In order to take advantage of the strength of both languages, developers use Python bindings which allows them to call C/C++ libraries from python.
Now, the question arises that why there is a need for doing this?
- As we know that, C has faster execution speed and to overcome the limitation of Global Interpreter Lock(GIL) in python, python bindings are helpful.
- We have a large, stable and tested library in C/C++, which will be advantageous to use.
- For performing large scale testing of the systems using Python test tools.
Let’s see the C code which we want to execute with Python :
C++
#include <iostream>class Geek{ public: void myFunction(){ std::cout << "Hello Geek!!!" << std::endl; }};int main(){ // Creating an object Geek t; // Calling function t.myFunction(); return 0;} |
We have to provide those cpp declarations as extern “C” because ctypes can only interact with C functions.
C++
extern "C" { Geek* Geek_new(){ return new Geek(); } void Geek_myFunction(Geek* geek){ geek -> myFunction(); }} |
Now, compile this code to the shared library :
Finally, write the python wrapper:
Python3
# import the modulefrom ctypes import cdll # load the librarylib = cdll.LoadLibrary('./libgeek.so') # create a Geek classclass Geek(object): # constructor def __init__(self): # attribute self.obj = lib.Geek_new() # define method def myFunction(self): lib.Geek_myFunction(self.obj) # create a Geek class objectf = Geek() # object method callingf.myFunction() |
Output :
Hello Geek!!!
