[Numpy-discussion] Array of Callables

Michael McLay mclay@l2sg....
Wed Mar 21 09:53:36 CDT 2007


On Wednesday 21 March 2007 09:52, Andrew Corrigan wrote:
> Anne Archibald <peridot.faceted <at> gmail.com> writes:
> > Vectorizing apply is what you're looking for, by the sound of it:
> > In [13]: a = array([lambda x: x**2, lambda x: x**3])
> >
> > In [14]: b = arange(5)
> >
> > In [15]: va = vectorize(lambda f, x: f(x))
> >
> > In [16]: va(a[:,newaxis],b[newaxis,:])
> > Out[16]:
> > array([[ 0,  1,  4,  9, 16],
> >        [ 0,  1,  8, 27, 64]])
>
> Thanks for pointing that out.  Technically that works, but it doesn't
> really express this operation as concisely and as naturally as I'd like to
> be able to.
>
> What I really want is to be able to write:
> >>> a = array([lambda x: x**2, lambda x: x**3])
> >>> b = arange(5)
> >>> a(b)
>
> and get:
> array([[ 0,  1,  4,  9, 16],
>        [ 0,  1,  8, 27, 64]])
>
> instead of the present error message:
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
> TypeError: 'numpy.ndarray' object is not callable

You could create a class with a __call__ method like the following.


from numpy import arange, array,vectorize,newaxis

class VA:
    def __init__(self,arg):
        self.a = array(arg)
    def __call__(self,b):
        va = vectorize(lambda f, x: f(x))
        return va(self.a[:,newaxis],b[newaxis,:])
    
a = VA([lambda x: x**2, lambda x: x**3])
b = arange(5)

print a(b)



More information about the Numpy-discussion mailing list