Passing a set of NumPy arrays into C function for input and output

To do this specifically with Numpy arrays, you could use:

import numpy as npimport ctypescount = 5size = 1000#create some arraysarrays = [np.arange(size,dtype="float32") for ii in range(count)] #get ctypes handlesctypes_arrays = [np.ctypeslib.as_ctypes(array) for array in arrays]#Pack into pointer arraypointer_ar = (ctypes.POINTER(C.c_float) * count)(*ctypes_arrays)ctypes.CDLL("./libfoo.so").foo(ctypes.c_int(count), pointer_ar, ctypes.c_int(size))

Where the C side of things might look like:

# function to multiply all arrays by 2void foo(int count, float** array, int size){   int ii,jj;   for (ii=0;ii<count;ii++){      for (jj=0;jj<size;jj++)         array[ii][jj] *= 2;       }}

In C, float** points to first element in a table/array of float* pointers.

Presumably each of those float* points to first element in a table/array of float values.

Your function declaration has 1 count, however it's not clear what this count applies to:

void compute (int count, float** input, float** output)

• 2D matrix count x count in size?
• count -sized array of float* each somehow terminated, e.g. with nan?
• null-terminated array of float* each of count elements (reasonable assumption)?

Please clarify your question and I will clarify my answer :-)

Assuming the last API interpretation, here's my sample compute function:

/* null-terminated array of float*, each points to count-sized array*/extern void compute(int count, float** in, float** out){    while (*in)    {        for (int i=0; i<count; i++)        {            (*out)[i] = (*in)[i]*42;        }        in++; out++;    }}

Test code for the sample compute function:

#include <stdio.h>extern void compute(int count, float** in, float** out);int main(int argc, char** argv){#define COUNT 3    float ina[COUNT] = { 1.5, 0.5, 3.0 };    float inb[COUNT] = { 0.1, -0.2, -10.0 };    float outa[COUNT];    float outb[COUNT];    float* in[] = {ina, inb, (float*)0};    float* out[] = {outa, outb, (float*)0};    compute(COUNT, in, out);    for (int row=0; row<2; row++)        for (int c=0; c<COUNT; c++)            printf("%d %d %f %f\n", row, c, in[row][c], out[row][c]);    return 0;}

And how you use same via ctypes in Python for count == 10 float subarrays and size 2 float* array, containing 1 real subarray and NULL terminator:

import ctypesinnertype = ctypes.ARRAY(ctypes.c_float, 10)outertype = ctypes.ARRAY(ctypes.POINTER(ctypes.c_float), 2)in1 = innertype(*range(10))in_ = outertype(in1, None)out1 = innertype(*range(10))out = outertype(out1, None)ctypes.CDLL("./compute.so").compute(10, in_, out)for i in range(10): print in_[0][i], out[0][i]

Numpy interface to ctypes is covered here http://www.scipy.org/Cookbook/Ctypes#head-4ee0c35d45f89ef959a7d77b94c1c973101a562f, arr.ctypes.shape[:] arr.ctypes.strides[:] and arr.ctypes.data are what you need; you might be able to feed that directly to your compute.

Here's an example:

In [55]: a = numpy.array([[0.0]*10]*2, dtype=numpy.float32)In [56]: ctypes.cast(a.ctypes.data, ctypes.POINTER(ctypes.c_float))[0]Out[56]: 0.0In [57]: ctypes.cast(a.ctypes.data, ctypes.POINTER(ctypes.c_float))[0] = 1234In [58]: aOut[58]: array([[ 1234.,     0.,     0.,     0.,     0.,     0.,     0.,     0.,            0.,     0.],       [    0.,     0.,     0.,     0.,     0.,     0.,     0.,     0.,            0.,     0.]], dtype=float32)