[SciPy-Dev] scipy.stats: some questions/points about distributions.py + reply on ticket 1493

nicky van foreest vanforeest@gmail....
Thu Apr 26 16:19:35 CDT 2012


> L1745 are necessary now because moments are still non-central, but
> allow for non default loc and scale.
> _munp still does the raw quad( lamdba x: x**k *_pdf(x)) or similar if
> it is not specifically defined, i.e. can be calculated from explicit
> _stats.
>
> as aside: whenever I try to go through generic _stats and moments I
> have a hard time to follow in what is going on in all the different
> cases
>
> I don't see where we could save much. I didn't go through the math.

Perhaps we should postpone the problems with the moments computations
for the moment. I'll come across this point in due time and bug you
again about it.


> from numpy import asarray
> or
> from numpy import asarray as arr
> doesn't make any difference, but I usually like full namespaces
> np.asarray

About replacing arr by asarray: do I get it right that you change
this? Or should  I try making a pull request for this?

> https://github.com/scipy/scipy/blob/master/scipy/stats/distributions.py#L538
>>>>
>>>> contains a typo. It should be Weisstein.
>>>
>>> should be fixed then
>>
>> Should this become a ticket, or is it too minor?
>
> too minor for a ticket, but sneaking it into a pull request, or making
> a separate pull request (to increase your Karma) would be useful

I just got a local version of scipy and wrote a patch for this. For
today I'll try to turn it into a pull request. As this is my first
attempt, I have no idea how long it will take to set everything up.
The spin off will be, hopefully, that I can add to the code at other
places too. I suppose you will see the patch appearing.

Nicky


More information about the SciPy-Dev mailing list