Perhaps you can do something along the following lines to get around
this limitation:
#
# parameterizes the original function by delta, size.
def parameterized_function(delta, size, function):
center = (size-1)/2.0
return lambda i: function( (i-center)*delta )
# the func
I've been using something I wrote:
coef_from_function (function, delta, size)
which does (c++ code):
double center = double(size-1)/2;
for (int i = 0; i < size; ++i)
coef[i] = call (func, double(i - center) * delta);
I thought to translate this to np.fromfunction. It seems fromfunction
