传递 boost::p ython::numpy::ndarray 作为 boost::p ython 函数的(默认或非
Passing boost::python::numpy::ndarray as (default or not) argument of a boost::python function?
是否可以将boost：:p ython：：numpy：：ndarray作为boost：:p ython函数的(默认与否(参数传递？
Dummy没有ndarray。 作为一个 ndarray 参数很愚蠢，但没有默认值。 Silly 有一个 ndarray 作为默认值。
>> more dummy.cpp stupid.cpp silly.cpp
::::::::::::::
dummy.cpp
::::::::::::::
#include <boost/python.hpp>
namespace bp = boost::python;
int f(double x, double y=1.0) {return (int)(x+y);};
BOOST_PYTHON_MODULE(dummy)
{
bp::def("f", f, ( bp::arg("x"), bp::arg("y")=1.0 ) );
}
::::::::::::::
stupid.cpp
::::::::::::::
#include <boost/python.hpp>
#include <boost/python/numpy.hpp>
namespace bp = boost::python;
namespace np = boost::python::numpy;
int f(np::ndarray x, double y=1.0) {return (int)(x.shape(0)+y);};
BOOST_PYTHON_MODULE(stupid)
{
bp::def("f", f, ( bp::arg("x"), bp::arg("y")=1.0 ) );
}
::::::::::::::
silly.cpp
::::::::::::::
#include <boost/python.hpp>
#include <boost/python/numpy.hpp>
namespace bp = boost::python;
namespace np = boost::python::numpy;
int f(np::ndarray x, np::ndarray * y=nullptr) {return (int)(y ? x.shape(0)+y->shape(0) : x.shape(0));};
BOOST_PYTHON_MODULE(silly)
{
bp::def("f", f, ( bp::arg("x"), bp::arg("y")=nullptr ) );
}
>> make
g++ -I /usr/include/python2.7 -o dummy.so -fPIC -shared dummy.cpp -lboost_python -lboost_numpy -lpython2.7
g++ -I /usr/include/python2.7 -o stupid.so -fPIC -shared stupid.cpp -lboost_python -lboost_numpy -lpython2.7
g++ -I /usr/include/python2.7 -o silly.so -fPIC -shared silly.cpp -lboost_python -lboost_numpy -lpython2.7
>> python
Python 2.7.17 (default, Oct 19 2019, 23:36:22)
[GCC 9.2.1 20191008] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import dummy; dummy.f(1)
2
>>> import numpy; import stupid; stupid.f(numpy.array([1, 2, 3]))
Segmentation fault
更新
试图在f中添加Py_Initialize(); np::initialize();但没有成功
>> more stupid.cpp silly.cpp
::::::::::::::
stupid.cpp
::::::::::::::
#include <boost/python.hpp>
#include <boost/python/numpy.hpp>
namespace bp = boost::python;
namespace np = boost::python::numpy;
int f(np::ndarray x, double y=1.0) {
Py_Initialize();
np::initialize();
return (int)(x.shape(0)+y);
};
BOOST_PYTHON_MODULE(stupid)
{
bp::def("f", f, ( bp::arg("x"), bp::arg("y")=1.0 ) );
}
::::::::::::::
silly.cpp
::::::::::::::
#include <boost/python.hpp>
#include <boost/python/numpy.hpp>
namespace bp = boost::python;
namespace np = boost::python::numpy;
int f(np::ndarray x, np::ndarray * y=nullptr) {
Py_Initialize();
np::initialize();
return (int)(y ? x.shape(0)+y->shape(0) : x.shape(0));
};
BOOST_PYTHON_MODULE(silly)
{
bp::def("f", f, ( bp::arg("x"), bp::arg("y")=nullptr ) );
}
>> make
g++ -I /usr/include/python2.7 -o stupid.so -fPIC -shared stupid.cpp -lboost_python -lboost_numpy -lpython2.7
g++ -I /usr/include/python2.7 -o silly.so -fPIC -shared silly.cpp -lboost_python -lboost_numpy -lpython2.7
>> python
Python 2.7.17 (default, Oct 19 2019, 23:36:22)
[GCC 9.2.1 20191008] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy; import stupid; stupid.f(numpy.array([1, 2, 3]))
Segmentation fault
更新
好的，让它在BOOST_PYTHON_MODULE的呼叫中工作.仍然带有默认参数的 KO(silly示例(。
>> more stupid.cpp silly.cpp
::::::::::::::
stupid.cpp
::::::::::::::
#include <boost/python.hpp>
#include <boost/python/numpy.hpp>
namespace bp = boost::python;
namespace np = boost::python::numpy;
int f(np::ndarray x, double y=1.0) {
return (int)(x.shape(0)+y);
};
BOOST_PYTHON_MODULE(stupid)
{
Py_Initialize();
np::initialize();
bp::def("f", f, ( bp::arg("x"), bp::arg("y")=1.0 ) );
}
::::::::::::::
silly.cpp
::::::::::::::
#include <boost/python.hpp>
#include <boost/python/numpy.hpp>
namespace bp = boost::python;
namespace np = boost::python::numpy;
int f(np::ndarray x, np::ndarray * y=nullptr) {
return (int)(y ? x.shape(0)+y->shape(0) : x.shape(0));
};
BOOST_PYTHON_MODULE(silly)
{
Py_Initialize();
np::initialize();
bp::def("f", f, ( bp::arg("x"), bp::arg("y")=nullptr ) );
}
>> make
g++ -I /usr/include/python2.7 -o stupid.so -fPIC -shared stupid.cpp -lboost_python -lboost_numpy -lpython2.7
g++ -I /usr/include/python2.7 -o silly.so -fPIC -shared silly.cpp -lboost_python -lboost_numpy -lpython2.7
>> python
Python 2.7.17 (default, Oct 19 2019, 23:36:22)
[GCC 9.2.1 20191008] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy; import stupid; stupid.f(numpy.array([1, 2, 3]))
4
>>> import numpy; import silly; silly.f(numpy.array([1, 2, 3]))
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: No to_python (by-value) converter found for C++ type: decltype(nullptr)
解决方法
>> more silly.cpp
#include <boost/python.hpp>
#include <boost/python/numpy.hpp>
namespace bp = boost::python;
namespace np = boost::python::numpy;
int f(np::ndarray x, bp::object y) {
np::ndarray yy = np::array(bp::list());
if (!y.is_none()) yy = bp::extract<np::ndarray>(y);
return (int)(x.shape(0)+yy.shape(0));
};
BOOST_PYTHON_MODULE(silly)
{
Py_Initialize();
np::initialize();
bp::def("f", f, ( bp::arg("x"), bp::arg("y") ) );
}
>> make
g++ -I /usr/include/python2.7 -o silly.so -fPIC -shared silly.cpp -lboost_python -lboost_numpy -lpython2.7
>> python
Python 2.7.17 (default, Oct 19 2019, 23:36:22)
[GCC 9.2.1 20191008] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy; import silly; silly.f(numpy.array([1, 2, 3]), numpy.array([1, 2]))
5
>>> import numpy; import silly; silly.f(numpy.array([1, 2, 3]), None)
3
要使用numpy首先初始化Python运行时和numpy模块，请执行以下操作：
Py_Initialize();
np::initialize();
未能调用这些会导致分段错误。
====
===========================================================================对于silly.cpp不需要解决方法。以下是实现：
#include <boost/python.hpp>
#include <boost/python/numpy.hpp>
namespace bp = boost::python;
namespace np = boost::python::numpy;
int f(np::ndarray x, np::ndarray y = np::array(bp::list()) ) {
return (int)(x.shape(0)+y.shape(0));
};
BOOST_PYTHON_MODULE(silly)
{
Py_Initialize();
np::initialize();
bp::def("f", f, ( bp::arg("x"), bp::arg("y") = np::array(bp::list()) ) );
}
而测试结果：
>>> import numpy, silly; silly.f(numpy.array([1, 2, 3]), numpy.array([1, 2]))
5
>>> import numpy, silly; silly.f(numpy.array([1, 2, 3]))
3
相关文章：
- 相当于 pybind11 中的 boost::p ython py::scope(）.attr(）
- 传递 boost::p ython::numpy::ndarray 作为 boost::p ython 函数的(默认或非
- boost::p ython:PyErr_Fetch始终返回 NULL 回溯
- 如何通过 boost::p ython 重命名构造函数的关键字参数
- 链接问题 boost::p ython::numpy.
- 在 boost::p ython 中的 python 对象中运行 python 代码
- 你如何在 boost：:p ython 中"from __future__ import division"？
- 使用 boost::p ython 将数据缓冲区放入C++中
- 如何返回包含 PyObject* 的 boost::p ython::元组?
- 如何正确访问结构值并将其传递给函数 - boost::p ython
- Boost：:P ython，将元组转换为Python有效，矢量<tuple>不
- 如何使用 boost::p ython 创建和使用 Python 对象的实例
- 使用 boost::p ython，如何将结构体向量作为字典列表返回给 Python
- 使用 boost::p ython 从C++创建 python collections.namedtuple
- 如何使用 boost::p ython 从 Python 类型中提取包装的C++类型
- Boost::p ython 静态方法返回类的实例
- boost::p ython 纯虚拟基类，具有静态工厂构造函数和 std::unique_ptr
- 在类中公开公共结构以 boost::p ython
- 如何使用Boost::P ython打印到Python终端
- 如何使用 boost::p ython::迭代器与return_internal_reference
