Weave stuff. (insanely obsolete)

Accessing an array and its metadata through C/C++ in scipy.weave.inline()

• Accessing an array foo with foo.ndim == 3:
  • The array itself will be accessable through the FOO3(i, j, k) function.
  • Array assignments will be the same. FOO3(i, j, k) = bar;
  • foo.shape will be accessable through the Nfoo[] array.

The inline code:

long dim1, dim2, dim3;

dim1 = Nfoo[0]; // dimension access.
dim2 = Nfoo[1];
dim3 = Nfoo[2];

FOO3(0,0,0) = 1; // array assignment
return_val = FOO3(0,0,0); // array access.

The Python code:

import scipy
import scipy.weave
inline_code = r"""
put C/C++ code in here
"""
foo = scipy.arange(3**3).reshape((3,3,3))
scipy.weave.inline(inline_code, foo)

Using Python functions in weave.inline

Call foo.call(py:tuple arg).

Links

• At Weave and Numpy array arguments: Scipy's Weave page
• Calling Python functions from inline C with scipy.weave (stackoverflow)