Fast interpolation of grid data

Question

I have a large 3d np.ndarray of data that represents a physical variable sampled over a volume in a regular grid fashion (as in the value in array[0,0,0] represents the valu

Answer 1

Great answer by Joe. Based on his suggestion, I created the regulargrid package (https://pypi.python.org/pypi/regulargrid/, source at https://github.com/JohannesBuchner/regulargrid)

It provides support for n-dimensional Cartesian grids (as needed here) via the very fast scipy.ndimage.map_coordinates for arbitrary coordinate scales.