matplotlib-users Mailing List for matplotlib (Page 4)

On 14 November 2012 21:05, Bror Jonsson <bro...@gm...> wrote:
> Dear all,
>
> I'm trying to to show where one set of values have NaN's on the contour
> plot of another set of values. I do this by creating a mask as such:
>
> fld = randn(4,4)
> fld[:2,:2] = np.nan
> mask[mask==0] = np.nan
> contourf(arange(4),arange(4),fld)
> contourf(arange(4),arange(4),mask)
>
> The problem is that the mask patch doesn't cover the empty space in the
> fld contour. Is there any way to make this happen?
>
> My ultimate goal is something like this:
>
> fld2 = randn(4,4)
> contourf(arange(4),arange(4),fld2)
> contourf(arange(4),arange(4),mask,[1,1], extend='both',
> colors='w', alpha=0.5)
>
> to present where fld has NaN's on the fld2 plot.
>
>
> Many thanks in advance!
>
> Bror Jonsson
>
Hello Bror,
It is not clear from your code snippets exactly what you are asking for.
Please can you post a full runnable example?
Ian Thomas

Thanks for reporting. It seems this file didn't make it over during the 
transition from Sourceforge to Github web hosting. It's been restored.
Mike
On 11/14/2012 04:45 PM, william ratcliff wrote:
> Hi! I was looking through the sample doc tutorial:
> http://matplotlib.org/sampledoc/
>
> and found that the link to the hard copy of the documentation is 
> missing. Is there a more recent link?
>
>
> Best,
> William
>
>
> ------------------------------------------------------------------------------
> Monitor your physical, virtual and cloud infrastructure from a single
> web console. Get in-depth insight into apps, servers, databases, vmware,
> SAP, cloud infrastructure, etc. Download 30-day Free Trial.
> Pricing starts from 795ドル for 25 servers or applications!
> http://p.sf.net/sfu/zoho_dev2dev_nov
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hi! I was looking through the sample doc tutorial:
http://matplotlib.org/sampledoc/
and found that the link to the hard copy of the documentation is missing.
Is there a more recent link?
Best,
William

Dear all,
I'm trying to to show where one set of values have NaN's on the contour plot of another set of values. I do this by creating a mask as such:
fld = randn(4,4)
fld[:2,:2] = np.nan
mask[mask==0] = np.nan
contourf(arange(4),arange(4),fld)
contourf(arange(4),arange(4),mask)
The problem is that the mask patch doesn't cover the empty space in the fld contour. Is there any way to make this happen?
My ultimate goal is something like this:
fld2 = randn(4,4)
contourf(arange(4),arange(4),fld2)
contourf(arange(4),arange(4),mask,[1,1], extend='both', 
 colors='w', alpha=0.5)
to present where fld has NaN's on the fld2 plot.
Many thanks in advance!
Bror Jonsson
"If you have a garden and a Library, You have everything you need." -Cicero
==============================================================
Associate Research Scholar
Princeton University
Department of Geosciences
113 Guyot Hall
Princeton, NJ 08544-1003
USA
AIM, Skype, gTalk: brorfred
Phone: +1-617-818-1096

Hello again
> expecting the transparency to "stop" at the layer below the plot and
> therefore see the.
Sorry, I meant "therefore see the panel".
-- 
Sylvain

Hello
I would like some help to understand a problem with matplotlib and 
wxpython.
I am developping a GUI where my plots are embedded on wxPanels on a 
wxNotebook (tabs). Under Windows, some themes don't use a single colour 
but a gradient as the tab background. Therefore, I'd like to make the 
background of my plots transparent.
Under Windows XP (whatever the theme), when I set the facecolor of the 
plot to 'none', the plot background becomes transparent, but the parts 
of the panel and of the notebook below as well, and I end up seeing 
other windows behind my GUI or the Windows desktop. I was expecting the 
transparency to "stop" at the layer below the plot and therefore see 
the.
I did a second experiment, where I overlayed two plots. The top one is 
larger than the one below. I make the top one partially transparent, to 
see the one below. The transparency is "stopped" in the area of the 
inferior plot, I see the desktop on the remaining parts, and where there 
is no plot the background of my panel.
I'm attaching the code for the second experiment.
I'm running XP 32bits with the Classic theme, python 2.7.3, matplotlib 
1.2.0 and wxpython 2.9.4-msw.
Thanks for your help
-- 
Sylvain

Hi All,
Hoping someone can help me get a definitive answer to this question.
Is draw_if_interactive bad to have in library plotting code?
Based on this thread [1], we've been working under the assumption that
calling draw_if_interactive in plotting code is bad. Though I'm
skeptical that this is the takeaway that we should have. I also asked
this question on the IPython mailing list [2] since the recommendation
comes from their type of usage, but I'm still not clear.
I'll repeat the gist of the question here.
We have plotting functions that are designed to update a given axes. I
often work in interactive mode, and I'd like it if these functions
updated my axes in the way that I expect (and an R user doing plotting
in Python would expect). But now I'm forced to litter my user scripts
with draw_if_interactive after I call a function I expect to update a
plot - say updating a scatter plot with a regression line. Would be
harmful to just include these draw_if_interactive calls in our plot
functions. To be clear, I never have to call show or draw because I'm
working in interactive mode, so the recommendation to just call show()
at the end of a script is not what I want.
My understanding of the pitfalls is 1) there's a performance hit to
calling draw instead of just making one call. This is moot because
we're only calling draw_if_interactive - so we assume the user is
working interactively and actually wants to do the drawing and doesn't
care about the performance hit. And 2) we are assuming that the user
has imported and is using pyplot and there are possible side effects.
A user wouldn't be using pyplot in a GUI or in some sort of embedded
plotting framework. However, my intuition says that if this is the
case, draw_if_interactive won't do anything because interactive will
be False in these cases.
Can someone please help clear this up? Thanks,
Skipper
[1] https://groups.google.com/forum/#!msg/pystatsmodels/biNlCvJPNNY/BT7bQJmOa1cJ
[2] http://python.6.n6.nabble.com/IPython-User-using-matplotlib-draw-if-interactive-in-library-code-td4991275.html

On Tue, Nov 13, 2012 at 8:38 AM, David Brunell <qua...@gm...> wrote:
> Hello, I have what I hope is a simple question. When producing a
> figure/plot, I have a window which pops up with the figure inside and a few
> tool buttons along the bottom, including "Zoom to rectangle." Clicking the
> Zoom tool button, I'm presented with a black crosshair to select my zoom
> rectangle. Many of the images I work with are predominantly black; is
> there any way to change the color of the crosshair so as to make it more
> visible? Thanks.
Hi David,
Unfortunately, those widgets are backend specific, so changing them is not
trivial in general, since each toolkit has its own way of specifying the
cursor. With that said, you can try to figure out if there's a way to do it
for your backend `import matplotlib as mpl; mpl.get_backend()` will tell
you which backend you're using, and then you'll need to look in the
relevant source code for where the cursor is define.
If you don't know where your matplotlib code lives, you can the path of the
relevant files using this:
 import matplotlib.backends as b
 import os
 os.path.dirname(b.__file__)
There, you'll find files for all of the backends, and the `cursord`
dictionary in most of them is what specifies how the widgets look. I'm not
sure which toolkits allow one to change the color of the default cursors,
but some of them allow you to even specify your own color images, so it
should be possible.
An alternative, of course, would be to change the colormap you're plotting
with, or add an alpha value to the images you're plotting so that the black
widgets can be seen. Maybe it's inelegant, but looks like the path of least
resistance...
best,
-- 
Paul Ivanov
314 address only used for lists, off-list direct email at:
http://pirsquared.org | GPG/PGP key id: 0x0F3E28F7

In article 
<CAJSg89LEe=HCx...@ma...>,
 Alexey Shamrin <sh...@gm...> 
 wrote:
> Thank you for 1.2.0 release!
> 
> Could you please make it clear that matplotlib requires
> python.org-Python sourceforge.net-NumPy? Telling about it during
> installation would be great.
This is described in three places:
- The description of the file on the download page
- The name of the file on the download page
- The ReadMe file in the binary installer
Note that the official binary installers for numpy and scipy are also 
for python.org python, and as far as I know they do no more than the 
matplotlib installer as far as informing the user of this fact.
It is bdist_mpkg that makes these installers, and it could be better 
about checking compatibility. But that is a known issue.
I don't know about messages about "system python", though that vaguely 
rings a bell as a bdist_mpkg issue.
I'll add information about numpy to the ReadMe for future binary 
installers. Aside from that, I believe I've done everything I reasonably 
can to clarify the requirements for the binary installer.
-- Russell

All,
I am attempting to plot a base map with extents which are outside of the
figure using the following code:
#Map
lon_min = -101.5
lon_max = -94.5
lat_min = -32.5
lat_max = -27.5
m = Basemap(projection='aeqd',llcrnrlat=lat_min,urcrnrlat=lat_max,
 llcrnrlon=lon_min,urcrnrlon=lon_max,lon_0=-97.7328,
lat_0=-30.0906,resolution=None, rsphere=(1737400.0,1737400.0))
 #Read the input image
input_basemap = gdal.Open('Mare_Orientale_Volc_AzEqui.png')
input_band = input_basemap.GetRasterBand(1)
bmap = input_band.ReadAsArray()
 #The bounds of the input image using gdalinfo
LL = (-204690.290, -162184.543)
UR = (200909.710, 176915.457)
I know that I need to use pcolormesh() to get my map visualized. I also
believe that I need to use the transform_scalar function to get from pixel
space to map projected space. My input image is not in Lat/Lon, but in
pixel space. Any suggestions on getting my image to display in projected
space?
Best,
Jay

Hello, I have what I hope is a simple question. When producing a
figure/plot, I have a window which pops up with the figure inside and a few
tool buttons along the bottom, including "Zoom to rectangle." Clicking the
Zoom tool button, I'm presented with a black crosshair to select my zoom
rectangle. Many of the images I work with are predominantly black; is
there any way to change the color of the crosshair so as to make it more
visible? Thanks.

2012年11月13日 Benjamin Root <ben...@ou...>
>
>
> On Tue, Nov 13, 2012 at 6:16 AM, Francesco Montesano <
> fra...@gm...> wrote:
>
>> Dear matplolibers,
>>
>> when dealing with multi-axes plot sometimes would be nice to use
>> figure-wide x and y labels.
>> On the web I've found some suggestion on how to do this, but I found
>> no solution valid in the general case and that integrate in the
>> matplotlib ecosystem.
>> The ideal would be to have a "set_xlabel" and "set_ylabel" method in the
>> Figure class, with the same api of the corresponding Axes methods.
>>
>> As a proof of concept I've written a class derived from Figure , which
>> implements the two methods simply adding a horizontal (vertical) text below
>> (left of) the lowest (leftmost) axes.
>> The class together with a short example is attached.
>> I'm aware that the current implementation is really poor (no integration
>> with tight_layout, the padding must be adjusted by hand, a problem in
>> particular for the y label).
>>
>> The best is to use "self.xaxis.set_label_text(xlabel, fontdict,
>> **kwargs)" as in the Axis set_xlabel (as I gather this create a label that
>> is rendered in the correct position accounting for ticklabels, ticks,
>> tight_layout, etc). To do this one would have to create:
>>
>> - a figure-wide invisible axes that encloses all the other
>> axes/subplots, and whose dimension has to be updated every time a new
>> axis/subplot is added (this should be easily done) with only the label
>> visible. This could also allow to use axis features, like twin axis.
>> - just the required axis (invisible) that hosts the labels. I think
>> that this approach is less demanding computationally, but I don't know how
>> much sense have two axis not attached to axes.
>>
>> Any suggestions/hints on how to implement these methods in a better way
>> is very welcome.
>>
>> If there is no opposition, later in the day I'll submit PR on github with
>> the two new method and see if we can get something out of this idea.
>>
>> Cheers,
>> Francesco
>>
>>
> I am not exactly sure if this is the same as what you are thinking, but
> the axes objects have a "label_outer()" method that would turn on and off
> the visibility of various axis components based on their location in a
> subplot grid. You call it for each axes in a subplot grid.
>
> Cheers!
> Ben Root
>
> Hi Ben,
sorry that I'm not being clear. My scope is to have the a unique x and y
label as in "figure_label.png" instead a x and y label for each outer axes
as in "axes_label.png".
This could be done not writing axes labels and then using a simple text on
the left and bottom, but I think that set_[xy]label method in class Figure
(as the method legend) is much neater.
Besides can improve readability of plots with lots of panels showing the
same quantities.
Is it clearer now?
Francesco

On Tue, Nov 13, 2012 at 6:16 AM, Francesco Montesano <
fra...@gm...> wrote:
> Dear matplolibers,
>
> when dealing with multi-axes plot sometimes would be nice to use
> figure-wide x and y labels.
> On the web I've found some suggestion on how to do this, but I found
> no solution valid in the general case and that integrate in the
> matplotlib ecosystem.
> The ideal would be to have a "set_xlabel" and "set_ylabel" method in the
> Figure class, with the same api of the corresponding Axes methods.
>
> As a proof of concept I've written a class derived from Figure , which
> implements the two methods simply adding a horizontal (vertical) text below
> (left of) the lowest (leftmost) axes.
> The class together with a short example is attached.
> I'm aware that the current implementation is really poor (no integration
> with tight_layout, the padding must be adjusted by hand, a problem in
> particular for the y label).
>
> The best is to use "self.xaxis.set_label_text(xlabel, fontdict, **kwargs)"
> as in the Axis set_xlabel (as I gather this create a label that is rendered
> in the correct position accounting for ticklabels, ticks, tight_layout,
> etc). To do this one would have to create:
>
> - a figure-wide invisible axes that encloses all the other
> axes/subplots, and whose dimension has to be updated every time a new
> axis/subplot is added (this should be easily done) with only the label
> visible. This could also allow to use axis features, like twin axis.
> - just the required axis (invisible) that hosts the labels. I think
> that this approach is less demanding computationally, but I don't know how
> much sense have two axis not attached to axes.
>
> Any suggestions/hints on how to implement these methods in a better way is
> very welcome.
>
> If there is no opposition, later in the day I'll submit PR on github with
> the two new method and see if we can get something out of this idea.
>
> Cheers,
> Francesco
>
>
I am not exactly sure if this is the same as what you are thinking, but the
axes objects have a "label_outer()" method that would turn on and off the
visibility of various axis components based on their location in a subplot
grid. You call it for each axes in a subplot grid.
Cheers!
Ben Root

Dear matplolibers,
when dealing with multi-axes plot sometimes would be nice to use
figure-wide x and y labels.
On the web I've found some suggestion on how to do this, but I found
no solution valid in the general case and that integrate in the
matplotlib ecosystem.
The ideal would be to have a "set_xlabel" and "set_ylabel" method in the
Figure class, with the same api of the corresponding Axes methods.
As a proof of concept I've written a class derived from Figure , which
implements the two methods simply adding a horizontal (vertical) text below
(left of) the lowest (leftmost) axes.
The class together with a short example is attached.
I'm aware that the current implementation is really poor (no integration
with tight_layout, the padding must be adjusted by hand, a problem in
particular for the y label).
The best is to use "self.xaxis.set_label_text(xlabel, fontdict, **kwargs)"
as in the Axis set_xlabel (as I gather this create a label that is rendered
in the correct position accounting for ticklabels, ticks, tight_layout,
etc). To do this one would have to create:
 - a figure-wide invisible axes that encloses all the other
 axes/subplots, and whose dimension has to be updated every time a new
 axis/subplot is added (this should be easily done) with only the label
 visible. This could also allow to use axis features, like twin axis.
 - just the required axis (invisible) that hosts the labels. I think that
 this approach is less demanding computationally, but I don't know how much
 sense have two axis not attached to axes.
Any suggestions/hints on how to implement these methods in a better way is
very welcome.
If there is no opposition, later in the day I'll submit PR on github with
the two new method and see if we can get something out of this idea.
Cheers,
Francesco

I've just run into this problem myself. I think I've tracked down the
offending code to lines 1910-1916 of
/usr/lib/pymodules/pythn2.7/matplotlib/axes.py
this is within the function definition for draw()
-----
if self.axison and not inframe:
 if self._axisbelow:
 self.xaxis.set_zorder(0.5)
 self.yaxis.set_zorder(0.5
 else:
 self.xaxis.set_zorder(2.5)
 self.yaxis.set_zorder(2.5)
-----
In particular, the zorder of 2.5 is being set by lines 1915-1916 (the last
of the lines copied above.
Seems the source of the bug to me, but I have no idea what the procedure is
for getting it logged and fixed.
-Michael Woods

 http://shrinktofit.dk/wp-content/plugins/ugoogle.html

If you're using pyplot.specgram (i.e. "from pylab import *;
specgram(...)"), note that the plot is in dB, hence the negative
values.
I'm surprised this fact isn't mentioned in the documentation:
http://matplotlib.org/api/pyplot_api.html?highlight=specgram#matplotlib.pyplot.specgram
However, when in doubt, look at the code.
On Mon, Nov 12, 2012 at 12:28 PM, Paul Anton Letnes
<pau...@gm...> wrote:
> Heh,
>
> that's funny. Now then, why do my plots come out with negative values all over the place? That's why I started digging around. After all, X * conj(X) should be equal to the absolute square of X, right?
>
> Paul
>
>
> On 12. nov. 2012, at 21:00, G Jones wrote:
>
>> Hi,
>> If you trace back into the code further, you will see that the Pxx is
>> computed as X = fft(x), Pxx = X * conj(X) which is real, but the data
>> type will be complex with a ~0 imaginary part (up to floating point
>> precision). Thus the Pxx.real is just to ensure that the resulting
>> data type is real instead of complex to save memory.
>> Glenn
>>
>> On Mon, Nov 12, 2012 at 11:42 AM, Paul Anton Letnes
>> <pau...@gm...> wrote:
>>> Hi,
>>>
>>> not 100% sure this is a bug, but here goes:
>>>
>>> In file matplotlib/lib/matplotlib/mlab.py, the functions psd (power spectral density) and specgram returns the real part of the fourier transform.
>>> % grep -n Pxx.real mlab.py
>>> 390: return Pxx.real,freqs
>>> 470: Pxx = Pxx.real #Needed since helper implements generically
>>> (git version 4f902fac1c5bf267e3fdeb4c2045926d7498e85a, cloned from github today)
>>>
>>> This all means that the specgram plot routine yields the real part of the Fourier transform, rather than its absolute square (forgetting normalization for simplicity of discussion). The definition of the PSD is that it is the absolute square of the Fourier transform:
>>> https://en.wikipedia.org/wiki/Power_spectral_density#Energy_spectral_density
>>>
>>> Hence, I believe this is a bug which should be fixed.
>>>
>>> Cheers
>>> Paul
>>> ------------------------------------------------------------------------------
>>> Monitor your physical, virtual and cloud infrastructure from a single
>>> web console. Get in-depth insight into apps, servers, databases, vmware,
>>> SAP, cloud infrastructure, etc. Download 30-day Free Trial.
>>> Pricing starts from 795ドル for 25 servers or applications!
>>> http://p.sf.net/sfu/zoho_dev2dev_nov
>>> _______________________________________________
>>> Matplotlib-users mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hi,
If you trace back into the code further, you will see that the Pxx is
computed as X = fft(x), Pxx = X * conj(X) which is real, but the data
type will be complex with a ~0 imaginary part (up to floating point
precision). Thus the Pxx.real is just to ensure that the resulting
data type is real instead of complex to save memory.
Glenn
On Mon, Nov 12, 2012 at 11:42 AM, Paul Anton Letnes
<pau...@gm...> wrote:
> Hi,
>
> not 100% sure this is a bug, but here goes:
>
> In file matplotlib/lib/matplotlib/mlab.py, the functions psd (power spectral density) and specgram returns the real part of the fourier transform.
> % grep -n Pxx.real mlab.py
> 390: return Pxx.real,freqs
> 470: Pxx = Pxx.real #Needed since helper implements generically
> (git version 4f902fac1c5bf267e3fdeb4c2045926d7498e85a, cloned from github today)
>
> This all means that the specgram plot routine yields the real part of the Fourier transform, rather than its absolute square (forgetting normalization for simplicity of discussion). The definition of the PSD is that it is the absolute square of the Fourier transform:
> https://en.wikipedia.org/wiki/Power_spectral_density#Energy_spectral_density
>
> Hence, I believe this is a bug which should be fixed.
>
> Cheers
> Paul
> ------------------------------------------------------------------------------
> Monitor your physical, virtual and cloud infrastructure from a single
> web console. Get in-depth insight into apps, servers, databases, vmware,
> SAP, cloud infrastructure, etc. Download 30-day Free Trial.
> Pricing starts from 795ドル for 25 servers or applications!
> http://p.sf.net/sfu/zoho_dev2dev_nov
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hi,
not 100% sure this is a bug, but here goes:
In file matplotlib/lib/matplotlib/mlab.py, the functions psd (power spectral density) and specgram returns the real part of the fourier transform.
% grep -n Pxx.real mlab.py
390: return Pxx.real,freqs
470: Pxx = Pxx.real #Needed since helper implements generically
(git version 4f902fac1c5bf267e3fdeb4c2045926d7498e85a, cloned from github today)
This all means that the specgram plot routine yields the real part of the Fourier transform, rather than its absolute square (forgetting normalization for simplicity of discussion). The definition of the PSD is that it is the absolute square of the Fourier transform:
https://en.wikipedia.org/wiki/Power_spectral_density#Energy_spectral_density
Hence, I believe this is a bug which should be fixed.
Cheers
Paul

On 29 October 2012 14:50, Daryl Herzmann <ak...@gm...> wrote:
> I've been attempting to get basemap to clip a contourf display. I have not
> had any luck! Attached is a self contained example. Could somebody kindly
> point out what I am doing wrong!?!
>
Hi Daryl,
You were almost there. Remove the call to mask_outside_polygon and replace
it with
for collection in cs.collections:
 collection.set_clip_path(patch)
Attached is your corrected example and the output produced.
Ian

On Mon, Nov 12, 2012 at 11:43 AM, Nils Wagner <ni...@go...>wrote:
> Hi all,
>
> how can I hide ticks and/or labels in the presence of sharex=ax. Only
> the last subplot 313 should have ticks and labels.
>
> import matplotlib.pyplot as plt
> fig=plt.figure(0,figsize=(16,24))
>
> ax = fig.add_subplot(311)
> ax.set_xticks([])
> ax.set_xticklabels('')
>
> ax1 = fig.add_subplot(312,sharex=ax)
> ax1.set_xticks([])
> ax1.set_xticklabels('')
>
> ax2 = fig.add_subplot(313,sharex=ax)
> ax2.set_xticks(ind+width)
> ax2.set_xticklabels( contname,rotation='90',fontsize=8)
>
> Nils
>
>
ax.label_outer() will set the appropriate visibility settings for the
particular subaxes depending on where it is in the grid. Just call it for
each subplot being shared and you are good to go.
Cheers!
Ben Root

Hi all,
how can I hide ticks and/or labels in the presence of sharex=ax. Only
the last subplot 313 should have ticks and labels.
import matplotlib.pyplot as plt
fig=plt.figure(0,figsize=(16,24))
ax = fig.add_subplot(311)
ax.set_xticks([])
ax.set_xticklabels('')
ax1 = fig.add_subplot(312,sharex=ax)
ax1.set_xticks([])
ax1.set_xticklabels('')
ax2 = fig.add_subplot(313,sharex=ax)
ax2.set_xticks(ind+width)
ax2.set_xticklabels( contname,rotation='90',fontsize=8)
Nils

On Sun, Nov 11, 2012 at 1:42 PM, Andrew Dawson <da...@at...> wrote:
> Hi
>
> I'm trying to plot the trajectory of a particle in 3d using mplot3d. I
> tried to follow the example of an animated 3d plot on the matplotlib
> website but I'm having trouble with the updating of the data point being
> plotted at each frame. Does anyone know how to do this?
>
> So far I have:
>
> import numpy as np
> import matplotlib.pyplot as plt
> from mpl_toolkits.mplot3d.axes3d import Axes3D
> from matplotlib.animation import FuncAnimation
>
>
> def update_plot(num, data, sc):
> sc.set_array(data[num])
> return sc
>
>
> def main():
> numframes = 2
> data = np.random.rand(10, 3)# a (time, position) array
>
> fig = plt.figure()
> ax = fig.add_subplot(111, projection='3d')
>
> ix, iy, iz = data[0]
> sc = ax.scatter(ix, iy, iz, c='k')
>
> ani = FuncAnimation(fig, update_plot, frames=numframes,
> fargs=(data,sc))
> plt.show()
>
>
> if __name__ == '__main__':
> main()
>
>
> This just changes the color of the initial marker. I also tried to use
> sc.set_3d_properties but it is not clear to me what the arguments should be
> here, I kept getting an error... If anyone has done this before I'd love to
> see an example.
>
> Thanks,
> Andrew
>
>
Andrew,
For scatter objects (which are PatchCollection), the get/set_data() refers
to the scalar mappable part of things, which is why the color kept
changing. It does not seem to be an easy way to adjust the position data
for a Patch3DCollection (or a Line3DCollection for that matter...). I
would suggest filing a feature request about that on github. In coming up
with an example for your use-case, I have come across a couple of minor
bugs in mplot3d that I am going to need to resolve as well. In the
meantime, I think the following version of the code:
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d.axes3d import Axes3D
from matplotlib.animation import FuncAnimation
def update_plot(num, data, sc):
 print sc._offsets3d
 sc._offsets3d = data[num]
 return sc
def main():
 numframes = 10
 data = np.random.rand(numframes, 3, 1)# a (time, position) array
 fig = plt.figure()
 ax = fig.add_subplot(111, projection='3d')
 ix, iy, iz = data[0]
 sc = ax.scatter(ix, iy, iz, c='k')
 ani = FuncAnimation(fig, update_plot, frames=numframes,
 fargs=(data,sc))
 plt.show()
if __name__ == '__main__':
 main()
Essentially, there is no nice way to set the 3d position data, and the
easiest way is to just go to the internal _offsets3d variable. Second,
there seems to be an issue with array/scalar data in Patch3DCollection that
I had to make the random number generation be 3D, rather than 2D as you
originally had it.
Cheers!
Ben Root

Hi
I'm trying to plot the trajectory of a particle in 3d using mplot3d. I
tried to follow the example of an animated 3d plot on the matplotlib
website but I'm having trouble with the updating of the data point being
plotted at each frame. Does anyone know how to do this?
So far I have:
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d.axes3d import Axes3D
from matplotlib.animation import FuncAnimation
def update_plot(num, data, sc):
 sc.set_array(data[num])
 return sc
def main():
 numframes = 2
 data = np.random.rand(10, 3)# a (time, position) array
 fig = plt.figure()
 ax = fig.add_subplot(111, projection='3d')
 ix, iy, iz = data[0]
 sc = ax.scatter(ix, iy, iz, c='k')
 ani = FuncAnimation(fig, update_plot, frames=numframes,
 fargs=(data,sc))
 plt.show()
if __name__ == '__main__':
 main()
This just changes the color of the initial marker. I also tried to use
sc.set_3d_properties but it is not clear to me what the arguments should be
here, I kept getting an error... If anyone has done this before I'd love to
see an example.
Thanks,
Andrew

Hello all,
It turns out the command-(APPLE) -TAB does work to change the focus from
page to page but only if the mouse is centered on top of the actual tab
area . In windows mouse focus anywhere on the figure allows CTRL-TAB page
flipping. But in OSX it only works when focussed on the acutal tab
rectangle.
I checked what this was the behavior with the official wxpython
wxAUINotebook demo application as well.
Thanks
Hari
On Sun, Oct 21, 2012 at 7:25 AM, hari jayaram <ha...@gm...> wrote:
> Hi I am using
> wxpython : 2.9.4.0
> matplotlib : 1.3
> osx Lion
>
> In my application I have a number of matplotlib figure objects, one on
> each page of the wx.aui.AuiNotebook .The pages are each a figure and
> arranged as tabs on the top of the wxpython frame like embedding in wx5
> example from the matplotlib gallery.
>
> On Windows I can navigate from page to page of the Notebook using
> CTRL-TAB and CTRL-SHIFT-TAB.
>
> However on OSX -Lion , neither the CTRL-TAB, nor Alt/Tab navigate from
> page to page.
>
> Instead what happens is that the "mouse selection" moves from icon to icon
> i.e from the "Home" to the "Pan-zoom " icon on the bottom of the matplotlib
> figure. The wxAuiNotebook is oblvious of these mouse events.
>
> Does anyone know how to restore the windows os behavior where CTRL-TAB
> changes the page of the Notebook on OSX. How do I prevent the matplotlib
> figure object from intercepting these events.
>
> Thanks
> Hari
>
>
>
>
>
>