Wed, Nov 11, 2009, 2:29am Additional mathy-Pythonic touches
Programming » Math-Science
(Last updated: Thu, Nov 12, 2009, 3:09am)
ith things running smoothly with my SciPy setup, I have been having fun trying out new libraries. Here's just a short list of some that I have been playing with.

  • SymPy — Symbolic mathematics system. Very powerful, and gives me quick access to a lot of what I missed from the tools in other systems. Sadly, the Plot functions in SymPy fail if they try to do any linking to the QuickTime 7 libraries, which are stuck in 32-bit land. SymPy also includes mymath as part of it, but it's neat enough to merit individual mention…
  • Mpmath — A multiprecision floating-point arithmetic library. I'm very impressed with it. The speed with which it lets me play with hypergeometric functions fills my mathy heart with joy.
  • GMPY — A C-coded Python extension that can be used by mpmath. I was unable to get v1.10 to work, but v1.04 did me just fine. When I ran the mpmath test suite before gmpy it took 32.03s on a 3GHz Xeon core, and with gmpy that time was reduced to 13.61s, so well worth installing. (One caveat was that my libgmp was 64-bit, so I only bothered installing this in 64-bit form instead of dual 32/64-bit, since I've been unable to get python to load up NumPy in 32-bit mode anyway.)
  • (I was also going to pop in Psycho, as mentioned in the mpmath manual, but it's 32-bit only.)
  • PIL — Not the band, but the Python Imaging Library, for image processing. Not sure if I'll use this one or not, but it's super easy to install and there if I want it.
  • and another good one that was added a day later:
  • FiPy — A finite volume PDE solver, which is very neat. See a quick-n-dirty demo of it in action which I made via QuickTime Player X here. (It requires that you also install Pysparse)
  • There are many others to try out, but this stuff was nice to try out and get to know the current state of what I can play with along these lines. It's a wonderful freedom made available by these projects for all those doing research, especially home researchers, a catagory rarely even acknowledged by the main math software companies these days.

    Any other gems that others want to give a shout out to?

    Leave a comment