matplotlib-users Mailing List for matplotlib

hu...@ya... wrote:
> 				Hello,
> 
> I have some stupid questions about how to use colorbar.
> 
> 1) I would like to be able to put the colorbar where I went: top, bottom, 
> left, right. For what I see I can do only a vertical left and horizontal 
> bottom ones (with a simple use of the function).
> I know that it's possible to do a colorbar only but it's difficult to place 
> them a little bit automatically and that bring my second question.
This looks like a feature request: instead of specifying "horizontal" or 
"vertical", be able to specify "top", "bottom", "left", or "right". 
This is fairly easy in principle, but there are some wrinkles such that 
I am not sure it would be satisfactory in practice--some custom 
adjustments might usually be needed depending on whether the main axes 
have a title (in the case of "top") or an xlabel (in the case of 
"left"). You may be better off simply figuring out what main and 
colorbar axes positions work best for your particular case and then 
making those axes explicitly.
> 
> 2) I would like that the colorbar with exactly the same size than the image I 
> plot with pcolor. It's working when I'm not using the aspect='equal' in the 
> subplot but not with this option. Naturally I want to not change the aspect 
> of my image and so it's a big problem for me. The only solution I found is to 
> manipulate the colorbar with the shrink valu but it's a little bit painfull 
> especially the value change for each pcolor image (they doesn't have always 
> the same size).
This is a common request, and a reasonable one, but unfortunately it is 
not straightforward to implement automatically given mpl's internal 
structure. Again, however, you get the desired result quite easily by 
creating your own axes explicitly.
> 
> 3) I would like to give a label to the colorbar (ie if the colorbar is at the 
> bottom a label under the colorbar etc). To tell people what means the colors 
> but I didn't find anyway to do it. xlabel work for the pcolor object (as 
> expected) and there are no colorbar.xlabel function.
Use the set_label() method.
...
cbar = fig1.colorbar(im,orientation='horizontal')
cbar.set_label('label')
...
Eric

				Hello,
I have some stupid questions about how to use colorbar.
1) I would like to be able to put the colorbar where I went: top, bottom, 
left, right. For what I see I can do only a vertical left and horizontal 
bottom ones (with a simple use of the function).
I know that it's possible to do a colorbar only but it's difficult to place 
them a little bit automatically and that bring my second question.
2) I would like that the colorbar with exactly the same size than the image I 
plot with pcolor. It's working when I'm not using the aspect='equal' in the 
subplot but not with this option. Naturally I want to not change the aspect 
of my image and so it's a big problem for me. The only solution I found is to 
manipulate the colorbar with the shrink valu but it's a little bit painfull 
especially the value change for each pcolor image (they doesn't have always 
the same size).
3) I would like to give a label to the colorbar (ie if the colorbar is at the 
bottom a label under the colorbar etc). To tell people what means the colors 
but I didn't find anyway to do it. xlabel work for the pcolor object (as 
expected) and there are no colorbar.xlabel function.
Thank you for any help.
N.
ps: i join a simple script to show my problem
#!/usr/bin/env python
import pylab
import numpy
a = numpy.arange(100).reshape((10,10))
fig1 = pylab.figure()
ax = fig1.add_subplot(141)
im = ax.pcolor(a)
fig1.colorbar(im,orientation='horizontal') 
pylab.xlabel('label')
ax = fig1.add_subplot(142)
im = ax.pcolor(a)
fig1.colorbar(im,orientation='vertical')
pylab.xlabel('label')
ax = fig1.add_subplot(143,aspect='equal')
im = ax.pcolor(a)
fig1.colorbar(im,orientation='vertical')
pylab.xlabel('label')
ax = fig1.add_subplot(144,aspect='equal')
im = ax.pcolor(a)
fig1.colorbar(im,orientation='horizontal')
pylab.xlabel('label')
pylab.show()

Hi ,
I am trying to plot a filled triangle (or if possible any filled
polygon) in 3D. I am afraid, I am missing something very basic here.
I'd like to do something like:
x=[0,1,1] #coordintes of the three corners
y=[0,0,1]
z=[3,8,1]
ax.plane(x,y,z,facecolor='r')
So far I tried the "plot_surface", but this requires the points to be on
a rectangular grid. The solution proposed in "*Re: [Matplotlib-users] plot 3D with filled curves is similar to what i am searching for,
 but allows the planes to be only along one of the main x-y-z axes. I tried to write this myself by
extracting info from axes3d.plot_surface and the art3d, but unfortunately failed miserable.
Anyone go a clue?
Thanks a lot in advance!
Cheers,
Ulrich 

Gary Ruben wrote:
> The attached test.py
Oops. Here it is.
Gary R.

The attached test.py produces the attached png files with the Agg and 
WXAgg backends respectively. I've just started using the latest 
Enthought all-in-one beta which ships with matplotlib 0.91.2 and 
defaults to using WXAgg rather than TkAgg, so I've only just noticed 
this behaviour. I edited rcsetup.py, making the changes described in 
<http://www.nabble.com/Re%3A-pylab-axis-query-and-possible-bug-p14459682.html> 
to enable this to run.
The other backends I was able to test (PDF, PS, SVG, TkAgg) all produce 
correct output (same as test_Agg.png), so the scaling problem seems 
looks like it's specific to WXAgg.
Gary R.

I use the python-netcdf4 package in linux.
http://code.google.com/p/netcdf4-python/
You can make a netcdf3-compatible version of the library by installing 
with:
python setup-nc3.py install
Jordan