Re: [Matplotlib-users] plot_surface does not work

[Jody K. <jk...@uv...>]

Hi All,


On Dec 11, 2012, at  16:59 PM, Damon McDougall <dam...@gm...> wrote:

> On Tue, Dec 11, 2012 at 1:16 PM, Benjamin Root <ben...@ou...> wrote:
>> 
>> 
>> On Tue, Dec 11, 2012 at 2:08 PM, Chloe Lewis <ch...@be...> wrote:
>>> 
>>> Would it be workable for the default to be proportional to the size of the
>>> array passed in? (suggested only because I do that myself, when deciding how
>>> coarse an investigative plot I can get away with.)
>>> 
>>> &C
>>> 
>> 
>> That is pretty much what the PR I was referring to does:
>> 
>> https://github.com/matplotlib/matplotlib/pull/1040
>> 
>> It makes it so that the behavior of both plot_surface and plot_wireframe is
>> the same in this respect.  So, by default, the rstride and cstride would be
>> 1% of the size of your data array.  This would make the default for the
>> recent example be 1, therefore showing every point.  I wonder if a
>> logarithmic default would make sense to better handle large data arrays?
>> 
>> Thoughts?
>> Ben Root
> 
> I hope nobody minds if I chime in here.
> 
> I'm in favour of making the defaults a little more intelligent that
> what is implemented at present, i.e, a constant stride for any
> surface. Any non-trivial scaling law to determine what stride to use
> will result in more expected behaviour than what our users are
> currently seeing.
> 
> Could we do better? Could we have plot_surface try and estimate the
> stride based on the 'roughness' of the surface to be plotted? This
> method would grind to a halt for very rough surfaces, so we could
> default to a scaling law in these cases.
> 

OK, way late here, but

1) I wasted an hour today before I discovered what "rstride" and "cstride" were.  Reading the documentation, I still don't actually know what they are, except that if I want to see all my data I need to set them to 1. "Array row stride (step size)", is pretty enigmatic!  "stride" is a term I've never heard before except is reference to walking. I see it is used in computer science, but to refer to the byte-wise distance between array elements, so not very analogous. 

Can I suggest the docs be improved to say exactly what these do (I assume either average over cstride columns and rstride rows, or subsample on that frequency, not clear which)? Can I also suggest the default is 1?  Its pretty frustrating for large a chunk  of your data to not show up for no logical reason.  If my data set is too large, I am smart enough to subsample it myself before I plot it.   

2) Can I suggest this example be added to the tutorial?  http://stackoverflow.com/questions/6539944/color-matplotlib-plot-surface-command-with-surface-gradient
None of the other examples explain how to colour your surface with data, which is what I wanted.  

3) I think plot_surface should accept a fourth (optional) argument C for colouring the faces: plot_surface(X,Y,Z,C).  I do this a lot if I want to make a 3-D plot, and normalizing C, clipping it, and indexing a colormap seem clunky, when the routine could do it for me.  

Thanks,   Jody

--
Jody Klymak    
http://web.uvic.ca/~jklymak/