https://sanzo.top/Default/pybind11
使用pybind11可以将C++库绑定到Python,即在Python内调用C++代码,同样也可以在C++中调用Python代码。
# install pytest
pip install pytest
# install pybind11
git clone https://github.com/pybind/pybind11.git
cd pybind11 && mkdir build && cd build
cmake .. && make && sudo make install
接下来使用pybind11演示下如何实现C++调用Python,以及C++调用Python。
#include <pybind11/pybind11.h>
/***********************调用普通函数***********************/
template <typename T>
T add(T a, T b) {
return a + b;
}
PYBIND11_MODULE(pyadd, m) {
m.doc() = "test for add()";
m.def("add", &add<int>, "add two number.");
m.def("add", &add<double>, "add two number.");
m.def("add", &add<long long>, "add two number.");
m.attr("__version__") = "dev";
/**
* >>> import pyadd
* >>> pyadd.__version__
* 'dev'
* >>> pyadd.add(1.1, 2.2)
* 3.3000000000000003
*/
}
/***********************调用类***********************/
namespace test_class {
class Person {
public:
Person() {
name = "Sanzo";
age = 21;
}
Person(std::string name, int age) {
this->name = name;
this->age = age;
}
std::string getName() {
return this->name;
}
void setName(std::string name) {
this->name = name;
}
int getAge() {
return age;
}
std::string name;
private:
int age;
};
}
PYBIND11_MODULE(test_class, m) {
m.doc() = "test_class::person";
pybind11::class_<test_class::Person>(m, "Person")
.def(pybind11::init())
.def(pybind11::init<std::string, int>())
.def("getAge", &test_class::Person::getAge)
.def("getName", &test_class::Person::getName)
.def("setName", &test_class::Person::setName)
.def_readwrite("name", &test_class::Person::name)
;
/*
>>> a = test_class.Person()
>>> a.getName()
'Sanzo'
>>> a.getAge()
21
>>> a.setName("Sanzo00")
>>> a.getName()
'Sanzo00'
>>> b = test_class.Person("Sazo00", 21)
>>> b.getName()
'Sazo00'
>>> b.getAge()
21
*/
}
/***********************调用stl***********************/
#include <pybind11/stl.h>
void printV(std::vector<int> &v) {
for (auto &it : v) {
std::cout << it << " ";
}
std::cout << std::endl;
}
PYBIND11_MODULE(stl, m) {
m.doc() = "test for stl.";
m.def("printV", &printV);
/*
>>> import stl
>>> stl.printV([1, 2, 3, 4])
1 2 3 4
*/
}
// file name: CMakeLists.txt
cmake_minimum_required(VERSION 3.5)
project(test_pybind)
add_subdirectory(pybind11)
pybind11_add_module(pyadd py_call_cpp.cc)
pybind11_add_module(test_class py_call_cpp.cc)
pybind11_add_module(stl py_call_cpp.cc)
// file name: cpp_call_py.ccc
#include <iostream>
#include <pybind11/pybind11.h>
#include <pybind11/embed.h>
namespace py = pybind11;
int main() {
py::scoped_interpreter python;
py::module t = py::module::import("test_add");
auto result = t.attr("add")(1, 2);
int sum = result.cast<int>();
std::cout << "sum = " << sum << std::endl;
return 0;
}
/*
vim ~/.bashrc # 将工作目录添加到环境变量中
export PYTHONPATH=/home/ubuntu/test_pybind:$PATH
*/
// file name: test_add.py
def add(a, b):
print(f"add({a}, {b})")
return a + b
// file name: CMakeLists.txt
cmake_minimum_required(VERSION 3.5)
project(test_pybind)
add_subdirectory(pybind11)
add_executable(cpp_call_py cpp_call_py.cc test_add.py)
target_include_directories(cpp_call_py
PUBLIC
/usr/include/python3.8
)
target_link_libraries(cpp_call_py
PUBLIC
/home/ubuntu/miniconda3/lib/libpython3.8.so
)
pybind11
放到工作目录下PYTHONPATH
中,这样python就可以找到自定义的包。