The definitive method to use NumPy and SciPy from IronPython

Question

There is a way to use NumPy/SciPy in IronPython, using IronClad to execute/communicate with the CPython binaries of the same. A newer project, Python Tools for VS allows for a faster integration with .NET IronPython programs because most of the NumPy/SciPy library functionality has been manually ported into IronPython. Comments on the same page point to this blogpost which links to github projects for the same.

As of today (Oct 2012), what is the definitive method to integrate/use these 2 libraries from IronPython NET python programs?

Ramachandra

Answer 1

Scipy for .NET looks like a dream come true to me. MSVS roundhouse kicks every other python IDE in the face, and seemless interop with all other .NET languages would be awesome (looking at you in particular, F#... you could really be the functional gateway drug... but I can't leave my scientific python toolstack behind!)

Very sad to see that it appears the project is not very actively supported, to put it mildly. First thing on my todo list if I win the lottery... Supporting this project would be a staggering reward/efford project for Microsoft to boost its standing in the scientific computing sphere; going from being almost entirely of the radar, to the very top of the stack. Just in case anyone with a budget at Microsoft is looking at this: seriously. You are already giving MSVS for free to this crummy academic anyway, but I'd force my future boss to buy my MSVS at gunpoint*, if you realized the full potential of this. (*not that my amazing productivity wouldn't convince him anyway...)

So no, not much of a direct answer to your question, but maybe this rant will pay off sometime in the future.

Unparliamentary

Answer 2

The Enthought blog post you cited is still the best way to set it up.

Paedo

a year later and this still seems true (even though the last commit there is from 2011) - at least I found no other way – Weightless