem on Windows
LICENSE	2012-10-30	Thomas Natschlger	[15b705] added copyright and license terms
README	2015-11-09	Thomas Natschlaeger	[79ed6e] Fixed a typo

Read Me

What is ArmaNpy?
----------------

ArmaNpy is a set of SWIG interface files which allows generating
Python bindings to C++ code which uses the Armadillo matrix library.
From within Python any Armadillo matrices are represented as NumPy matrices.
This is possible due to the same memory layout used.
Copying of memory is avoided whenever possible.
It also supports boost::shared_ptr wrapped return values of Armadillo matrices.

Usage
-----

For general information on how to use SWIG to generate python bindings see
http://www.swig.org/Doc2.0/Python.html.

See the file example/example.i and test/armanpy_test_lib.i how to use ArmaNpy
to generate bindings for a set of classes defined in example/example.hpp and
test/armanpy_test_lib.hpp.

See example/example.py and test/armanpy_test_run.py for the Python side of
usage (note that armanpy_test_run.py is just unit tests).

How to compile and run the example and the tests?
-------------------------------------------------

1) Use CMake to generate your build system based on the provided CMakeLists.txt file
2) Build the resulting project.
3) Run the Python script example.py or armanpy_tests.py

From the command line this can be done as follows for the example dir (adapt to your path settings and use '\' instedead of '/' on Windows). The same holds for the test dir.

cd /path/to/armanpy
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release -DARMADILLO_INCLUDE_DIR=/path/to/armadillo/include -DARMADILLO_LIBRARIES=/path/to/armadillo/library

# e.g. under Linux: 
#   cmake .. -DCMAKE_BUILD_TYPE=Release -DARMADILLO_INCLUDE_DIR=/usr/include -DARMADILLO_LIBRARIES=/usr/lib64/libarmadillo.so
# e.g. under Windows: 
#   cmake ../git -DCMAKE_BUILD_TYPE=Release -G "Visual Studio 12 2013 Win64"  -DARMADILLO_INCLUDE_DIR=C:\Users\natschlaeger\Programs\armadillo-6.200.2\include -DARMADILLO_LIBRARIES=C:/Users/natschlaeger/Programs/OpenBLAS-v0.2.15-Win64-int32/lib/libopenblas.lib

cmake --build . --config Release

The above commands will generate all necessary files in build/bin.

To run the tests and examples either run

ctest 

or 

cd bin
python example.py
python armanpy_test_run.py


*Note* : -DCMAKE_BUILD_TYPE=Release is used bcs on Windows a typical Python distribution does not
come with the debug version of the Python libraries. Hence, the build would fail for the Debug
configuration.

*Remark* : To use 64bit under Windows add e.g. -G "Visual Studio 11 Win64" to the first cmake command.

Credits
-------

The idea of wrapping armadillo matrices py NumPy arrays was taken from Michael Forbes`
https://bitbucket.org/mforbes/swig_cpp_python/src/cd7cef820273?at=default.

Copyright
---------

Copyright (C) 2012-2014 Thomas Natschläger (thomas.natschlaeger@gmail.com)

License
-------

ArmaNpy is provided without any warranty of fitness
for any purpose. You can redistribute this file
and/or modify it under the terms of the GNU
Lesser General Public License (LGPL) as published
by the Free Software Foundation, either version 3
of the License or (at your option) any later version.
(see http://www.opensource.org/licenses for more info)