[Distutils] Improved Stats Features in Python

[Ralf Gommers]

On Sun, Sep 11, 2016 at 11:26 AM, Rui Sarmento <rui_sarmento at hotmail.com>
wrote:

> Hi Ralf,
>
> Yes, in fact I was trying to submit in git but I have some doubts. I will
> explore it more tomorrow (it is late here).
>
> For example I still have doubts with this:
>
> "If you are adding new functionality, you need to add it to the
> documentation by editing (or creating) the appropriate file in docs/source
> ."
>
> What exactly is "the appropriate file"?
>
> and also "Open the docs/source/release/versionX.X.rst file that has the
> version number of the next release and add your changes to the appropriate
> section", I see that the last version in the repository is version0.8.rst
> but I'm not sure this is the file I should edit...
>
> Maybe tomorrow with a good night sleep I'll figure it out.
>

Okay, I'd say just don't worry too much about those details if it's not
clear. Here's what you do:
1. Send an email to the statsmodels mailing list saying you want to add KMO
and Bartlett's spericity, ask if that's welcomed and in what file to put
that.
2. Add your functions in that file, and add tests for that (for function in
fname.py, tests go in tests/test_fname.py).
3. Commit that and put it up on your own GitHub account.
4. From there, send a pull request to statsmodels. Or if it's really not in
good enough shape, Cc me (@rgommers) and I'll give you a few pointers.

Let's take the discussion off this list, this is very off-topic.

Cheers,
Ralf



> Cheers,
>
> Rui
> Às 22:56 de 10-09-2016, Ralf Gommers escreveu:
>
>
>
> On Sun, Sep 11, 2016 at 6:24 AM, Rui Sarmento <rui_sarmento at hotmail.com>
> wrote:
>
>> Dear Ralf,
>>
>> No problem, its always nice to discover something new. In fact I've seen
>> the statsmodel page you sent, talking about submitting with git. I'm not
>> familiar with these procedures. Is there a How-To you would suggest me to
>> read? It is the first time I submit to these repositories. My goal is to
>> submit two functions, one for Bartlett and another for KMO.
>>
>
> Did you see that this page expands at the bottom? This is pretty much a
> walkthrough of how you go about submitting a PR: http://statsmodels.
> sourceforge.net/devel/dev/git_notes.html. It also has links to a couple
> of other helpful tutorials.
>
> Cheers,
> Ralf
>
>
> Cheers,
>>
>> Rui
>>
>> Às 23:07 de 09-09-2016, Ralf Gommers escreveu:
>>
>>
>>
>> On Sat, Sep 10, 2016 at 10:01 AM, Rui Sarmento <rui_sarmento at hotmail.com>
>> wrote:
>>
>>> Dear Ralf,
>>>
>>> Thank you for your suggestions. About the Bartlett test I'm aware that
>>> one of his tests (equal variance of samples) is already available.
>>> Nonetheless, I'm not talking about that particular test but about other
>>> Bartlett's test. The test I wish to contribute is directed to Factor
>>> Analysis and is related to the test for sphericity. I'll try to submit both
>>> to the statsmodel.
>>>
>>
>> Ah okay, thanks - learned something new. For Bartlett's sphericity test
>> statsmodels is probably also the best place indeed.
>>
>> Cheers,
>> Ralf
>>
>>
>>
>>> Best Regards,
>>>
>>> Rui
>>> Às 22:46 de 09-09-2016, Ralf Gommers escreveu:
>>>
>>>
>>>
>>> On Fri, Sep 9, 2016 at 6:29 PM, Ronny Pfannschmidt <
>>> opensource at ronnypfannschmidt.de> wrote:
>>>
>>>> Hello Rui,
>>>>
>>>> this mailing list deal with tools you can use to publish 3rd party
>>>> packages to something like the pypi package index,
>>>>
>>>> if you want to add to the python stdlib, you need to get started with
>>>> python-ideas, python-dev and very likely write a PEP that will have to get
>>>> accepted.
>>>>
>>>> if you just want to publish your own library, you just need to upload
>>>> it to pypi and make it known.
>>>>
>>>> -- Ronny
>>>>
>>>> On 06.09.2016 17:06, Rui Sarmento wrote:
>>>>
>>>>> Dear Sirs,
>>>>>
>>>>> I've implemented some stats functions related to Factor Analysis in
>>>>> the statistics area. Specifically, the KMO test and the Bartlett test also.
>>>>> At this time I do not seem to find any module performing these tests. Is
>>>>> there any chance I could add these functions to a package in Python. What
>>>>> is the procedure to perform such contribution.
>>>>>
>>>>
>>> Barlett is already implemented in SciPy: http://docs.scipy.org/doc/scip
>>> y/reference/generated/scipy.stats.bartlett.html
>>>
>>> KMO isn't available anywhere as far as I can tell; statsmodels would be
>>> the best place if you would like to contribute your implementation there.
>>> See http://statsmodels.sourceforge.net/devel/dev/ for how to go about
>>> that. I wouldn't bother proposing that for stdlib inclusion, it's way too
>>> specialized for that.
>>>
>>> Cheers,
>>> Ralf
>>>
>>>
>>>
>>>>> Thank you very much in advance for the suggestions.
>>>>>
>>>>> Best Regards,
>>>>>
>>>>> Rui
>>>>>
>>>>> _______________________________________________
>>>>> Distutils-SIG maillist  -  Distutils-SIG at python.org
>>>>> https://mail.python.org/mailman/listinfo/distutils-sig
>>>>>
>>>>
>>>> _______________________________________________
>>>> Distutils-SIG maillist  -  Distutils-SIG at python.org
>>>> https://mail.python.org/mailman/listinfo/distutils-sig
>>>>
>>>
>>>
>>>
>>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/distutils-sig/attachments/20160911/3016018d/attachment-0001.html>