Hi Travis, Thanks for your work on this--it's very useful to me. I found 2 issues. I'm including a test and a potential fix for the first issue, which seems to be an end-point problem. Under some circumstances, the endpoints aren't properly detected. I didn't attempt to comprehend everything going on in this function, but the patch I made apparently works. Please review it and apply it if it's acceptable. The second issue is that the x array cannot be integers (a TypeError gets raised). There doesn't seem to be any good reason for this (why can't splines exist over integers?), so I submit that it's also a bug. Unfortunately, that didn't look as easy for me to fix, so I leave it for now. Cheers! Andrew Travis Oliphant wrote:

This example needs the SVN version of scipy (or you need to get the cspline1d_eval function out of SVN):

I sent a smaller image hoping it would make it to the list...

Example;

from numpy import r_, sin from scipy.signal import cspline1d, cspline1d_eval

x = r_[0:10] dx = x[1]-x[0] newx = r_[-3:13:0.1] # notice outside the original domain y = sin(x) cj = cspline1d(y) newy = cspline1d_eval(cj, newx, dx=dx,x0=x[0])

from pylab import plot plot(newx, newy, x, y, 'o')

Have fun,

-Travis