Skip to content

Latest commit

 

History

History
100 lines (70 loc) · 3.19 KB

interfacing-c-with-cython.md

File metadata and controls

100 lines (70 loc) · 3.19 KB
Raw

Interfacing C code with Cython

As Cython code compiles down to C code, it is relatively straightforward to utilize also Cython for interfacing with C.

In order to use C functions and variables from the library, one must provide external declarations for them in the Cython .pyx file. While normal cdef declarations refer to functions and variables that are defined in the same file, by adding extern keyword one can specify that they are defined elsewhere. As an example, by having in a .pyx file the statement

cdef extern void add(double *a, double *b, double *c, int n)

one could call add function within that file. In addition, the actual library or source implementing the function needs to be included in setup.py when building the extension module with Cython.

With the above construct, Cython will add the declaration to the generated .c file. However, when using libraries it is preferable to have the actual library header included in the generated file. This can be achieved with the following construct:

cdef extern from "mylib.h"
    void add(double *a, double *b, double *c, int n)
    void subtract(double *a, double *b, double *c, int n)

Now, mylib.h header is included in the generated .c file, while the statements in the following block specify that functions add and subtract can be used within the .pyx file. It is important to understand that Cython does not itself read the C header file, so you still need to provide declarations from it that you use.

Including the library in setup.py

Compared to building simple pure Cython modules, one has to provide some extra information in the setup.py. If the code to be interfaced is in a library (i.e. .so) one can use the following type of setup.py:

from distutils.core import setup, Extension
from Cython.Build import cythonize

ext = Extension("module_name",
                sources=["cython_source.pyx",],
                libraries=["name",], # Cython module is linked against
                library_dirs=[".",]) # libname.so, looked in "."

setup(ext_modules=cythonize(ext))

One can also use direct C source files if more appropriate:

from distutils.core import setup, Extension
from Cython.Build import cythonize

# Specify all sources in Extension object
ext = Extension("module_name",
                sources=["cython_source.pyx", "c_source.c"])

setup(ext_modules=cythonize(ext))

Passing NumPy arrays from Cython to C

Similarly as when using CFFI to pass NumPy arrays into C, also in the case of Cython one needs to be able to pass a pointer to the "data area" of an array. For arrays that are declared as type of ndarray, Cython supports similar & syntax as in C:

import numpy as np
cimport numpy as np

cdef extern from "myclib.h":

    void add(double *a, double *b, double *c, int n)

    void subtract(double *a, double *b, double *c, int n)

def add_py(np.ndarray[cnp.double_t,ndim=1] a,
           np.ndarray[cnp.double_t,ndim=1] b):

    add(&a[0], &b[0], &c[0], len(a))