[Numpy-discussion] Toward release 1.0 of NumPy

Tim Hochberg tim.hochberg at cox.net
Thu Apr 13 08:44:47 CDT 2006


Alan G Isaac wrote:

>On Thu, 13 Apr 2006, Charles R Harris apparently wrote: 
>  
>
>>The Kronecker product (aka Tensor product) of two 
>>matrices isn't a matrix. 
>>    
>>
>
>That is an unusual way to describe things in
>the world of econometrics.  Here is a more
>common way:
>http://planetmath.org/encyclopedia/KroneckerProduct.html
>I share Sven's expectation.
>  
>
mathworld also agrees with you. As does the documentation (as best as I 
can tell) and the actual output of kron. I think Charles must be 
thinking of the tensor product instead.

In fact, if you look at the code you see this:

    # TODO:  figure out how to keep arrays the same

I think that in general this is going to be a bit of an issue whenever 
we have multiple arguments. Let me propose the world's second dumbest 
(in a good way, maybe) procedure:

    def kron(a,b):
        wrappers = [(getattr(x, '__array_priority__', 0),
    x.__array_wrap__) for x in [a,b]
                             if hasattr(x, '__array_wrap__')]
        if wrappers:
            priority, wrap = wrappers[-1]
        else:
            wrap = None
        # ....
        result = concatenate(concatenate(o, axis=1), axis=1)
        if wrap is not None:
            result = wrap(result)
        return result

   
This generalizes what _wrapit does for arbitrary arguments. It breaks 
'ties' where more than one argument wants to wrap something by using 
__array_priority__.  You'd actually want to factor out the wrapper 
finding code. This generalized what _wrapit does to multiple dimensions.

Thought?

Better plans?

-tim





More information about the Numpy-discussion mailing list