matplotlib-devel Mailing List for matplotlib

Hi,
I've built and tested some binary wheels for OSX - available here:
https://nipy.bic.berkeley.edu/scipy_installers/
I'd really like to upload these to pypi so that people get them by
default with 'pip install matplotlb'. Is that OK? Can someone give
me permission to do that?
Thanks a lot,
Matthew

Hi,
On Mon, Mar 24, 2014 at 6:29 AM, Benjamin Root <ben...@ou...> wrote:
> I thought we fixed this one...
>
> Seems like we haven't as there is an open issue for it:
> https://github.com/matplotlib/matplotlib/issues/2842
Sorry - I didn't say - but the wheels are for the 1.3.1 release...
Cheers,
Matthew

Hi,
Working on mpld3, I've hit something I don't quite know how to handle. I'm
trying to render legend components in d3, and the strange thing is that
markers within the legend have empty paths. Consider the script below:
#------------------------------------------------------
import matplotlib.pyplot as plt
fig, ax = plt.subplots()
line, = ax.plot(range(5), range(5), 'o', label='dots')
ax.legend()
print("here is the marker as it appears in the plot:")
print(line._marker.get_path())
leg = ax.get_legend()
children = leg.get_children()
print("here is the marker as it appears in the legend")
print(children[1]._marker.get_path())
#------------------------------------------------------
On the plot itself, the marker is a circle. On the legend, the marker is an
empty path. I've tried poking around in the backend code to figure out how
the renderers know what path to use when drawing the legend, but the answer
is not obvious to me.
Does anyone have insight into what special magic happens here when the
legend is rendered? Thanks,
 Jake
 Jake VanderPlas
 Director of Research - Physical Sciences
 eScience Institute, University of Washington
 http://www.vanderplas.com

Wow. Thanks so much, Stan! This is a huge help and works just as I need it
to.
Much appreciated!
On Mon, Mar 24, 2014 at 11:26 AM, Stan West <sta...@nr...> wrote:
> On 2014年03月24日 14:08, Stan West wrote:
>
> May I suggest that you look at the mailing list thread from that time [1],
> try the patch in the thread, and see whether your issue is resolved? This
> solution doesn't provide a work-around in your code, but it may fix the
> problem at the root.
>
> Paul, I just found the work-around that I used in my code. Define the
> following function:
>
> def set_spine_position(spine, position):
> """
> Set the spine's position without resetting an associated axis.
>
> As of matplotlib v. 1.0.0, if a spine has an associated axis, then
> spine.set_position() calls axis.cla(), which resets locators, formatters,
> etc. We temporarily replace that call with axis.reset_ticks(), which is
> sufficient for our purposes.
> """
> axis = spine.axis
> if axis is not None:
> cla = axis.cla
> axis.cla = axis.reset_ticks
> spine.set_position(position)
> if axis is not None:
> axis.cla = cla
>
> (The mention of v. 1.0.0 in the docstring is just the version I was using
> at the time, not necessarily the earliest version with this issue.) Then
> replace method calls like "spine.set_position(pos)" with the function call
> "set_spine_position(spine, pos)". I hope this helps.
>