Hi,
It turns out to be a bug (new_vertical works, but new_horizontal does not).
To work around this
right after
cax = divider.new_horizontal(size="5%", pad=0.05, pack_start=True)
add these lines
locator = divider.new_locator(nx=0, ny=0)
cax.set_axes_locator(locator)
These two lines only need to be executed when "new_horizontal" is
called with pack_start=True.
The code below does some monkey patching to fix this bug. Use this if
it fits your need.
Regards,
-JJ
# work around new_horizontal bug
from mpl_toolkits.axes_grid.axes_divider import AxesDivider
if not hasattr(AxesDivider, "_new_horizontal_bug_fixed"):
new_horizontal_orig = AxesDivider.new_horizontal
def new_horizontal(self, size, pad=None, pack_start=False, **kwargs):
ax = new_horizontal_orig(self, size, pad=pad,
pack_start=pack_start, **kwargs)
if pack_start:
locator = self.new_locator(nx=0, ny=0)
ax.set_axes_locator(locator)
return ax
AxesDivider.new_horizontal = new_horizontal
AxesDivider._new_horizontal_bug_fixed = True
On Fri, Mar 5, 2010 at 9:05 AM, Thomas Robitaille
<thomas.robitai...@gmail.com> wrote:
> Hi Jae-Joon,
>
> Thanks! This is exactly what I needed. Putting the colorbar on the right or
> bottom works great - however, I am running into issues with trying to put the
> colorbar on the left or bottom (which, from my understanding, is controlled
> by using pack_start=True?). Should the following code work?
>
> import matplotlib.pyplot as mpl
> import numpy as np
> from mpl_toolkits.axes_grid import make_axes_locatable
>
> fig = mpl.figure()
> ax = fig.add_subplot(1,1,1)
> divider = make_axes_locatable(ax)
> cax = divider.new_horizontal(size="5%", pad=0.05, pack_start=True)
> fig.add_axes(cax)
> image = ax.imshow(np.random.random((100,100)))
> cb = fig.colorbar(image, cax=cax)
>
> Cheers,
>
> Thomas
>
> On Mar 4, 2010, at 10:28 PM, Jae-Joon Lee wrote:
>
>> see
>>
>> http://www.mail-archive.com/matplotlib-users@lists.sourceforge.net/msg15919.html
>>
>> axes_grid toolkit provides some helper function that utilizes
>> axes_locator (take a look at demo_locatable_axes_easy function in the
>> example below)
>>
>> http://matplotlib.sourceforge.net/examples/axes_grid/demo_axes_divider.html
>>
>> -JJ
>>
>>
>>
>>
>> On Thu, Mar 4, 2010 at 9:05 PM, Thomas Robitaille
>> <thomas.robitai...@gmail.com> wrote:
>>> Hi,
>>>
>>> I am trying to set up a colorbar that automatically resizes if I zoom in to
>>> an image (which changes the aspect ratio of the axes, so I want the
>>> colorbar to get resized too). Let's say I have two Axes instances, say ax
>>> (for the main image) and cax (for the colorbar). I can set up a callback if
>>> the view limits in one axes change, for example
>>>
>>> ax.callbacks.connect('xlim_changed', update_colorbar)
>>> ax.callbacks.connect('ylim_changed', update_colorbar)
>>>
>>> Now I can store a reference to cax inside ax:
>>>
>>> ax._cax = cax
>>>
>>> And I can now define update_colorbar so that it basically changes the
>>> position of cax:
>>>
>>> def update_colorbar(ax):
>>>
>>> # Get current position
>>> xmin = ax..get_position().xmin
>>> ...
>>>
>>> # Compute new colorbar position
>>> ...
>>>
>>> # Set new position
>>> ax._cax.set_position(...)
>>>
>>> # Return axes instance
>>> return ax
>>>
>>> Now the issue is that if I select a region of the image to zoom into, then
>>> as soon as I've selected the region, update_colorbar gets called, but by
>>> then, the aspect ratio of ax hasn't changed, and so the position I find
>>> when I do xmin = ax..get_position().xmin in update_colorbar is the *old*
>>> position of ax, not the new one. So the colorbar position is always one
>>> step behind compared to the main image axes.
>>>
>>> Can anyone think of any way that would avoid this issue, and to be able to
>>> use the *new* position of ax inside update_colorbar?
>>>
>>> Thanks in advance for any help,
>>>
>>> Thomas
>>> ------------------------------------------------------------------------------
>>> Download Intel® Parallel Studio Eval
>>> Try the new software tools for yourself. Speed compiling, find bugs
>>> proactively, and fine-tune applications for parallel performance.
>>> See why Intel Parallel Studio got high marks during beta.
>>> http://p.sf.net/sfu/intel-sw-dev
>>> _______________________________________________
>>> Matplotlib-users mailing list
>>> Matplotlib-users@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>
>
>
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-users
