[SciPy-Dev] Makes stats._support functions more visible

Bruce Southey bsouthey at gmail.com
Sat Jul 3 22:12:49 EDT 2010


On Sat, Jul 3, 2010 at 12:02 PM, Benjamin Root <ben.root at ou.edu> wrote:
> unique() is in numpy.  As for the rest of them, these functions appear to be
> very specific to the stats package.  In particular, unique() in the support
> package is a very restricted version of the more generic unique() in numpy.
> Some of the other functions state that they are like other functions in
> numpy, but... different...
>
> I think they should stay right where they are.
>
> Ben Root
>
>
> On Fri, Jul 2, 2010 at 6:11 PM, Vincent Davis <vincent at vincentdavis.net>
> wrote:
>>
>> There are several useful functions in stats._support I don't think I
>> would have found them if I had  been looking for such tools. I think
>> it would be good for these to be more visible. Primarily I am thinking
>> of
>> abut()
>> adm()
>> colex()
>> unique()
>> linexand()
>> collapse()
>>
>> Actually most/all of these probably belong in numpy. I don't think
>> they duplicate any np functions.
>>
>> Vincent
>> _______________________________________________
>> SciPy-Dev mailing list
>> SciPy-Dev at scipy.org
>> http://mail.scipy.org/mailman/listinfo/scipy-dev
>
>
> _______________________________________________
> SciPy-Dev mailing list
> SciPy-Dev at scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-dev
>
>

Actually for the most part the stats functions need lots of attention
because these have not kept pace with numpy. Most are specific to very
specific inputs or perhaps can be one-liners (trimmed ones are an
example of that). So these would not be considered by numpy if the
functions are not sufficiently general.

So what you have to do is write new code for these functions.Then
submit a patch to scipy and perhaps an enhancement patch to numpy.

Bruce



More information about the SciPy-Dev mailing list