How to write an easy to use, interactive physics/science/engineering simulator leveraging ctypes, numba, tkinter and scipy.ndimage + PIL

Abstract

Python gives so many powerful, easy-to-use libraries, it should be easy to build simulator software for our physics or other sciences or engineering simulations using basic coding skill. But it may seem too hard or time-consuming to try. That's a shame! Instead we pay lots of money and become addicted to powerful, expensive simulation software that mostly does what we don't need and not quite what we do. So I'll tell my story of how I wrote a skeleton script for the equations of my problem, wrapped a GUI around it with tkinter, and made the numpy math (and therefore response) much, much faster by using the @jit compiler effectively and correctly. For those operations I couldn't easily cast as numpy array manipulations, I'll show how I put my "loopy", branching bits into C++ and linked with ctypes. Finally, since my data involves images, I'll show how numpy methods, scipy.ndimage and PIL provides a wealth of processing functionality.

Description