Why pybind11 converts double to int?

Question

I created.pyd file:

#include <pybind11/pybind11.h>
#include <iostream>
#include <typeinfo>
namespace py = pybind11;
int add(int num) {
    float a = 1.0;  
    for (int i = 0; i <= num; i = i + 1) {
        a = (a + i)/a;
    }        
    std::cout << "dll is typing: " << a << '\n';
    std::cout << typeid(a).name() << std::endl;
    return a;
}
PYBIND11_MODULE(py_dll, m) {
    m.doc() = "pybind11 py_dll plugin"; // optional module docstring
    m.def("add", &add, "Add function", py::arg("num"));
}

I call it from python:

import py_dll
num = 500
a = py_dll.add(num)
print ('python is typing: ', a)

It prints:

Why digit became int? I speak about 22. I expect it to be float 22.8722

Answer 1

This function

int add(int num)

takes an int as parameter and returns an int . The "problem" is not with pybind. Try this to see:

int main() {
    auto x = add(42);
    std::cout << "return value: " << x << '\n';
    std::cout << typeid(x).name() << std::endl;
}

If the function should return a float you have to declare that it returns a float :

float add(int num)