matplotlib-users — Discussion related to using matplotlib

Re: [Matplotlib-users] griddata fails

[Shahar Shani-K. <ka...@po...>]

Yes, you are absolutely correct. I did not realize that I did not actually evaluate the function over a grid, makes sense that interpolation fails. I thought that since I created the two axis vectors the function evaluation occurs over the entire domain, meshgrid is what I was missing.

thanks,
Shahar
On Jan 9, 2013, at 8:45 PM, Ian Thomas wrote:

> On 9 January 2013 09:32, Shahar Shani-Kadmiel <ka...@po...> wrote:
> Hi,
> 
> I'm trying to contour some data that I have and the griddata line fails. I tried running it on some synthetically generated data and I get the same IndexError. Any Ideas?
> 
> Here is the example with the synthetic data:
> 
> x = y = arange(-10,10,0.01)
> 
> z = x**2+y**3
> 
> xi = yi = linspace(-10.1, 10.1, 100)
> 
> zi = griddata(x, y, z, xi, yi)
> ---------------------------------------------------------------------------
> IndexError                                Traceback (most recent call last)
> <ipython-input-52-0458ab6ea672> in <module>()
> ----> 1 zi = griddata(x, y, z, xi, yi)
> 
> /Library/Frameworks/Python.framework/Versions/7.3/lib/python2.7/site-packages/matplotlib/mlab.py in griddata(x, y, z, xi, yi, interp)
>    2766             xi,yi = np.meshgrid(xi,yi)
>    2767         # triangulate data
> -> 2768         tri = delaunay.Triangulation(x,y)
> 
> Hello Shahar,
> 
> I think that your simple example is probably not what you intended.  Your (x,y) points are all defined on the straight line from (-10,-10) to (10,10).  The Delaunay triangulation of these points (which is what griddata does) is not very interesting!  Perhaps you wanted (x,y) defined on the 2D grid from (-10,-10) to (10,10), in which case you should follow the x = y ... line with, for example:
>     x, y = meshgrid(x, y)
> (see numpy.meshgrid for further details).
> 
> You may still obtain the same IndexError, and the traceback shows this is happening in the delaunay.Triangulation function call.  The matplotlib delaunay package is not particularly robust, and can have problems handling regularly-spaced data points.  The griddata documentation explains some of this, see http://matplotlib.org/api/mlab_api.html#matplotlib.mlab.griddata.
> 
> To avoid the problem, the griddata documentation explains one possible way that uses the natgrid algorithm.  A simpler solution that I often use is to add a very small amount of noise to my regularly-spaced (x,y) points using the numpy.random module.  I can give more details if you wish.
> 
> Ian

[Matplotlib-users] (no subject)

[Neacsa B. V. <nea...@ya...>]

http://krishnacement.com/boycharityleemacdonald/

Re: [Matplotlib-users] mathtext and fonts under Windows 8

[Paul H. <pmh...@gm...>]

Sounds like it might have something to do with your Latex installation (if
any) or the barebones Latex-rendering done by MPL alone. Namely, they
simply don't have the characters for mathematical Arial available.

Not too sure though. Hopefully someone more knowledgeable responds.
-paul


On Tue, Jan 8, 2013 at 9:31 PM, CAB <ca...@ya...> wrote:

> Hi, All,
>
> I am encountering a thorny problem when trying to run matplotlib under
> Windows 8.  If I label an axis using a command like
>
> ax.set_ylabel(r'time (s)', name='Arial'),
>
> all is well.  But if  try to add mathtext to that, as in
>
> ax.set_ylabel(r'time ($s$)', name='Arial'),
>
> mathtext.py throws an error (a very long stream) ending in "RuntimeError:
> Face has no glyph names".  If I remove the "name='Arial'" above and let the
> program default to Bitstream Vera Sans, the mathtext works.
>
> This problem does not occur under Windows 7 or XP; only under two
> different Windows 8 installations.  Any ideas what's going on?
>
> Chad
>
>
> ------------------------------------------------------------------------------
> Master Java SE, Java EE, Eclipse, Spring, Hibernate, JavaScript, jQuery
> and much more. Keep your Java skills current with LearnJavaNow -
> 200+ hours of step-by-step video tutorials by Java experts.
> SALE $49.99 this month only -- learn more at:
> http://p.sf.net/sfu/learnmore_122612
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

[Matplotlib-users] Incorrect ylabel placement in twinx

[Goyo <goy...@gm...>]

I'm using matplotlib master from
https://launchpad.net/%7Etakluyver/+archive/matplotlib-daily, ubuntu
12.10 and python 2.7 and sometimes I get misplaced y labels for twinx
plots. I file a bug at github with a sample script and figure images:
https://github.com/matplotlib/matplotlib/issues/1608
A guy commented there saying he can't replicate the issue. I wonder if
anyone using ubuntu 12.10 can try my code with the matplotlib version
from that repo and share the output.

Best regards

Goyo

Re: [Matplotlib-users] griddata fails

[Ian T. <ian...@gm...>]

On 9 January 2013 09:32, Shahar Shani-Kadmiel <ka...@po...>wrote:

> Hi,
>
> I'm trying to contour some data that I have and the griddata line fails. I
> tried running it on some synthetically generated data and I get the same
> IndexError. Any Ideas?
>
> Here is the example with the synthetic data:
>
> x = y = arange(-10,10,0.01)
>
> z = x**2+y**3
>
> xi = yi = linspace(-10.1, 10.1, 100)
>
> zi = griddata(x, y, z, xi, yi)
> ---------------------------------------------------------------------------
> IndexError                                Traceback (most recent call last)
> <ipython-input-52-0458ab6ea672> in <module>()
> ----> 1 zi = griddata(x, y, z, xi, yi)
>
> /Library/Frameworks/Python.framework/Versions/7.3/lib/python2.7/site-packages/matplotlib/mlab.py
> in griddata(x, y, z, xi, yi, interp)
>    2766             xi,yi = np.meshgrid(xi,yi)
>    2767         # triangulate data
> -> 2768         tri = delaunay.Triangulation(x,y)
>

Hello Shahar,

I think that your simple example is probably not what you intended.  Your
(x,y) points are all defined on the straight line from (-10,-10) to
(10,10).  The Delaunay triangulation of these points (which is what
griddata does) is not very interesting!  Perhaps you wanted (x,y) defined
on the 2D grid from (-10,-10) to (10,10), in which case you should follow
the x = y ... line with, for example:
    x, y = meshgrid(x, y)
(see numpy.meshgrid for further details).

You may still obtain the same IndexError, and the traceback shows this is
happening in the delaunay.Triangulation function call.  The matplotlib
delaunay package is not particularly robust, and can have problems handling
regularly-spaced data points.  The griddata documentation explains some of
this, see http://matplotlib.org/api/mlab_api.html#matplotlib.mlab.griddata.

To avoid the problem, the griddata documentation explains one possible way
that uses the natgrid algorithm.  A simpler solution that I often use is to
add a very small amount of noise to my regularly-spaced (x,y) points using
the numpy.random module.  I can give more details if you wish.

Ian

Re: [Matplotlib-users] Interactive mode broken in Qt4Agg backend?

[Yannick C. <yan...@la...>]

Hi,

Yannick Copin <yannick.copin@...> writes:
> Is it a bug in Qt4Agg, or an unimplemented feature? Is there any way 
> other than switching back to GTKAgg specifically?

just a follow-up and conclusion: I create a matplotlib issue for this
(https://github.com/matplotlib/matplotlib/issues/1646), and the use of
pyplot.pause in place of pyplot.draw seems to help.

Cheers,

Yannick

[Matplotlib-users] griddata fails

[Shahar Shani-K. <ka...@po...>]

Hi,

I'm trying to contour some data that I have and the griddata line fails. I tried running it on some synthetically generated data and I get the same IndexError. Any Ideas?

Here is the example with the synthetic data:

x = y = arange(-10,10,0.01)

z = x**2+y**3

xi = yi = linspace(-10.1, 10.1, 100)

zi = griddata(x, y, z, xi, yi)
---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-52-0458ab6ea672> in <module>()
----> 1 zi = griddata(x, y, z, xi, yi)

/Library/Frameworks/Python.framework/Versions/7.3/lib/python2.7/site-packages/matplotlib/mlab.py in griddata(x, y, z, xi, yi, interp)
   2766             xi,yi = np.meshgrid(xi,yi)
   2767         # triangulate data
-> 2768         tri = delaunay.Triangulation(x,y)
   2769         # interpolate data
   2770         if interp == 'nn':

/Library/Frameworks/Python.framework/Versions/7.3/lib/python2.7/site-packages/matplotlib/delaunay/triangulate.py in __init__(self, x, y)
     88             self.triangle_neighbors = delaunay(self.x, self.y)
     89 
---> 90         self.hull = self._compute_convex_hull()
     91 
     92     def _collapse_duplicate_points(self):

/Library/Frameworks/Python.framework/Versions/7.3/lib/python2.7/site-packages/matplotlib/delaunay/triangulate.py in _compute_convex_hull(self)
    113 
    114         edges = {}
--> 115         edges.update(dict(zip(self.triangle_nodes[border[:,0]][:,1],
    116                               self.triangle_nodes[border[:,0]][:,2])))
    117         edges.update(dict(zip(self.triangle_nodes[border[:,1]][:,2],

IndexError: invalid index

[Matplotlib-users] mathtext and fonts under Windows 8

[CAB <ca...@ya...>]

Hi, All,

I am encountering a thorny problem when trying to run matplotlib under Windows 8.  If I label an axis using a command like

ax.set_ylabel(r'time (s)', name='Arial'),

all is well.  But if  try to add mathtext to that, as in

ax.set_ylabel(r'time ($s$)', name='Arial'),

mathtext.py throws an error (a very long stream) ending in "RuntimeError: Face has no glyph names".  If I remove the "name='Arial'" above and let the program default to Bitstream Vera Sans, the mathtext works.


This problem does not occur under Windows 7 or XP; only under two different Windows 8 installations.  Any ideas what's going on?

Chad

Re: [Matplotlib-users] Bar chart corruption when plotting multiple subplots in MATPLOTLIB

[p.bingham <p.b...@ke...>]

Hi Ben and thanks for the reply

The matplotlib stuff above is an optional extra to an application that I've
created. I don't want the users to have to install the matplotlib package
unless they have to. The only way I could figure of achieving this is to
import within the class scope. They are assigned as members to allow later
reference. Without the assignment I cannot access the matplot lib module. If
you know a better way of achieving what I'm looking for please let me know.
 
I've deprioritised this bug as it only occurs when the bar data vacillates
between positive and negative values and I'm just looking to plot FFTs. When
I get some time I plan to reduce the problem down and will post here when I
do so. Sorry but I really shouldn't have posted my class directly.
 
Thanks for the feedback. 



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Bar-chart-corruption-when-plotting-multiple-subplots-in-MATPLOTLIB-tp40023p40171.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Re: [Matplotlib-users] Persisting navigation toolbar configuration / settings

[ateamdev <cs...@ho...>]

Hi
Do  you have any documentation about how to implement such functionality? 
Could you perhaps suggest a suitable example?



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Persisting-navigation-toolbar-configuration-settings-tp40124p40153.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

[Matplotlib-users] Fwd:

[David C. <dcd...@gm...>]

http://smallshop.lt/gbwpmas.php

Re: [Matplotlib-users] legend on a plot with broken_barh()

[gsal <sal...@gm...>]

Great, thank you very much...that did the trick.

I don't know why broken_barh does not come in the legend, either; maybe is
because of the ambiguity as to which color to use in the legend...a single
broken barh command can have several bars with different colors. 

gsal



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/legend-on-a-plot-with-broken-barh-tp40145p40151.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Re: [Matplotlib-users] legend on a plot with broken_barh()

[Paul H. <pmh...@gm...>]

[Forgot to reply-all, sorry for the dup, gsal]

On Fri, Jan 4, 2013 at 1:22 PM, gsal <sal...@gm...> wrote:

> can you provide an example? The reference help is only two lines!
>
> Given:
> [code]
> import numpy as np
> import matplotlib.pyplot as plt
>
> fig = plt.figure()
> ax = fig.add_subplot(111)
> ax.broken_barh([ (110, 30), (150, 10) ] , (10, 9), facecolors='b',
> label='barh')
>
> ax.set_xlim((0,200))
> ax.set_ylim((0,50))
>
> ax.legend()
>
> plt.show()
> [/code]
>
> How do I import what in order to, say, create a plot
> "plot([0.0],[0.0],'bs')" so that I can at least plot a marker of the same
> color as my broken_barh so that when the legend is added, the correct icon
> precedes the label?
>
> I tried adding
>
> pp = plt.plot([80],[40],'bs', label='proxy artist')
>
> to the previous program, right before the legend command, but it actually
> plots the marker, too.
>
> Is there a way to import "plot" or "Line2D" or something so that I can
> produce an artist that is NOT related to the plot and, hence, not plotted?
> (is that what "proxy artist" means?).
>

Yes. Proxy artists are created in memory but never added to the axes
object. Here's an expanded version of the example:

import matplotlib.patches as mpatch
import matplotlib.pyplot as plt
fig, ax = plt.subplots()
patch = mpatch.Rectangle((0, 0), 1, 1, fc="r")
ax.legend([patch], ["Proxy artist"])
plt.show()

So for your example, I'd do the following:

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as mpatch
fig = plt.figure()
ax = fig.add_subplot(111)
ax.broken_barh([ (110, 30), (150, 10) ] , (10, 9), facecolors='b',
label='barh')
ax.set_xlim((0,200))
ax.set_ylim((0,50))
fakeredbar = mpatch.Rectangle((0, 0), 1, 1, fc="r")
fakebluebar = mpatch.Rectangle((0, 0), 1, 1, fc="b")
ax.legend([fakeredbar, fakebluebar], ['Red Data', 'Blue Data'])
plt.show()

Now to me, it seems very strange that broken_barh doesn't generate any
items in a legend. Not sure why that is but it seems like a bug.

Hope that helps,
-paul

Re: [Matplotlib-users] legend on a plot with broken_barh()

[gsal <sal...@gm...>]

can you provide an example? The reference help is only two lines!  

Given:
[code]
import numpy as np
import matplotlib.pyplot as plt

fig = plt.figure()
ax = fig.add_subplot(111)
ax.broken_barh([ (110, 30), (150, 10) ] , (10, 9), facecolors='b',
label='barh')

ax.set_xlim((0,200))
ax.set_ylim((0,50))

ax.legend()

plt.show()
[/code]

How do I import what in order to, say, create a plot
"plot([0.0],[0.0],'bs')" so that I can at least plot a marker of the same
color as my broken_barh so that when the legend is added, the correct icon
precedes the label?

I tried adding 

pp = plt.plot([80],[40],'bs', label='proxy artist')

to the previous program, right before the legend command, but it actually
plots the marker, too.

Is there a way to import "plot" or "Line2D" or something so that I can
produce an artist that is NOT related to the plot and, hence, not plotted?
(is that what "proxy artist" means?).





--
View this message in context: http://matplotlib.1069221.n5.nabble.com/legend-on-a-plot-with-broken-barh-tp40145p40149.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Re: [Matplotlib-users] legend on a plot with broken_barh()

[Paul H. <pmh...@gm...>]

On Fri, Jan 4, 2013 at 8:41 AM, gsal <sal...@gm...> wrote:

> So, it looks like broken_barh's do not show up on the legend...is there
> work
> around for this?
>
> Or,
>
> Is there a way to fake a legend? A way to set legend to whatever I want?
>
> Thanks,
>
> gsal
>
>
To fake a legend, try using so-called proxy artists:
http://matplotlib.org/users/legend_guide.html#using-proxy-artist

-paul

Re: [Matplotlib-users] Synching sub-plots

[gsal <sal...@gm...>]

Maybe you can use the  zoom example
<http://matplotlib.org/examples/pylab_examples/axes_zoom_effect.html>   and
the binding method idea so that when you click on some point in the main
graph, the zooming graph updates and plots from -50 to +50 days from the
point that you clicked on the main graph...just an idea.



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Synching-sub-plots-tp40125p40147.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Re: [Matplotlib-users] Persisting navigation toolbar configuration / settings

[gsal <sal...@gm...>]

Hhhmmm...I am thinking to may have to implement some interactivity
(callbacks) to your plot and be able to, say, click or right click and
demand that the current settings be stored or restored from choices
previously stored...but this has to be all your doing.



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Persisting-navigation-toolbar-configuration-settings-tp40124p40146.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

[Matplotlib-users] legend on a plot with broken_barh()

[gsal <sal...@gm...>]

So, it looks like broken_barh's do not show up on the legend...is there work
around for this?

Or,

Is there a way to fake a legend? A way to set legend to whatever I want? 

Thanks,

gsal



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/legend-on-a-plot-with-broken-barh-tp40145.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Re: [Matplotlib-users] subplots [xy]labels

[<gia...@po...>]

Alex Goodman <goo...@il...>
writes:

> Try using the set_ylabel() and set_xlabel() methods for each Axes instance
> instead, eg:
>
> a[0].set_ylabel('f1')
> ...

works as intended, tx Alex

Re: [Matplotlib-users] subplots [xy]labels

[Alex G. <goo...@il...>]

Hi Giacomo,

Try using the set_ylabel() and set_xlabel() methods for each Axes instance
instead, eg:

a[0].set_ylabel('f1')
a[0].set_xlabel('t')

a[1].set_ylabel('f2')
a[1].set_xlabel('t')


On Fri, Jan 4, 2013 at 5:44 AM, <gia...@po...> wrote:

> two plots in a figure:
> ########################################################################
> from pylab import *
> ...
> f,a = subplots(nrows=2, sharex=False, sharey=False)
> a[0].plot(x,f0(x))
> ylabel('f1')
> xlabel('t')
> ...
> a[1].plot(x,f1(x))
> ylabel('f2')
> xlabel('t')
> ...
> show()
> ########################################################################
> but all i can get are labels for ONLY the lower subplot, what shoud I do?
>
> any help will be appreciated
>
> tia
>                                                         gb
>
>
>
> ------------------------------------------------------------------------------
> Master HTML5, CSS3, ASP.NET, MVC, AJAX, Knockout.js, Web API and
> much more. Get web development skills now with LearnDevNow -
> 350+ hours of step-by-step video tutorials by Microsoft MVPs and experts.
> SALE $99.99 this month only -- learn more at:
> http://p.sf.net/sfu/learnmore_122812
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>



-- 
Alex Goodman
University of Illinois at Urbana-Champaign
Email: goo...@il...

[Matplotlib-users] subplots [xy]labels

[<gia...@po...>]

two plots in a figure:
########################################################################
from pylab import *
...
f,a = subplots(nrows=2, sharex=False, sharey=False)
a[0].plot(x,f0(x))
ylabel('f1')
xlabel('t')
...
a[1].plot(x,f1(x))
ylabel('f2')
xlabel('t')
...
show()
########################################################################
but all i can get are labels for ONLY the lower subplot, what shoud I do?

any help will be appreciated

tia
                                                        gb

Re: [Matplotlib-users] open circles aren't showing up in the legend box

[Joe L. <jgl...@gm...>]

Thank you everyone for your answers.  My legend is no longer defaced since
I de-faced the call to scatter() :-)

Re: [Matplotlib-users] open circles aren't showing up in the legend box

[Jonathan S. <js...@cf...>]

Joe,

I think the problem is the edgecolor='face' in the scatter call for the
open circles.  For me when I omit that, it all works.  I'd also note
that calling legend after draw results in the legend showing the circles
correctly colored.

Regards,
Jon

On Wed, 2013-01-02 at 16:46 -0800, Joe Louderback wrote:
> Using matplotlib.pyplot I create a scatter plot using two symbols, an
> open circle, and a filled circle.  The legend for this plot shows both
> labels, but there is no symbol for the open circles (see
> missing_legend_circles.png).  Switching to matplotlib.pylab yields the
> same result.  However, from an "ipython --pylab" session the open
> circles are displayed in the legend.  This has me more baffled than
> usual.  Here's the code to reproduce the enclosed figure:
> 
> import matplotlib.pyplot as plt
> 
> 
> fig = plt.figure()
> plot = fig.add_subplot(111)
> 
> 
> plot.scatter([1, 2, 3], [4, 5, 6], c = [0.2, 0.4, 0.6], label =
> 'one', 
>     cmap = 'jet', marker = 'o', edgecolor = 'face') 
> 
> 
> plot.scatter([1, 2, 3], [7, 6, 5], c = [0.2, 0.4, 0.6], label = 'two',
>               cmap = 'jet', marker = 'o', edgecolor = 'face', 
>      facecolor = 'none')
> plot.legend()
> plt.show()
> 
> 
> 
> 
> This is with matplotlib 1.1 and 1.2 on Windows 7 with the TkAgg, WxAgg
> and WX backends.
> 
> 
> Thank you for any help.
> 
> 
> -- Joe Louderback

-- 
______________________________________________________________
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
______________________________________________________________

Re: [Matplotlib-users] open circles aren't showing up in the legend box

[ChaoYue <cha...@gm...>]

I doubt there is a color called 'face',  according to documentation:

In [2]: setp(mat.collections.PathCollection,'edgecolor')
  edgecolor: matplotlib color arg or sequence of rgba tuples     

probably because matplotlib doesn't know how rend the color 'face'?

Chao



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/open-circles-aren-t-showing-up-in-the-legend-box-tp40136p40138.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Re: [Matplotlib-users] open circles aren't showing up in the legend box

[Paul H. <pmh...@gm...>]

On Wed, Jan 2, 2013 at 4:46 PM, Joe Louderback <jgl...@gm...>wrote:

> import matplotlib.pyplot as plt
>
> fig = plt.figure()
> plot = fig.add_subplot(111)
>
> plot.scatter([1, 2, 3], [4, 5, 6], c = [0.2, 0.4, 0.6], label = 'one',
>      cmap = 'jet', marker = 'o', edgecolor = 'face')
>
> plot.scatter([1, 2, 3], [7, 6, 5], c = [0.2, 0.4, 0.6], label = 'two',
>               cmap = 'jet', marker = 'o', edgecolor = 'face',
>       facecolor = 'none')
> plot.legend()
> plt.show()
>

I'm not quite sure what /should/ happen when you set the `facecolor` to
'none' and the `edgecolor` to 'face'. I'm surprised anything shows up at
the axes at all. But if you remove the `edgecolor` kwarg things look better:

import matplotlib.pyplot as plt
fig = plt.figure()
plot = fig.add_subplot(111)
plot.scatter([1, 2, 3], [4, 5, 6], c = [0.2, 0.4, 0.6], label = 'one',
     cmap = 'jet', marker = 'o', edgecolor = 'face')

plot.scatter([1, 2, 3], [7, 6, 5], c = [0.2, 0.4, 0.6], label = 'two',
              cmap = 'jet', marker = 'o', facecolor='none')
plot.legend()
plt.show()