matplotlib-users Mailing List for matplotlib (Page 8)

Hello,
I need to prepare a black&white bar chart with several sets of bars.
Now, I have white bars and black bars but I need more colorings that are
distinguishable in black&white.
Is it possible to make bars filled with some texture or tiling or a grid?
Excel has this function but I could not find a way to do it in
python/matplotlib.

Thanks,
Eli

Hello,
I need to prepare two versions of figures: color and BW(Black&White).
Is there an easy way to produce just the colored version and than use some
command or script to turn it to BW or grayscale?
I thought that converting from color to BW really means: "in all object in
the figure, turn any color that is not white to black".
Is there an easy way to implement this?

Thanks
Eli

  Is there a way to put the legend for a graph inside the margin instead 
of on the graph, in other words to put the legend where excel would?  I 
have a stacked bar graph with many categories and so the legend is vary 
large and there is little dead space on the graph.  As a result if the 
legend is on the graph it covers up one of the bars.

Hello,

I have just chaged over to the Qt4Agg backend so I can use Pierre Raybaut's
addition of editing cuves and parameters (thanks for this functionality).  I
have found a small issue.  Whenever I apply changes with more than one
trace/curve, the colour of the second curve changes.  The example below is
enough to show the issue:

from matplotlib import pyplot
import numpy as np

a = np.random.rand(100)
b =a*1.5
fig = pyplot.figure()
ax = fig.add_subplot(111)
ax.plot(a)
ax.plot(b)
pyplot.show()

The  example above changes second trace from a default of green to a light
green when I use the edit button.  Just pressing 'apply' without any changes
should be enough to trigger the change.  It appears that the addin is not
getting the correct color from the plot and so apply is appying the
incorrect color back into the plot.

Also if I try to set color for a trace like this:
fig = pyplot.figure()
ax = fig.add_subplot(111)
ax.plot(a,'r')
pyplot.show()
It works but then when I press the edit curves button, I get:

Traceback (most recent call last):
  File "C:\Python26\lib\site-packages\matplotlib\backends\backend_qt4.py",
line 469, in edit_parameters
    figureoptions.figure_edit(axes, self)
  File
"C:\Python26\lib\site-packages\matplotlib\backends\qt4_editor\figureoptions.py",
line 154, in figure_edit
    icon=get_icon('qt4_editor_options.svg'), apply=apply_callback)
  File
"C:\Python26\lib\site-packages\matplotlib\backends\qt4_editor\formlayout.py",
line 460, in fedit
    dialog = FormDialog(data, title, comment, icon, parent, apply)
  File
"C:\Python26\lib\site-packages\matplotlib\backends\qt4_editor\formlayout.py",
line 384, in __init__
    parent=self)
  File
"C:\Python26\lib\site-packages\matplotlib\backends\qt4_editor\formlayout.py",
line 362, in __init__
    widget = FormComboWidget(data, comment=comment, parent=self)
  File
"C:\Python26\lib\site-packages\matplotlib\backends\qt4_editor\formlayout.py",
line 344, in __init__
    widget = FormWidget(data, comment=comment, parent=self)
  File
"C:\Python26\lib\site-packages\matplotlib\backends\qt4_editor\formlayout.py",
line 233, in __init__
    self.setup()
  File
"C:\Python26\lib\site-packages\matplotlib\backends\qt4_editor\formlayout.py",
line 256, in setup
    selindex = value.pop(0)
AttributeError: 'tuple' object has no attribute 'pop'

I am on version 1.0.0 and XP


Thanks,
Bevan




-- 
View this message in context: http://old.nabble.com/Qt4Agg-backend---edit-curves-and-axis-parameters-tp29754925p29754925.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

On Sat, Sep 18, 2010 at 1:03 PM, Benjamin Root <ben...@ou...> wrote:
>
> I'll admit that I am not very familiar with how these step plots are done.
> Maybe it should be the 'drawstyle' kwarg that should be over-riden, not
> 'linestyle'?
>
> Ben Root
>

This seems like more reasonable to me. When I update
/matplotlib/lib/axes.py:

kwargs['linestyle'] = 'steps-' + where -> kwargs['drawstyle'] = 'steps-' +
where

then I get the expected behavior.

-- 
Gökhan

Thanks, that works fantastically !

-- Forest.

On Fri, Sep 17, 2010 at 9:23 PM, Benjamin Root <ben...@ou...> wrote:
> On Fri, Sep 17, 2010 at 8:05 PM, Forest Yang <yzi...@gm...> wrote:
>>
>> Hi
>>
>>   I have a function z(x, y) on a regular grid. But some of the value
>> z are not defined on (x,y). I want to plot the contour or contourf of
>> z on (x,y) but exclude specific (x,y) points.
>> How can I do it ? Right now I just draw small colored square
>> (rectangular) around defined (x,y) the color is not smooth since no
>> interpolation like contour or contourf.
>>
>> Thanks.
>>
>> Forest.
>>
>
> Forest,
>
> There are a few ways to do this.  If you have a recent enough version of
> matplotlib, you can use masked arrays, and the contourf will just ignore
> those data points.  One could also use NaNs and make sure that the clim (the
> limits on z that you wish to display a color for) is defined.
>
> To make a masked array is easy.  Imagine you wish to exclude any value less
> than zero (assume z is defined):
>
> import numpy.ma as ma
> z_masked = ma.masked_array(z, mask=(z < 0.))
>
> And then just use the  masked array in your contourf as you would the
> regular numpy array.
>
> I hope that helps!
> Ben Root
>
>

Many thanks Ben!

I went and removed all fonts but regular helvetica and vera, also I
removed all but the .png files under images, and now my mpl-data is
about 250 kb.

cheers

Carlos



On Thu, Sep 16, 2010 at 21:39, Benjamin Root <ben...@ou...> wrote:
> On Thu, Sep 16, 2010 at 5:33 PM, Carlos Grohmann <car...@gm...>
> wrote:
>>
>> Hello all,
>>
>> Is it OK to remove the fonts I don't use? (I use only sans-serif) By
>> Ok I mean not only from the practical poin tof view (that is, will the
>> app run?) but also from the _legal_ point of view (am I obliged to
>> distribute all those fonts?)
>>
>
> Carlos,
>
> Just to make it very clear, matplotlib is open sourced.  You are free to
> modify the package to your heart's content, however you see fit.  Matplotlib
> only asks that you keep the copyright notice with the distributed software
> (in particular, the matplotlib/license/LICENSE file applies here, as well as
> others.)
>
> With regards to fonts, refer to the matplotlib/license/LICENSE_STIX file.
> My understanding of that license (though, IANAL), is that you don't have to
> worry about anything above and beyond just simply including the license file
> unless you are eliminating individual glyphs from a font (or adding
> glyphs).  However, I don't see any reason why you can't constrain yourself
> to a particular font.  Note that you are not allowed to sell any particular
> font in the package, though you are allowed to charge a distribution fee for
> the "font software".
>
> Generally speaking, my rule of thumb is that if you are distributing
> open-source software in the same spirit you have received it, you are
> satisfying the spirit of the licenses.  The only thing remaining is whether
> the source code has to accompany the software or not.  The core part of
> matplotlib is BSD licensed (or similar) and does not require that (although
> it is encouraged!).
>
> Important! Note that the basemap package is GPL-licensed, and is required to
> have its source code accompany its software.  However, unless your program
> *depends* on basemap for it to function, the source code to your program is
> not required to be GPL-ed.
>
> And, as always, I am not a lawyer.  I am merely conveying my understanding
> and experience with software licensing.  Anyone else is free to add to
> and/or correct what I have said here.
>
> I hope this helps!
> Ben Root
>



-- 
Prof. Carlos Henrique Grohmann - Geologist D.Sc.
Institute of Geosciences - Univ. of São Paulo, Brazil
http://www.igc.usp.br/pessoais/guano
http://lattes.cnpq.br/5846052449613692
Linux User #89721
________________
Can’t stop the signal.

That works perfectly. Thank you all so much! 


Gökhan SEVER-2 wrote:
> 
> On Fri, Sep 17, 2010 at 1:13 PM, musik <xi....@gm...> wrote:
> 
>>
>> Is there a way to set the legend font color? I am plotting multiple sets
>> of
>> data using different colors. I basically want to set each legend font
>> color
>> the same as the corresponding data line color. Here is an example.
>>
>> x = arange(0,10,0.1)
>> y1 = sin(x)
>> y2 = cos(x)
>>
>> plot(x,y1,'r-',x,y2,'b--')
>>
>> I want the legend font for y1 to be in red and the legend font for y2 to
>> be
>> in blue.
>>
>> Can anybody help? Thanks.
>>
> 
> I initially read like Ted did :) Here is a simple demonstration of Ben's
> explanation:
> 
> x = np.arange(0,10,0.1)
> y1 = np.sin(x)
> y2 = np.cos(x)
> fig = plt.figure()
> ax = fig.add_subplot(111)
> p1, = ax.plot(x, y1, label="y1")
> p2, = ax.plot(x, y2, label="y2")
> leg = ax.legend()
> text1, text2 = leg.get_texts()
> # this part can be turned into a loop depends on the number of text
> objects
> text1.set_color(p1.get_color())
> text2.set_color(p2.get_color())
> plt.show()
> 
> 
> -- 
> Gökhan
> 
> ------------------------------------------------------------------------------
> Start uncovering the many advantages of virtual appliances
> and start using them to simplify application deployment and
> accelerate your shift to cloud computing.
> http://p.sf.net/sfu/novell-sfdev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
> 

-- 
View this message in context: http://old.nabble.com/legend-font-color-tp29741260p29747823.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

On Fri, Sep 17, 2010 at 11:33 PM, Jeremy Conlin <jlc...@gm...> wrote:

> On Fri, Sep 17, 2010 at 12:44 AM, Scott Sinclair
> <sco...@gm...> wrote:
> > On 16 September 2010 22:52, Jeremy Conlin <jlc...@gm...> wrote:
> >> I have a colorbar which has some ticks, but I would like to add my own
> >> ticks without replacing any of the existing ones.  In addition, I
> >> would like to give the ticks a different labels like "min" and "max".
> >> Can someone show how this might be done?
> >
> >
> http://matplotlib.sourceforge.net/examples/pylab_examples/colorbar_tick_labelling_demo.html
> > should give you some ideas to get started.
>
> Thanks for pointing me to that demo, it's a good one.  However, I want
> the ticks that are generated automatically, in addition to the extra
> ones I add.  The demo seems to show how I can define my own only.
>
> Jeremy
>
>
While I haven't tried this, so there might be a few extra things to do to
make this work... You could first call

ticks = cbar.ax.get_yticks()
tickStrs = [label.get_text() for label in cbar.ax.get_yticklabels()]

to get the list of tick locations and the strings.
Then, insert any additional ticks at whatever locations you want in 'ticks'.
Next, insert the desired strings at the same locations in 'tickStrs'.
Then, update the tick locations and strings in cbar:

cbar.set_ticks(ticks, update_ticks=False)
cbar.set_ticklabels(tickStrs, update_ticks=True)

The use of update_ticks=False is to defer the actual processing of the tick
data until later.  This is to avoid conflicts with the differing lengths of
tick locations and tick labels.  The update would then occur at the end of
set_ticklabels().

Note, I have not tested this, so there is no guarantee that this would
work.  Good luck!

Ben Root

On Sat, Sep 18, 2010 at 10:39 AM, Gökhan Sever <gok...@gm...>wrote:

> On Fri, Sep 17, 2010 at 8:59 PM, Benjamin Root <ben...@ou...> wrote:
>>
>> Not a bug, but a "feature".  In step(), the 'linestyle' kwarg gets
>> over-ridden with a value of 'steps-' + kwargs['where'].
>>
>> Ben Root
>>
>
> Hard to interpret that way at the first look. This does what I want:
>
> plt.plot(range(10), 'g--', drawstyle='steps-mid')
>
>
I'll admit that I am not very familiar with how these step plots are done.
Maybe it should be the 'drawstyle' kwarg that should be over-riden, not
'linestyle'?

Ben Root

On Fri, Sep 17, 2010 at 8:59 PM, Benjamin Root <ben...@ou...> wrote:
>
> Not a bug, but a "feature".  In step(), the 'linestyle' kwarg gets
> over-ridden with a value of 'steps-' + kwargs['where'].
>
> Ben Root
>

Hard to interpret that way at the first look. This does what I want:

plt.plot(range(10), 'g--', drawstyle='steps-mid')


-- 
Gökhan

Hi Mario,

(Sorry for the reply to a reply, but I was not on the list when the original message was posted)

This may not be useful for Basemap, but I did this with the matplotlib.projections.geo HammerAxes projection.  I got it to work by subclassing that and modifying the _get_affine_transform() to negate the x-axis scale.  I also seem to have had to fiddle with _set_lim_and_transforms() to account for this, though I don't remember the details.

In any case, I don't know how Basemap works and whether this is helpful.  Maybe as inspiration...  The code is below if it helps.

joey


On Sep 18, 2010, at 6:02 AM, Jeff Whitaker wrote:

>  On 9/16/10 12:38 PM, Mario Juric wrote:
>> Hi,
>> 	I'm looking at Basemap as a backend for plotting maps of the sky in
>> different projections, and so far it seems like a really good match!
>> Excellent work!
>> 
>> 	The only problem that I don't know how to solve is that in astronomy
>> the longitude on maps typically increases from right to left (we're
>> looking at the celestial sphere from the "inside"). Is there any way (or
>> a trick) to make Basemap do this?
>> 
>> Regards,

class AstroHammerAxes(HammerAxes):
    """Astronomical hammer projection.

    This is just a Hammer projection with the longitude axis
    reversed left-right.

    """
    name='astrohammer'

    def _get_affine_transform(self):
        """get affine transform

        This is cribbed from GeoAxes, but negates the xscale
        to put positive longitude to the left.  Not sure this
        is actually robust.

        """
        transform=self._get_core_transform(1)
        xscale,_=transform.transform_point((pi,0))
        _,yscale=transform.transform_point((0,pi/2.0))
        return Affine2D()\
               .scale(-0.5/xscale, 0.5/yscale)\
               .translate(0.5,0.5)

    def _set_lim_and_transforms(self):
        HammerAxes._set_lim_and_transforms(self)
        yaxis_stretch = Affine2D().scale(pi * 2.0, 1.0).translate(pi, 0.0)
        yaxis_space = Affine2D().scale(1.0, 1.1)
        yaxis_text_base = \
            yaxis_stretch + \
            self.transProjection + \
            (yaxis_space + \
             self.transAffine + \
             self.transAxes)
        self._yaxis_text1_transform = \
            yaxis_text_base + \
            Affine2D().translate(-8.0, 0.0)
        self._yaxis_text2_transform = \
            yaxis_text_base + \
            Affine2D().translate(8.0, 0.0)

# register the projection on import
register_projection(AstroHammerAxes)

  On 9/16/10 12:38 PM, Mario Juric wrote:
> Hi,
> 	I'm looking at Basemap as a backend for plotting maps of the sky in
> different projections, and so far it seems like a really good match!
> Excellent work!
>
> 	The only problem that I don't know how to solve is that in astronomy
> the longitude on maps typically increases from right to left (we're
> looking at the celestial sphere from the "inside"). Is there any way (or
> a trick) to make Basemap do this?
>
> Regards,
Mario:  Are you asking whether the basemap coordinate system can be 
reversed, or just the longitude labelling?  If it's the former, the 
answer is no.  If you just want the labels reversed,  you can pass the 
drawmeridians method a custom formatting function with the 'fmt' keyword.

-Jeff

On Fri, Sep 17, 2010 at 12:44 AM, Scott Sinclair
<sco...@gm...> wrote:
> On 16 September 2010 22:52, Jeremy Conlin <jlc...@gm...> wrote:
>> I have a colorbar which has some ticks, but I would like to add my own
>> ticks without replacing any of the existing ones.  In addition, I
>> would like to give the ticks a different labels like "min" and "max".
>> Can someone show how this might be done?
>
> http://matplotlib.sourceforge.net/examples/pylab_examples/colorbar_tick_labelling_demo.html
> should give you some ideas to get started.

Thanks for pointing me to that demo, it's a good one.  However, I want
the ticks that are generated automatically, in addition to the extra
ones I add.  The demo seems to show how I can define my own only.

Jeremy

On 9/17/10 9:08 PM, Fernando Perez wrote:
> Hi Luke,
>
> On Fri, Sep 17, 2010 at 5:49 PM, Dale Lukas Peterson
> <haz...@gm...>  wrote:
>>
>>   I'm not sure I understand how I would make use of my function then.
>>   My function needs to be evaluated over a 3-d mesh (x, y, and z) , and then the
>>   level surfaces (not contour lines) calculated.  I guess I could treat
>>   z as a parameter, then plot the zero level contour lines of my function for
>>   a discrete number of z values, but then I would need to adjust the
>>   height that each countour line is plotted at when I do the 3-d plot.
>>   This still would only give bunch of vertically stacked contour
>>   lines, rather than a nice smooth 3-d surface.
>>
>>   If I'm misunderstanding what you meant, perhaps you could point me
>>   to an example of something that makes a level surface of a function
>>   of 3 (not 2) variables?
>
> You're looking for an isosurface; as far as I know matplotlib does not
> have isosurface modules, only 2-d contours embedded in 3d (such as
> those illustrated in
> http://matplotlib.sourceforge.net/examples/mplot3d/contourf3d_demo.html).
>
> VTK does have powerful isosurface capabilities, nicely exposed by mayavi:
>
> http://code.enthought.com/projects/mayavi/docs/development/html/mayavi/auto/mlab_helper_functions.html#contour3d


Sage will also do this sort of thing, though it's not as powerful as 
VTK/Mayavi in this functionality:

http://www.sagemath.org/doc/reference/sage/plot/plot3d/implicit_plot3d.html

Here are lots of sheets on sagenb.org that use implicit_plot3d somewhere:

http://sagenb.org/pub/?typ=pub&search=implicit_plot3d


Thanks,

Jason

Hi Luke,

On Fri, Sep 17, 2010 at 5:49 PM, Dale Lukas Peterson
<haz...@gm...> wrote:
>
>  I'm not sure I understand how I would make use of my function then.
>  My function needs to be evaluated over a 3-d mesh (x, y, and z) , and then the
>  level surfaces (not contour lines) calculated.  I guess I could treat
>  z as a parameter, then plot the zero level contour lines of my function for
>  a discrete number of z values, but then I would need to adjust the
>  height that each countour line is plotted at when I do the 3-d plot.
>  This still would only give bunch of vertically stacked contour
>  lines, rather than a nice smooth 3-d surface.
>
>  If I'm misunderstanding what you meant, perhaps you could point me
>  to an example of something that makes a level surface of a function
>  of 3 (not 2) variables?

You're looking for an isosurface; as far as I know matplotlib does not
have isosurface modules, only 2-d contours embedded in 3d (such as
those illustrated in
http://matplotlib.sourceforge.net/examples/mplot3d/contourf3d_demo.html).

VTK does have powerful isosurface capabilities, nicely exposed by mayavi:

http://code.enthought.com/projects/mayavi/docs/development/html/mayavi/auto/mlab_helper_functions.html#contour3d

If the mlab helper isn't sufficient for you, you can create directly
VTK isosurfaces, the heart example is a good point to start learning:

http://code.enthought.com/projects/mayavi/docs/development/html/mayavi/example_heart.html

Regards,

f

On Fri, Sep 17, 2010 at 8:43 PM, Gökhan Sever <gok...@gm...> wrote:

> On Fri, Sep 17, 2010 at 3:34 PM, Gökhan Sever <gok...@gm...>wrote:
>
>> Hello,
>>
>> Can someone confirm me if this creates a dashed line for a simple step
>> plot?
>>
>> # this is fine
>> plt.plot(range(10), "g--")
>>
>> # plots solid line!
>> plt.step(range(10), "g--")
>>
>> Thanks,
>>
>> --
>> Gökhan
>>
>
> My version is v8624. Looks like a bug to me.
>


Not a bug, but a "feature".  In step(), the 'linestyle' kwarg gets
over-ridden with a value of 'steps-' + kwargs['where'].

Ben Root

On Fri, Sep 17, 2010 at 4:37 PM, Jeremy Lewi <jl...@in...> wrote:

>  HI,
>
>
>
> Is there a way to get the size of the bounding box for the axes which
> includes the axes labels and tick marks? It looks like
> Axes.get_position/set_position refers to the inner position (i.e the actual
> plot area).
>
>
>
>
>
> In matlab (http://www.mathworks.com/help/techdoc/ref/axes_props.html) this
> was the outer and inner position of an axes. I would like to get the outer
> position (i.e the yellow box in
> http://www.mathworks.com/help/techdoc/ref/axes_props.html)
>
>
>
> Thanks
>
> Jeremy
>
>
>
> Jeremy Lewi
>
> Engineering Scientist
>
> The Intellisis Corporation
>
> jl...@in...
>
>
>
>
Jeremy,

Yes, get/set_position refers to the position of the plot area within a
particular axes area in a figure.  I am not exactly aware of how one
determines its outer boundary box, because there really isn't an outer box.
Everything within an axes area is plotted within coordinates of 0 to 1.  The
plot box is set typically at 0.125 to 0.9.  Axes labels are then draw as an
offset relative to the plot area.  This can cause the axes labels and figure
titles to  overlap each other in neighboring subplots if they get too big.

Tony Yu did make some  code to help deal with display issues surrounding the
lack of a constrained outer box.  See this link for some discussion of it
and the code.  Maybe it can help you with whatever you need.

http://permalink.gmane.org/gmane.comp.python.matplotlib.general/24652

I hope that helps!
Ben Root

On Fri, Sep 17, 2010 at 3:34 PM, Gökhan Sever <gok...@gm...> wrote:

> Hello,
>
> Can someone confirm me if this creates a dashed line for a simple step
> plot?
>
> # this is fine
> plt.plot(range(10), "g--")
>
> # plots solid line!
> plt.step(range(10), "g--")
>
> Thanks,
>
> --
> Gökhan
>

My version is v8624. Looks like a bug to me.

On Fri, Sep 17, 2010 at 1:13 PM, musik <xi....@gm...> wrote:

>
> Is there a way to set the legend font color? I am plotting multiple sets of
> data using different colors. I basically want to set each legend font color
> the same as the corresponding data line color. Here is an example.
>
> x = arange(0,10,0.1)
> y1 = sin(x)
> y2 = cos(x)
>
> plot(x,y1,'r-',x,y2,'b--')
>
> I want the legend font for y1 to be in red and the legend font for y2 to be
> in blue.
>
> Can anybody help? Thanks.
>

I initially read like Ted did :) Here is a simple demonstration of Ben's
explanation:

x = np.arange(0,10,0.1)
y1 = np.sin(x)
y2 = np.cos(x)
fig = plt.figure()
ax = fig.add_subplot(111)
p1, = ax.plot(x, y1, label="y1")
p2, = ax.plot(x, y2, label="y2")
leg = ax.legend()
text1, text2 = leg.get_texts()
# this part can be turned into a loop depends on the number of text objects
text1.set_color(p1.get_color())
text2.set_color(p2.get_color())
plt.show()


-- 
Gökhan

On Fri, Sep 17, 2010 at 4:57 PM, John Hutchinson <jmh...@gm...> wrote:

> Hi:
>
> I am trying to make a 1row by 3 column plot with subplot, and I want
> the first plot (subplot(131)) to have equal aspect ratio, but the rest
> can auto scale.
>
> My code results in an empty plot for the 1st column subplot whenever I
> try to use the set_aspect('equal')
>
> Any ideas?
>
> Thanks
> John
>
>
> figure(1)
> hold(True)
> subplot(131)
> plot(RX,RY)
> title('Geometry')
> xlabel('um')
> ylabel('um')
> scatter(AX,AY,c='green',marker='s')
> plot([AX,0],[AY,0],linewidth=3,c='green')
> scatter(BX,BY,c='purple',marker='s')
> plot([0,BX],[0,BY],linewidth=3,c='purple')
> scatter(DX,DY,c='red',marker='o')
> axes().set_aspect('equal')
> hold(False)
> subplot(132)
> scatter(e1.real,gi,marker='s')
> title('E1 by tooth')
> subplot(133)
> scatter(e2.real,gi,marker='+')
> title('E2 by tooth')
> show()
>
>
This is most likely because of the call to axes().  I can't remember for
sure, but it might be automatically clearing out the axes.  Replace axes()
with subplot(131):

subplot(131).set_aspect('equal')

The subplot(131) call should recognize that that particular axes has already
been made and just simply retrieves it without modifying anything.

I hope this helps!
Ben Root

On Fri, Sep 17, 2010 at 8:05 PM, Forest Yang <yzi...@gm...> wrote:

> Hi
>
>   I have a function z(x, y) on a regular grid. But some of the value
> z are not defined on (x,y). I want to plot the contour or contourf of
> z on (x,y) but exclude specific (x,y) points.
> How can I do it ? Right now I just draw small colored square
> (rectangular) around defined (x,y) the color is not smooth since no
> interpolation like contour or contourf.
>
> Thanks.
>
> Forest.
>
>
Forest,

There are a few ways to do this.  If you have a recent enough version of
matplotlib, you can use masked arrays, and the contourf will just ignore
those data points.  One could also use NaNs and make sure that the clim (the
limits on z that you wish to display a color for) is defined.

To make a masked array is easy.  Imagine you wish to exclude any value less
than zero (assume z is defined):

import numpy.ma as ma
z_masked = ma.masked_array(z, mask=(z < 0.))

And then just use the  masked array in your contourf as you would the
regular numpy array.

I hope that helps!
Ben Root

I misunderstood your question, musik. My apologies.

Ted


On 18 September 2010 02:15, Benjamin Root <ben...@ou...> wrote:

> On Fri, Sep 17, 2010 at 1:13 PM, musik <xi....@gm...> wrote:
>
>>
>> Is there a way to set the legend font color? I am plotting multiple sets
>> of
>> data using different colors. I basically want to set each legend font
>> color
>> the same as the corresponding data line color. Here is an example.
>>
>> x = arange(0,10,0.1)
>> y1 = sin(x)
>> y2 = cos(x)
>>
>> plot(x,y1,'r-',x,y2,'b--')
>>
>> I want the legend font for y1 to be in red and the legend font for y2 to
>> be
>> in blue.
>>
>> Can anybody help? Thanks.
>>
>>
>>
> musik,
>
> I am not aware of any easy way to control the font properties for each
> individual item in a legend box.  I would suppose that it would be possible
> to obtain the list of Text objects from the legend object (*after* creating
> the legend object), and then modify their color properties based on the
> string that it contains.  Note that I have never personally tried this.
>
> As far as I can tell, the text for the legend elements are black by
> default.  Note that you can set the font properties for all Text objects in
> a legend when creating the legend with the 'prop' kwarg.
>
> I hope this helps,
> Ben Root
>
>
> ------------------------------------------------------------------------------
> Start uncovering the many advantages of virtual appliances
> and start using them to simplify application deployment and
> accelerate your shift to cloud computing.
> http://p.sf.net/sfu/novell-sfdev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

On Fri, Sep 17, 2010 at 1:13 PM, musik <xi....@gm...> wrote:

>
> Is there a way to set the legend font color? I am plotting multiple sets of
> data using different colors. I basically want to set each legend font color
> the same as the corresponding data line color. Here is an example.
>
> x = arange(0,10,0.1)
> y1 = sin(x)
> y2 = cos(x)
>
> plot(x,y1,'r-',x,y2,'b--')
>
> I want the legend font for y1 to be in red and the legend font for y2 to be
> in blue.
>
> Can anybody help? Thanks.
>
>
>
musik,

I am not aware of any easy way to control the font properties for each
individual item in a legend box.  I would suppose that it would be possible
to obtain the list of Text objects from the legend object (*after* creating
the legend object), and then modify their color properties based on the
string that it contains.  Note that I have never personally tried this.

As far as I can tell, the text for the legend elements are black by
default.  Note that you can set the font properties for all Text objects in
a legend when creating the legend with the 'prop' kwarg.

I hope this helps,
Ben Root

Hi

   I have a function z(x, y) on a regular grid. But some of the value
z are not defined on (x,y). I want to plot the contour or contourf of
z on (x,y) but exclude specific (x,y) points.
How can I do it ? Right now I just draw small colored square
(rectangular) around defined (x,y) the color is not smooth since no
interpolation like contour or contourf.

Thanks.

Forest.