Boost.python Numpy example

Before recent advances such as Cython, Boost.Python was one of the best ways to integrate C++ code into Python programs. The following C++ source code exposes a C++ function to Python. The function takes a NumPy array as an argument and extracts a C++ integer type.

#include <boost/python.hpp>
#include "boost/python/extract.hpp"
#include "boost/python/numeric.hpp"
#include <iostream>
 
using namespace boost::python;
 
// Functions to demonstrate extraction
void setArray(boost::python::numeric::array data) {
 
	// Access a built-in type (an array)
	boost::python::numeric::array a = data;
	// Need to <extract> array elements because their type is unknown
	std::cout << "First array item: " << extract<int>(a[0]) << std::endl;
}
 
 
// Expose classes and methods to Python
BOOST_PYTHON_MODULE(TestNumPy) {
	boost::python::numeric::array::set_module_and_type("numpy", "ndarray");
 
	def("setArray", &setArray);
 
}

2 thoughts on “Boost.python Numpy example

  1. ARAMBAM JAMES SINGH

    I get this error :

    import TestNumpy as tnp
    ImportError: dynamic module does not define init function (initTestNumpy)

    Reply
  2. James

    Hi ,

    I get the error “Segmentation fault: 11”

    when i run the python code after compiling above c++ code

    #————— #

    import TestNumPy as nt
    import numpy as np
    a = np.array([10,20,30])
    nt.setArray(a)

    Reply

Leave a Reply