So I'm all jazzed wrapping CLUTO's static library in a Python module using SWIG.Things are going smoothly when on a lark I decide to hit Google with "python clustering".
Lo and behold, it gives me Pycluster, a module that wraps a C level clustering library from UC Berkeley. There's also a bunch of other toys based on the same library, all cooked up by Michiel de Hoon and friends.
It looks like there's significant overlap in the clustering algorithms provided by CLUTO vs Pycluster, but they're different enough to both be useful. One nice thing is that Pycluster actually does self-organizing maps (SOM).
Might be time for a bake-off.