I’m trying to deal with an arbitrary-sized (NxMxZ) 3D matrix in Python, about 50MB of floating point numbers in total. I need to do simple as-efficient-as-possible sum and average calculations across axes and diagonals, but nothing very fancy, and the matrix is dense.
Anyone know if such a library exists? I’ve found a number of “3D matrix” libraries for python, but they’re all for 3D graphics, and are limited to, e.g. 4x4x4 matrices. Normally I’d use Numpy, but I’m on Google AppEngine, and can’t use a library that requires C extensions.
We just announced a trusted tester program for Python 2.7 support, which includes NumPy. You might want to consider signing up for it.