matplotlib-users Mailing List for matplotlib

Hi all,

On 6/3/06, Charlie Moad <cw...@gm...> wrote:
> On 6/2/06, David S. <dav...@al...> wrote:
> matplotlib-0.87.2 requires numpy-0.9.6.  You will get an error otherwise.

I also had problems installing matplotlib with numpy on windows.
Basicaly, I installed matplotlib, didn't create the matplotlib rc file
(it isn't created by the installer), and it turned out, that by
default, pylab needed  numeric, not numpy.

Now back to inspecting the internals of matplotlib :)

On 6/3/06, John Hunter <jdh...@ac...> wrote:
> >>>>> "Charlie" == Charlie Moad <cw...@gm...> writes:
>
>     Charlie> On 6/2/06, David S. <dav...@al...> wrote:
>     >> I have just installed numpy-0.9.8, scipy-0.4.9, and
>     >> matplotlib-0.87.2 on a Windows machine with Python 2.4.2.
>
>     Charlie> matplotlib-0.87.2 requires numpy-0.9.6.  You will get an
>     Charlie> error otherwise.
>
>
> This is starting to bite enough people that we should consider rolling
> out a release.  We had hope to get the 3d stuff working before the
> next release, but having a version up there that doesn't work with the
> latest numpy is a more pressing problem to me.  Does this seem like a
> good idea to you Charlie?  Are there other features or fixes that
> other developers would like to get in first?

Agreed.  I think we are tied closely enough to numpy that a minor
release is justified.  A 0.88 release doesn't seem justified yet, so I
could push a 0.87.3 barring any outstanding bugs.

- Charlie

>>>>> "Charlie" == Charlie Moad <cw...@gm...> writes:

    Charlie> On 6/2/06, David S. <dav...@al...> wrote:
    >> I have just installed numpy-0.9.8, scipy-0.4.9, and
    >> matplotlib-0.87.2 on a Windows machine with Python 2.4.2.

    Charlie> matplotlib-0.87.2 requires numpy-0.9.6.  You will get an
    Charlie> error otherwise.


This is starting to bite enough people that we should consider rolling
out a release.  We had hope to get the 3d stuff working before the
next release, but having a version up there that doesn't work with the
latest numpy is a more pressing problem to me.  Does this seem like a
good idea to you Charlie?  Are there other features or fixes that
other developers would like to get in first?

JDH

>>>>> "Willi" == Willi Richert <ri...@c-...> writes:

    Willi> Hi, is there a way to display confidence intervals other
    Willi> than using Polygon like, e.g.

    Willi> ax = subplot(111) # make the shaded region upperPoints =
    Willi> zip(runs, upperConf) upperPoints.reverse() verts =
    Willi> [(runs[0], upperConf[0])] + zip(runs, lowerConf) +
    Willi> upperPoints poly = Polygon(verts, facecolor=0.6,
    Willi> edgecolor=0.6) ax.add_patch(poly)

Have you seen fill?

  http://matplotlib.sourceforge.net/screenshots.html#fill_demo

For upper and lower bounds, you need to reverse the x values for the
lower bounds, so the polygon will fill in order.  Here's how you would
do it with it with lists (untested), where x are the x data points and
upper and lower are len(x) lists with the upper and lower confidence bounds

  xr = x[:] 
  xr.reverse()
  lower.reverse()
  fill(x + xr, upper + lower)

JDH