pybind11向C++ dll 传递数组 图像

 

python 通过pybind11向C++ dll 传递数组 图像

 

传递python中的List

pybind11 很贴心地帮你把 vector<T> 跟 python 的 list 做好了转换，你只需要 #include <pybind11/stl.h> 即可 [1]

C++端代码

#include <pybind11/pybind11.h>

#include <pybind11/stl.h>

#include <vector>

using std::vector;

namespace py = pybind11;

vector<float> ListMul(vector<float>& in_list, float coef) {

vector<float> ret_vec;

ret_vec.clear();

ret_vec.reserve(in_list.size()); // Requests that the vector capacity be at least enough to contain n elements.

for (size_t i = 0; i < in_list.size(); ++i) {

int v = in_list[i];

ret_vec.push_back(coef * v);

}

return ret_vec;

}

PYBIND11_MODULE(example, m) {

m.doc() = "pass and return a list";

m.def("ListMul", &ListMul, "example function");

}

python端调用

import example

print(example.ListMul([1.0, 2, 3, 4, 5, 6], 5))

 

python端的数据类型有时候不是那么严格，会进行自动转换，但是python float到c的int会出错，

类似地还有map<Tk, Tv>在 python 里对应的就是 Dict[Tk, Tv] [1]。

传递Numpy 数组

两个数组相加的案例 [2]

C++端代码

#include <pybind11/pybind11.h>

#include <pybind11/numpy.h>

 

 

namespace py = pybind11;

 

 

 

py::array_t<double> add_arrays(py::array_t<double> input1, py::array_t<double> input2) {

// read input arrays buffer_info

py::buffer_info buf1 = input1.request(), buf2 = input2.request();

 

if (buf1.size != buf2.size)

throw std::runtime_error("Input shapes must match");

 

// allocate the output buffer

py::array_t<double> result = py::array_t<double>(buf1.size);

 

py::buffer_info buf3 = result.request(); // acquire buffer info

 

double *ptr1 = (double *)buf1.ptr, *ptr2 = (double *)buf2.ptr, *ptr3 = (double *)buf3.ptr;

size_t high = buf1.shape[0];

size_t width = buf1.shape[1];

 

// Add both arrays

for (size_t idy = 0; idy < high; idy++)

{

for (size_t idx = 0; idx < width; idx++)

{

int curIdx = idy*width + idx;

ptr3[curIdx] = ptr1[curIdx] + ptr2[curIdx];

 

}

}

 

/* Reshape result to have same shape as input */

result.resize({ high, width });

 

return result;

}

PYBIND11_MODULE(example, m) {

m.doc() = "Add two vectors using pybind11"; // optional module docstring

m.def("add_arrays", &add_arrays, "Add two NumPy arrays");

}

python端调用

import numpy as np

import example

a = np.ones((10,3))

b = np.ones((10,3)) * 3

c = example.add_arrays(a, b)

print(c)

These numpy array arguments can either be generic py:array or typed py:array_t<double>. The properties of the numpy array can be obtained by calling its request method. This returns a struct of the following form [3]:

struct buffer_info {

void *ptr;

size_t itemsize;

std::string format;

int ndim;

std::vector<size_t> shape;

std::vector<size_t> strides;

};

python传递图像到c++ dll

根据上面的案例，很容易写出C++端代码

1.  
2.  

#include <pybind11/pybind11.h>

#include <pybind11/numpy.h>

 

 

#include <iostream>

using namespace std;

 

namespace py = pybind11;

 

#include <io.h>

#include <fcntl.h>

// display console

void OpenConsole()

{

// create a console

AllocConsole();

 

//

FILE* stream;

freopen_s(&stream, "CON", "r", stdin); //

freopen_s(&stream, "CON", "w", stdout); //

 

SetConsoleTitleA("Information output"); //

 

HANDLE _handleOutput;

_handleOutput = GetStdHandle(STD_OUTPUT_HANDLE);

 

// FreeConsole();

}

 

 

void GetImage(py::array_t<char> input1)

{

py::buffer_info buf1 = input1.request();

OpenConsole();

printf("dim:%d \n", buf1.ndim);

 

for (int cnt = 0; cnt < buf1.ndim; cnt++)

{

printf("dim size %d : %d\n", cnt, buf1.shape[cnt]);

}

unsigned char *ptr1 = (unsigned char *)buf1.ptr;

printf("image data: %d %d %d %d %d %d\n", ptr1[0], ptr1[1], ptr1[2], ptr1[3], ptr1[4], ptr1[5]);

}

 

PYBIND11_MODULE(example, m) {

m.doc() = "Get NumPy arrays"; // optional module docstring

 

m.def("GetImage", &GetImage, "Get NumPy arrays");

}

python端代码

import numpy as np

import example

from PIL import Image #PIL pakage name is Pillow

im = Image.open('leopard-2.png')

in_data = np.asarray(im, dtype=np.uint8)

in_data = np.rollaxis(in_data,2,0)

print(type(in_data))

example.GetImage(in_data)