如何将复杂数字从python numpy传递给c(此刻尝试使用SWIG)

时间:2012-08-09 08:15:24

标签: python c numpy swig

我想在复杂的编号输入和python中调用c编写的函数。我已经尝试过使用SWIG来生成一个包装器 - 但它似乎已经失败了。我想我需要找出适当的'宏'用于numpy.i - 但不确定它是什么 - 任何人都有这方面的经验 - 或者其他方式我可以解决这个问题?

numpy.i在底部显示了这一点 - 尽管它已经被注释掉了。我尝试使用这些宏 - 但它们失败了,SWIG抱怨我尝试过以下宏扩展的语法错误:

%numpy_typemaps(complex float, NPY_CFLOAT , int)
%numpy_typemaps(complex double, NPY_CDOUBLE, int)
%numpy_typemaps(complex long double, NPY_CLONGDOUBLE, int)

这些是我的文件:

ComplexNumbers.c

# include <math.h>
# include <complex.h>


double complex returnX(double complex X) 
/*
    fresnel reflection coefficient rs
*/
{
    return X;
}

ComplexNumbers.i:

%{
#define SWIG_FILE_WITH_INIT
%}
%include "numpy.i"
%init %{
import_array();
%}


%module ComplexNumbers

%inline %{
extern double complex returnX(double complex X);
%}

的Python:

#!/usr/bin/env python

"""
setup.py file for ComplexNumbers
"""

from distutils.core import setup
from distutils.extension import Extension

import numpy


ComplexNumbers_module = Extension('_ComplexNumbers',
                           sources=['ComplexNumbers_wrap.c', 
                                    'ComplexNumbers.c'],
                           include_dirs=[numpy.get_include()]
                           )

setup (name = 'ComplexNumbers',
       version = '1.0',
       author      = "JP Hadden jp.hadden@bristol.ac.uk",
       description = """Spectral Interfereometry functions""",
       ext_modules = [ComplexNumbers_module],
       py_modules = ["ComplexNumbers"],
       )

编译器输出错误

C:\MinGW32-xy\bin\gcc.exe -mno-cygwin -mdll -O -Wall -IC:\Python27\lib\site-pack
ages\numpy\core\include -IC:\Python27\include -IC:\Python27\PC -c ComplexNumbers
_wrap.c -o build\temp.win32-2.7\Release\complexnumbers_wrap.o
ComplexNumbers_wrap.c:2975:23: error: expected '=', ',', ';', 'asm' or '__attrib
ute__' before 'returnX'
ComplexNumbers_wrap.c: In function '_wrap_returnX':
ComplexNumbers_wrap.c:2982:18: error: expected '=', ',', ';', 'asm' or '__attrib
ute__' before 'arg1'
ComplexNumbers_wrap.c:2982:18: error: 'arg1' undeclared (first use in this funct
ion)
ComplexNumbers_wrap.c:2982:18: note: each undeclared identifier is reported only
 once for each function it appears in
ComplexNumbers_wrap.c:2986:18: error: expected '=', ',', ';', 'asm' or '__attrib
ute__' before 'result'
ComplexNumbers_wrap.c:2986:18: error: 'result' undeclared (first use in this fun
ction)
ComplexNumbers_wrap.c:2997:24: error: expected ')' before 'complex'
ComplexNumbers_wrap.c:2997:24: error: pointer value used where a floating point
value was expected
ComplexNumbers_wrap.c:2997:14: error: invalid type argument of unary '*' (have '
double')
ComplexNumbers_wrap.c:3000:20: error: expected ')' before 'complex'
ComplexNumbers_wrap.c:3000:20: warning: implicit declaration of function 'return
X'
ComplexNumbers_wrap.c:3001:15: error: expected ')' before 'complex'
ComplexNumbers_wrap.c:3001:15: error: expected ')' before 'complex'
ComplexNumbers_wrap.c:3001:15: error: expected ')' before 'complex'
ComplexNumbers_wrap.c:3001:15: error: pointer value used where a floating point
value was expected
ComplexNumbers_wrap.c:3001:15: error: expected ')' before 'complex'
ComplexNumbers_wrap.c:3001:15: error: incompatible type for argument 1 of 'memcp
y'
c:\mingw32-xy\bin\../lib/gcc/mingw32/4.5.2/../../../../include/string.h:38:40: n
ote: expected 'void *' but argument is of type 'double'
error: command 'gcc' failed with exit status 1

1 个答案:

答案 0 :(得分:3)

我不确定这是如何与Numpy交互的,但SWIG肯定包括对C99复杂类型的支持。我能够通过以下示例验证这一点:

%module test
%{
#include <complex.h>
%}

%include <complex.i>

%inline %{
  double complex test(double complex X) {
    return X;
  }
%}

这使用SWIG complex.i接口。

(注意,这里%inline用于在一个地方声明,定义和包装 - 方便测试,但对于“真正的”代码,我倾向于使用%include)。

我用(

)编译了这个(在Linux上)
swig -Wall -python test.i 
gcc -std=c99 -Wall -Wextra test_wrap.c -I/usr/include/python2.7 -shared -o _test.so

只用一个警告就建好了。一旦完成,我就可以做到:

LD_LIBRARY_PATH=.  python                       
Python 2.7.3 (default, Aug  1 2012, 05:16:07) 
[GCC 4.6.3] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import test
>>> test.test(complex(0,1))
1j
>>> test.test(complex(0,2))
2j
>>> test.test(complex(1,2))
(1+2j)
>>> 

似乎接受并返回原生Python复杂类型。

我还能够进行以下界面编译:

%{
#define SWIG_FILE_WITH_INIT
#include <complex.h>
%}
%include <complex.i>
%include <numpy.i>
%init %{
import_array();
%}

%numpy_typemaps(complex float, NPY_CFLOAT , int)
%numpy_typemaps(complex double, NPY_CDOUBLE, int)
%numpy_typemaps(complex long double, NPY_CLONGDOUBLE, int)

%module test

%inline %{
  double complex test(double complex X) {
    return X;
  }
%}

在编译器标志中添加%include <complex.i>-std=c99。我认为你可以使用像these options这样的东西来设置distutils。