How to build libraries in C++ and use them in Python

In this post I will expose one way to create libraries in C++ and expose them to Python.

This can be very useful when we want to integrate existing piece of code, or write new code that perform better than Python.

The way exposed in this post will make use of Boost, a peer reviewed C++ set of libraries. In paticular, we will use Boost-Python

Installing Boost

The first step is to install the package Boost:

wget https://dl.bintray.com/boostorg/release/1.75.0/source/boost_1_75_0.tar.gz
tar -xzf boost_1_75_0.tar.gz

We can try if everything is working correctly, with a simple program with Boost (copied from the original documentation)

#include <boost/lambda/lambda.hpp>
#include <iostream>
#include <iterator>
#include <algorithm>

int main()
{
    using namespace boost::lambda;
    typedef std::istream_iterator<int> in;

    std::for_each(
        in(std::cin), in(), std::cout << (_1 * 3) << " " );
}

And we compile

c++ -I ./boost_1_75_0 test-boost.cc -o test-boost

We can see that it works if we get the following output running the program

luis@BARE:~/projects/How to build libraries in CPP to use in Python$ echo 1 2 3 | ./test-boost 
3 6 9 
luis@BARE:~/projects/How to build libraries in CPP to use in Python$

Installing Boost-Python

First let’s enter into the boost directory boost_1_75_0 and we can run ./bootstrap.sh --show-libraries to see the non-header only libraries provided by Boost.

We should see something similar to this at the end

The Boost libraries requiring separate building and installation are:
    - atomic
    - chrono
    - container
    - context
    - contract
    - coroutine
    - date_time
    - exception
    - fiber
    - filesystem
    - graph
    - graph_parallel
    - headers
    - iostreams
    - json
    - locale
    - log
    - math
    - mpi
    - nowide
    - program_options
    - python
    - random
    - regex
    - serialization
    - stacktrace
    - system
    - test
    - thread
    - timer
    - type_erasure
    - wave

Now we can install Boost-Python prepare the environment to use Python:

./bootstrap.sh --with-libraries=python

And with the b2 tool generated while running bootstrap.sh, we can compile and install Boost-Python

sudo ./b2 install

If you ignore the step of specifying the libraries to build, the full build with take a long time….

Example

First, let’s copy the content of the tutorial included with Boost into the folder example-1

cp -Rf boost_1_75_0/libs/python/example/tutorial/* example-1/

We should have three files in the folder

luis@BARE:~/projects/How to build libraries in CPP to use in Python/example-1$ ls
hello.cpp  hello.py  Jamfile
luis@BARE:~/projects/How to build libraries in CPP to use in Python/example-1$

Now we execute bjam and next we execute python hello.py. If everything went well, we should see hello world! as the output.

In the future I will write a post and how to implement a library to operate with complex numbers in C++ and its use in Python