This is a simple example on how to use boost.python and boost.numpy to call c++ code form python and convert numpy arrays to Eigen.

To run (assuming you've boost.python already installed):

git submodule init
git submodule update
./build.sh      # will build boost.numpy
make
./run.sh hello.py
make test
./run.sh test_eigen_numpy.py

Links :

https://github.com/ndarray/Boost.NumPy

http://eigen.tuxfamily.org/index.php?title=Main_Page

There also is the 'ndarray' project, that aims at providing a multidimensionnal array library similar to numpy.ndarray for C++. Some of the code comes from ndarray : https://github.com/ndarray/ndarray

TODO: - The from python converts do some data copy. This could hurt for big

matrices