[Numpy-discussion] documenting index_tricks

Pauli Virtanen pav@iki...
Sun May 24 14:21:59 CDT 2009


Sun, 24 May 2009 14:29:30 -0400, Ralf Gommers wrote:

> Hi all,
> 
> I'm documenting the index_tricks module and am unsure how to handle r_
> and related objects. r_ is an instance of RClass, therefore ipython
> shows the RClass docstring for "r_?". RClass in the doc editor is also
> marked as "needs editing", and the page for r_ is completely empty (see
> http://docs.scipy.org/numpy/docs/numpy.r_/#numpy-r). It seems like only
> one of the two should be edited and then the docstring should be copied
> over. In Travis Oliphant's Guide to Numpy r_ is extensively documented,
> should I integrate part of that with one of the docstrings?

Yes, please reuse the r_ documentation, and write it on the
numpy.lib.index_tricks.r_ page. We already do special tricks to get the 
mgrid and ogrid docs to work as they should, and it's simple to do the 
same for r_ and c_.

There are many of the r_ object pages in the doc wiki: this should be 
fixed, but I don't have time to do much this in the following couple of 
weeks.

> About the code itself, is there a use case for having two RClass
> instances at the same time? If not maybe RClass should be a private
> class. Also a name like RowConcatenator (it is a subclass of
> AxisConcatenator) would make more sense. I'm not sure this change is
> worth the trouble though.

I think the RClass is a private implementation detail, and shouldn't be 
documented in the reference guide.

-- 
Pauli Virtanen



More information about the Numpy-discussion mailing list