[Python-modules-team] Issue in numpydoc (Was: Further help needed for statsmodels)

[Denis Laxalde]

Andreas Tille a écrit :
> Hi Denis,
>
> there seems to be an issue with numpydoc which breaks the build of
> statsmodels.  I'm on my way back home from DebConf and can't do any
> sensible research into this, but I'd consider it urgent to get
> statsmodels uploaded soonish.
>

https://github.com/statsmodels/statsmodels/blob/v0.8.0rc1/statsmodels/genmod/families/family.py#L92

There's a duplicate "Notes" section in Family.starting_mu(). Deleting 
lines 98-99 in statsmodels/genmod/families/family.py should fix the 
error raised by numpydoc.

> Any help would be really appreciated
>
>       Andreas.
>
> On Thu, Aug 17, 2017 at 02:58:57PM +0200, Julien Puydt wrote:
>> Hi,
>>
>> Le 17/08/2017 à 14:45, Andreas Tille a écrit :
>>> as it was discussed in bug #868977 the tests that are failing due to an
>>> issue in pandas were deactivated now.  Unfortunately there is another issue
>>> that seems to be caused by numpydoc.  My build ends up with:
>>
>>> Exception occurred:
>>>   File "/usr/lib/python2.7/dist-packages/numpydoc/docscrape.py", line 343, in _parse
>>>     raise ValueError(msg)
>>> ValueError: The section Notes appears twice in the docstring of None in None.
>>>
>>> Any help with this to save several rdepends of statsmodels from beeing
>>> removed from testing would be really welcome.
>>
>> It's annoying because something has been done in that part of the code
>> quite recently in numpydoc, and I think it's part of numpydoc 0.7.0 already:
>> https://github.com/numpy/numpydoc/commit/46036b0a4e458f79ee6b8980b9d0e15fe204b143
>>
>> May I suggest asking to the guy who did that commit?
>>
>> Snark on #debian-science
>>
>>
>