特征返回类型的 Pybind11 问题

Question

我正在尝试使用以下方法公开data_all哪个是datapython的向量pybind11：

struct data {

    std::vector<Eigen::ArrayXf> values;

    std::vector<int> indices;

    float x;

    float y;

    float z;

};

class dataBuffer {

public:

    std::vector<data> data_all;

    Eigen::ArrayXf getValues(ssize_t i, ssize_t j) { return data_all.at(i).values.at(j); };

};

我将pybind11包装器定义如下：

PYBIND11_MODULE(example, m) {

    py::class_<data>(m, "data")

        .def(py::init<>())

        .def_readonly("values", &data::values)

        .def_readonly("indices", &data::indices)

        .def_readonly("x", &data::x)

        .def_readonly("y", &data::y)

        .def_readonly("z", &data::z);

    py::class_<dataBuffer>(m, "dataBuffer")

        .def(py::init<>())

        .def("getValues", &dataBuffer::getValues);

}

我的 C++ 示例代码是

namespace py = pybind11;

int main()

{

    data d;

    d.x = 1.1;

    d.y = 2.1;

    d.z = 3.1;

    d.indices.push_back(4);

    d.values.push_back(Eigen::ArrayXf::LinSpaced(50, 0.0, 50 - 1.0));

    d.indices.push_back(5);

    d.values.push_back(Eigen::ArrayXf::LinSpaced(60, 0.0, 60 - 1.0));

    d.indices.push_back(11);

    d.values.push_back(Eigen::ArrayXf::LinSpaced(70, 0.0, 70 - 1.0));

    dataBuffer D;

    D.data_all.push_back(d);

    D.data_all.push_back(d);

    std::cout << D.getValues(0,0) << "\n";

    py::scoped_interpreter guard{};

    py::object result = py::module::import("pybind11_test").attr("testData")(0,0);

}

pybind11_test.py文件的内容

import numpy as np

import example as m

def testData(buffer):

    help(buffer)

    a = buffer.getValues(0,0) # trying to retrieve the data buffer created in C++

    print(a)

help(buffer)打印以下签名：

Help on method getValues in module example:

getValues(...) method of example.dataBuffer instance

    getValues(self: example.dataBuffer, arg0: int, arg1: int) -> Eigen::Array<float,-1,1,0,-1,1>

我相信EigenPython 不喜欢返回类型。有人可以帮我解决问题吗？我应该如何帮助 Python 理解返回类型，getValues以便我可以使用numpy库进一步处理它？

Answer 1

我假设这条线：py::object&nbsp;result&nbsp;=&nbsp;py::module::import("pybind11_test").attr("testData")(0,0);本来是：py::object&nbsp;result&nbsp;=&nbsp;py::module::import("pybind11_test").attr("testData")(D);除此之外，您所缺少的是：#include&nbsp;"pybind11/eigen.h"在包装代码的顶部。