matplotlib-users Mailing List for matplotlib

Dear matplotlib users,

is it possible to pick an individual artist (Polygon) within a 
PolyCollection?

Cheers, Florian

Thanks for the comment, Jonathan.

Yeah, I did not expect aspect='equal' to work but I tried it anyway ;-)
Removing the extent argument indeed produces a very nice output but I have
not tried yet to also get the tick labels right. Instead, I have now
reverted back to matplotlib.pylab's subplots method and an extra axes for
the colorbar (see code below just in case somebody else can use that). With
some fiddling with spacing it looks ok now. It's just a mess to produce
differently sized figures but I probably won't need to do that.

Best, Matthias


<http://matplotlib.1069221.n5.nabble.com/file/n41080/2imshow%2B1colorbar.png> 
###########################################################
import numpy as np
import numpy.random as npr
from scipy.interpolate import griddata
import matplotlib.pyplot as plt

fig, axes = plt.subplots(nrows=1, ncols=2, figsize=[12,5])

for ax in axes:
    xmin, xmax = 0., 1.
    ymin, ymax = 0., 0.5
    zmin, zmax = -1., 1.
    
    # create random data:
    N = 100
    X = xmin + (xmax-xmin)*npr.random((N,))          # x_i in [0, 1]
    Y = ymin + (ymax-ymin)*npr.random((N,))          # y_i in [0, 0.5]
    Z = zmin + (zmax-zmin)*npr.random((N,))          # z_i in [-1, 1]
    
    # generate griddata for contour plot:
    numspaces = np.sqrt(N)
    xi = linspace(xmin, xmax, numspaces)
    yi = linspace(ymin, ymax, numspaces)
    zi = griddata((X, Y), Z, (xi[None,:], yi[:,None]), method='nearest')
    norm = matplotlib.colors.normalize(vmin=zmin, vmax=zmax)
    
    im = ax.imshow(zi,
        extent = [xmin,xmax,ymin,ymax],  
        norm = norm, 
        vmin = zmin,
        vmax = zmax, 
        origin = 'lower', 
        aspect = 2., 
        interpolation = 'nearest')
    
    ax.grid(False)
    ax.set_xlabel('x')
    ax.set_ylabel('y')
    
# add a colorbar:
fig.subplots_adjust(left=0.05, bottom=0.2, right=0.8, top=0.95, wspace=0.2,
hspace=0.2)
cbar_ax = fig.add_axes([0.85, 0.2, 0.03, 0.75])
fig.colorbar(im, cax=cbar_ax)
cbar_ax.set_ylabel('color level')
fig.subplots_adjust(left=0.05, bottom=0.2, right=0.8, top=0.95, wspace=0.2,
hspace=0.2)
##############################################################



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Squashed-axes-with-AxesGrid-tp40699p41080.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Gregorio,

I'm glad that helped.  I have not reported it on the issue tracker, but your case certainly has more of a bug with it (where it works sometimes, but not always), so I would recommend it.  

-Sterling

On May 17, 2013, at 2:10AM, Gregorio Bastardo wrote:

> Hi Sterling,
> 
> Thanks for the hint, using "line._legmarker" attribute solved the problem.
> 
> I see the reason behind, however I still consider this as an incorrect
> behaviour, since marker toggling works in case the line is originally
> added to the legend without marker (so legend line and marker do not
> behave like separated). Is it worth reporting on the mpl issue tracker
> (or have you done it that time)?
> 
> Gregorio
> 
> 2013/5/16 Sterling Smith <sm...@fu...>:
>> Gregorio,
>> 
>> I experienced a similar issue with trying to change the marker color.
>> 
>> See below the previous response from JJ for accessing the legend marker or using a proxy artist.
>> 
>> -Sterling
>> 

Matthias,

It's clear to me why apect='equal' doesn't work for you.  That option
means to give the axes equal scaling -- i.e., the ratio of length in
axis units to length in the plot is the same for both axes, so that an
axis that goes from 0 to 1 will be twice as long as one that goes from 0
to 0.5.  What is quite unclear to me is why aspect=2 should give a
result like it does.  You can get the right image, though wrong tick
labeling, if you omit the extent argument to imshow.

Jon

On Thu, 2013-05-16 at 12:43 -0700, Matthias Flor wrote:
> Hi all,
> 
> it seems that I am experiencing the same problem here with imshow (rather
> than scatter) and AxesGrid. But calling imshow with aspect=False does not do
> the trick for me.
> I am trying to have two imshow subplots next to each other and a single
> colorbar at the right. The data underlying the imshow's have different x-
> and y-ranges but I want the x- and y-axis to have an aspect ratio of 1 (i.e.
> each imshow should produce a square). I've tried aspect=False,
> aspect='equal', and explicitely setting aspect=2 which should be the correct
> value. See the three images below the code example.
> 
> I've also tried Grid instead of AxesGrid as suggested but I didn't manage to
> achieve good results with the colorbar in that case.
> 
> I'd appreciate any help,
> 
> Matthias
> 
> 
> 
> Here's a more or less minimal code example:
> #############################################
> import numpy as np
> import numpy.random as npr
> from scipy.interpolate import griddata
> import matplotlib.pyplot as plt
> from mpl_toolkits.axes_grid1 import AxesGrid
> 
> fig = figure(1, figsize=[12,10])
> 
> grid = AxesGrid(fig, 111,
>     nrows_ncols = (1, 2),
>     axes_pad = 0.2,
>     share_all = False,
>     label_mode = 'L',
>     cbar_location = 'right',
>     cbar_mode = 'single',
>     cbar_pad = 0.2
> )
> 
> for i in range(2):  
>     xmin, xmax = 0., 1.
>     ymin, ymax = 0., 0.5
>     zmin, zmax = -1., 1.
>     
>     # generate random data:
>     N = 100
>     X = xmin + (xmax-xmin)*npr.random((N,))          # x_i in [0, 1]
>     Y = ymin + (ymax-ymin)*npr.random((N,))          # y_i in [0, 0.5]
>     Z = zmin + (zmax-zmin)*npr.random((N,))          # z_i in [-1, 1]
>     
>     # generate griddata for imshow plot:
>     numspaces = np.sqrt(N)
>     xi = linspace(xmin, xmax, numspaces)
>     yi = linspace(ymin, ymax, numspaces)
>     zi = griddata((X, Y), Z, (xi[None,:], yi[:,None]), method='nearest')
>     norm = matplotlib.colors.normalize(vmin=zmin, vmax=zmax)
>     
>     ax = grid[i]
>     im = ax.imshow(zi,
>         extent = [xmin,xmax,ymin,ymax],  
>         norm = norm, 
>         vmin = zmin,
>         vmax = zmax, 
>         origin = 'lower', 
>         aspect = 'equal',     # or False, or 'auto', or 2, or ...
>         interpolation = 'nearest')
>     
>     ax.grid(False)
>     ax.set_xlabel('x')
>     ax.set_ylabel('y')
>     
> # add a colorbar:
> cbar = plt.colorbar(im, cax=grid.cbar_axes[0])
> cbar.ax.set_ylabel('color level')
> ############################################
> 
> And here are the three resulting images:
> 
> aspect='equal':
> <http://matplotlib.1069221.n5.nabble.com/file/n41075/equal.png> 
> 
> aspect=False:
> <http://matplotlib.1069221.n5.nabble.com/file/n41075/False.png> 
> 
> aspect=2.:
> <http://matplotlib.1069221.n5.nabble.com/file/n41075/two.png> 
> 
> 
> 
> --
> View this message in context: http://matplotlib.1069221.n5.nabble.com/Squashed-axes-with-AxesGrid-tp40699p41075.html
> Sent from the matplotlib - users mailing list archive at Nabble.com.
> 
> 

-- 
______________________________________________________________
Jonathan D. Slavin              Harvard-Smithsonian CfA
js...@cf...         60 Garden Street, MS 83
phone: (617) 496-7981           Cambridge, MA 02138-1516
 cell: (781) 363-0035           USA
______________________________________________________________

Hi Sterling,

Thanks for the hint, using "line._legmarker" attribute solved the problem.

I see the reason behind, however I still consider this as an incorrect
behaviour, since marker toggling works in case the line is originally
added to the legend without marker (so legend line and marker do not
behave like separated). Is it worth reporting on the mpl issue tracker
(or have you done it that time)?

Gregorio

2013/5/16 Sterling Smith <sm...@fu...>:
> Gregorio,
>
> I experienced a similar issue with trying to change the marker color.
>
> See below the previous response from JJ for accessing the legend marker or using a proxy artist.
>
> -Sterling
>