Cython Tutorial: How to Use shared

Overview

Wrap the C ++ code in cython so that it can be called from python,

I wrote various things like that, This time, I will explain a little about how to write shared_ptr in cython.

I have a little habit, but once I get used to it, it's okay.

Code structure

As usual,

--Classes written in C ++ --Functions written in C ++

By wrapping with cython, it will be configured so that it can be called from python.

Here, as an example

--C ++ has a class called Test_CPP. `` --C ++ Test_CPP has a method called test_function`

I will say.

Actual code

It's probably easier to understand by looking at the code than it is explained. Py_Test_CPP is the class that wraps Test_CPP.

`test.pyx`



from cython.operator cimport dereference as deref
from libcpp.memory cimport shared_ptr, make_shared

cdef class Py_Test_CPP:
    cdef shared_ptr[Test_CPP] ptr

    def __cinit__(self, int arg1):
        self.ptr = make_shared[Test_CPP](arg1)
    
    def test_function(self):
        return deref(self.ptr).test_function()

Import what you need

First, I'm importing because I use dereference, shared_ptr, make_shared.

How to refer to an object on the C ++ side

At this time, it was necessary to hold the pointer of the class object on the C ++ side as a property, This time I write it as cdef shared_ptr [Test_CPP] ptr and keep it as shared_ptr.

constructor

In __cinit __ which is the constructoron thecython side, make_shared toself.ptr = make_shared [Test_CPP](arg1) By using like, self.ptr on the cython side is made a shared pointer.

Call a function

At def test_function (self), I wrap the function test_functionon theC ++ side, In that case, access the entity as deref (self.ptr) and call the function on the C ++ side from there, and it will work.

Summary

I summarized the writing method when using shared_ptr with cython as a memorandum.

This time around here.

end.

Cython Tutorial: How to Use shared_ptr