"Cython" tutorial to make Python explosive: Handling when a function on the C ++ side is passed by reference.

Overview

I want to make the code written in C ++ a library that can be called in Python. However, there was a big wall called Cython.

This is "Cython" tutorial to make Python explosive: when a function on the C ++ side has an overload. This is a continuation of.

The code is listed in "github here", so please take a look.

Last time, I explained Cythonization when a function on the C ++ side is overloaded (when the configuration has the same function name but different arguments).

This time, I would like to explain how to handle a function on the C ++ side on the Cython side when passing it by reference instead of passing it by value as an argument. It's not very new, but when you actually try to write Cython, you often get annoyed by errors, so be sure to check it out! !!

Situation explanation

--Since the function on the C ++ side was overloaded, it was solved using a cast like previous article. --The arguments of the function implemented on the C ++ side include those passed by reference. --If the argument is a reference type, an error will occur.

I will actually try it.

cpp_library/TestClass1.h

class TestClass1{
    private:
        TestClass2 property_test_class2;
        EnumForTestClass1 group;

    public:
        //~abridgement
        static void print_vector(vector<int> &x);
        static void print_vector(vector<double> &x);

The difference from the last time is that the argument `` `xof print_vector``` is changed from ** by value to passed by reference **.

cpp_library/TestClass1.cpp

void TestClass1::print_vector(vector<int> &x){
    for(int i=0; i<x.size(); i++){
        cout << x[i] << " ";
    }
    cout << endl;
}

void TestClass1::print_vector(vector<double> &x){
    for(int i=0; i<x.size(); i++){
        cout << x[i] << " ";
    }
    cout << endl;
}

When I try to build this with python setup.py install as usual, I get the following error:

./cython/my_library.cpp:2414:69: error: cannot bind non-const lvalue reference of type 'std::vector<int>&' to an rvalue of type 'std::vector<int>'
     __pyx_t_1 = __Pyx_void_to_None(__pyx_v_testclass1.print_vector(((std::vector<int> )__pyx_t_4))); if (unlikely(!__pyx_t_1)) __PYX_ERR(0, 53, __pyx_L1_error)

The cast should be neat, but I get angry when it comes to passing by reference.

So, as a solution

Create the entity on the spot, not the cast.

What does that mean?

cython/test_class1.pyx

    @staticmethod
    def print_vector(list x):
        cdef:
            TestClass1 testclass1
            vector[int] vec_int
            vector[double] vec_double
        
        if isinstance(x[0], int):
            vec_int.resize(len(x))
            for i in range(len(x)):
                vec_int[i] = <int>x[i]
            return testclass1.print_vector(vec_int)
        elif isinstance(x[0], float):
            vec_double.resize(len(x))
            for i in range(len(x)):
                vec_double[i] = <double>x[i]
            return testclass1.print_vector(vec_double)
        else:
            raise Exception('TypeError')

will do. Now, with `cdef```, create ``` vector [int] and ``` vector [double] `` on the fly and pass it to the function. Maybe the error I just got is passed by reference, but I don't know if it really exists! Don't worry! !! !! I interpret that it was an error.

Therefore, it is good to prepare the entity and reassure it.

Then, as usual, `python setup.py install` will pass the build.

By doing this, by writing the test code as follows,

test.py

import my_library as myl

if __name__ == "__main__":

    cl1 = myl.test()
    x = [1,2,3]
    y = [1.1, 2.2, 3.3]
    cl1.print_vector(x)
    cl1.print_vector(y)

I was able to call the function properly.

(myenv) root@e96f489c2395:/from_local/cython_practice# python test.py 
1 2 3 
1.1 2.2 3.3 

Summary

This time, I explained how to handle the function on the C ++ side when passing by reference instead of passing by value as an argument on the Cython side.

This time around.

end.