matplotlib-users Mailing List for matplotlib

On Aug 30, 2012, at 12:35PM, Benjamin Root wrote:

> 
> 
> On Thu, Aug 30, 2012 at 3:26 PM, Goyo <goy...@gm...> wrote:
> 2012/8/28 Sterling Smith <sm...@fu...>:
> > List,
> >
> > Consider the following script:
> >
> > import pylab
> > pylab.plot(pylab.linspace(0,1,100),label='Test',marker='o',ls='')
> > pylab.plot(pylab.linspace(0,1,100),label='Test2',marker='o',ls='-')
> > leg=pylab.legend(loc='best')
> > line=leg.get_lines()
> > line[0].set_color('black')
> > line[1].set_color('black')
> > pylab.draw()
> >
> > I am attaching the results I see with the TkAgg baackend.  I am using version 1.1.0.
> >
> > I would expect the markers to change colors also.  However, I can't seem to find the markers recorded in the legend object to be able to change them.  Is this a bug or a feature (resetting the color could lead to inconsistencies between legend and plot)?
> 
> 
> I think it's a design issue, there's no connection kept between plot
> lines and the legend. You can change colors in the plot an then call
> legend again instead.
> 
> Cheers
> 
> Goyo
> 
> 
> That is correct.  The lines that one obtains from the get_lines() method of legend are completely new line objects.  It is only upon the creation of those lines (at legend creation) that the lines inherit the plot's line properties.
> 
> Now, in the future, it would be nice for artist objects to have shareable "styles", in which case an edit to a style in one place effects all artists with the same style object.  But that might be something reserved for version 2+.
> 
> Cheers!
> Ben Root
> 

Ben, Goyo,

Thank you for taking the time to consider my question.  I'm sorry that I didn't pose my question correctly.  I should have said: 'Consider the _results_ of the following script:'  I originally tried to attach the results I obtained, which showed no change in color for the markers in the legend, while the line connecting the markers in the legend did change color.

I understand that the lines in the legend are new objects; this is desirable in my case because I have the same symbol with different colors, where the color means one thing (plasma condition), but the symbol means something else (simulated vs measured).  So it makes sense for the markers in the legend describing the meaning of the symbols to be black, as opposed to any of the colors of the markers.   However, setting the legend's new line2d object color properties does not change the appearance of the marker in the legend, only the appearance of the line.  I suspect that this is because the legend marker is drawn separately from the legend line to accommodate the numpoints argument of the legend functions.  Then the question is how to access these markers if they are separate from the line2d objects in the legend.  I didn't even see them in the children of the legend [legend.get_children()].

Thank you for your attention,
Sterling

On Thu, Aug 30, 2012 at 3:26 PM, Goyo <goy...@gm...> wrote:

> 2012/8/28 Sterling Smith <sm...@fu...>:
> > List,
> >
> > Consider the following script:
> >
> > import pylab
> > pylab.plot(pylab.linspace(0,1,100),label='Test',marker='o',ls='')
> > pylab.plot(pylab.linspace(0,1,100),label='Test2',marker='o',ls='-')
> > leg=pylab.legend(loc='best')
> > line=leg.get_lines()
> > line[0].set_color('black')
> > line[1].set_color('black')
> > pylab.draw()
> >
> > I am attaching the results I see with the TkAgg baackend.  I am using
> version 1.1.0.
> >
> > I would expect the markers to change colors also.  However, I can't seem
> to find the markers recorded in the legend object to be able to change
> them.  Is this a bug or a feature (resetting the color could lead to
> inconsistencies between legend and plot)?
>
>
> I think it's a design issue, there's no connection kept between plot
> lines and the legend. You can change colors in the plot an then call
> legend again instead.
>
> Cheers
>
> Goyo
>
>
That is correct.  The lines that one obtains from the get_lines() method of
legend are completely new line objects.  It is only upon the creation of
those lines (at legend creation) that the lines inherit the plot's line
properties.

Now, in the future, it would be nice for artist objects to have shareable
"styles", in which case an edit to a style in one place effects all artists
with the same style object.  But that might be something reserved for
version 2+.

Cheers!
Ben Root

2012/8/28 Sterling Smith <sm...@fu...>:
> List,
>
> Consider the following script:
>
> import pylab
> pylab.plot(pylab.linspace(0,1,100),label='Test',marker='o',ls='')
> pylab.plot(pylab.linspace(0,1,100),label='Test2',marker='o',ls='-')
> leg=pylab.legend(loc='best')
> line=leg.get_lines()
> line[0].set_color('black')
> line[1].set_color('black')
> pylab.draw()
>
> I am attaching the results I see with the TkAgg baackend.  I am using version 1.1.0.
>
> I would expect the markers to change colors also.  However, I can't seem to find the markers recorded in the legend object to be able to change them.  Is this a bug or a feature (resetting the color could lead to inconsistencies between legend and plot)?


I think it's a design issue, there's no connection kept between plot
lines and the legend. You can change colors in the plot an then call
legend again instead.

Cheers

Goyo

There is imshow(), which will plot an array into the existing axes, as well
as figimage(), which will insert the image into the figure in general.
Each of these function have ways to control placement, particularly
figimage().

Without more information, I can't help much further.

Cheers!
Ben Root

On Thu, Aug 30, 2012 at 5:44 PM, Christoph Gohlke wrote:
>
> Try to remove 'geos_c' from the list of libraries in setup.py. Works for me.
>
> -                                libraries=['geos_c','geos']))
> +                                libraries=['geos']))

Thanks, that was it. It builds fine afterwards :)

Using this opportunity to thank you once more, for publicly
maintaining large and great selection of ready-made Python packages
for Windows, thus lowering the entrance level to broader Python sci
userbase

Best wishes
Cheers

On 30/08/2012 19:00, Fabien Lafont wrote:
> Actually I just want to do it on that plot not on all my future plot.
>
> 2012/8/30 Fabrice Silva <si...@lm...>:
>> Le jeudi 30 août 2012 à 19:48 +0200, Fabien Lafont a écrit :
>>> I just create two vectors from a .txt file and I plot them.
>>> I think I have the latest version of matplotlib. I have at least the
>>> last version of python(x,y)
>>>
>>>
>>> from pylab import*
>>>
>>> import matplotlib
>>> matplotlib.rcParams['xtick.labelsize'] = 20.0
>>
>> In your matplotlib config file
>> axes.titlesize      : 10      # fontsize of the axes title
>> axes.labelsize      : 10      # fontsize of the x any y labels
>> (see http://matplotlib.sourceforge.net/users/customizing.html )
>>
>>
>>
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>

I think you're looking for this 
http://matplotlib.sourceforge.net/users/whats_new.html#tick-params

-- 
Cheers.

Mark Lawrence.

Actually I just want to do it on that plot not on all my future plot.

2012/8/30 Fabrice Silva <si...@lm...>:
> Le jeudi 30 août 2012 à 19:48 +0200, Fabien Lafont a écrit :
>> I just create two vectors from a .txt file and I plot them.
>> I think I have the latest version of matplotlib. I have at least the
>> last version of python(x,y)
>>
>>
>> from pylab import*
>>
>> import matplotlib
>> matplotlib.rcParams['xtick.labelsize'] = 20.0
>
> In your matplotlib config file
> axes.titlesize      : 10      # fontsize of the axes title
> axes.labelsize      : 10      # fontsize of the x any y labels
> (see http://matplotlib.sourceforge.net/users/customizing.html )
>
>
>

Le jeudi 30 août 2012 à 19:48 +0200, Fabien Lafont a écrit :
> I just create two vectors from a .txt file and I plot them.
> I think I have the latest version of matplotlib. I have at least the
> last version of python(x,y)
> 
> 
> from pylab import*
> 
> import matplotlib
> matplotlib.rcParams['xtick.labelsize'] = 20.0

In your matplotlib config file 
axes.titlesize      : 10      # fontsize of the axes title
axes.labelsize      : 10      # fontsize of the x any y labels
(see http://matplotlib.sourceforge.net/users/customizing.html )

I just create two vectors from a .txt file and I plot them.
I think I have the latest version of matplotlib. I have at least the
last version of python(x,y)


from pylab import*

import matplotlib
matplotlib.rcParams['xtick.labelsize'] = 20.0

B5= genfromtxt("2012-05-14_RC3D3D2C1_D2D1_m18T_18T_Vg32V_20K.dat", usecols =(2))
RH5 = genfromtxt("2012-05-14_RC3D3D2C1_D2D1_m18T_18T_Vg32V_20K.dat",
usecols =(3))
#plot(B5,RH5,"-o",label = "2012-05-14_RC3D3D2C1_D2D1_m18T_18T_Vg32V_20K.dat")


xlabel(u"$ B (T)$", size= 30)
ylabel(u"$R_H (\Omega)$", size= 30)
grid()
legend()
show()

2012/8/30 Benjamin Root <ben...@ou...>:
>
>
> On Thu, Aug 30, 2012 at 1:38 PM, Fabien Lafont <laf...@gm...>
> wrote:
>>
>> I'm just trying to plot a graph and add a label to each axis of that graph
>> and change the labelsize of the ticks.
>>
>
> Could you post your code?  What you are describing shouldn't happen.  Also,
> which version of matplotlib are you using?
>
> Ben Root
>

On Thu, Aug 30, 2012 at 1:38 PM, Fabien Lafont <laf...@gm...>wrote:

> I'm just trying to plot a graph and add a label to each axis of that graph
> and change the labelsize of the ticks.
>
>
Could you post your code?  What you are describing shouldn't happen.  Also,
which version of matplotlib are you using?

Ben Root

I'm just trying to plot a graph and add a label to each axis of that graph
and change the labelsize of the ticks.

2012/8/30 Damon McDougall <dam...@gm...>

> On Thu, Aug 30, 2012 at 07:28:40PM +0200, Fabien Lafont wrote:
> > Thanks,
> >
> > I've found the problem. I use xlabel("name of my axis", size= 30) after
> > matplotlib.rcParams['xtick.
> > labelsize'] = 12.0 and it cancel it! Is it possible to have a name on the
> > axe and  matplotlib.rcParams['xticks.
> > labelsize'] = 12.0 ??
> >
>
> Wait a minute. What exactly are you trying to do?
>
> --
> Damon McDougall
> http://www.damon-is-a-geek.com
> B2.39
> Mathematics Institute
> University of Warwick
> Coventry
> West Midlands
> CV4 7AL
> United Kingdom
>

On Thu, Aug 30, 2012 at 07:28:40PM +0200, Fabien Lafont wrote:
> Thanks,
> 
> I've found the problem. I use xlabel("name of my axis", size= 30) after
> matplotlib.rcParams['xtick.
> labelsize'] = 12.0 and it cancel it! Is it possible to have a name on the
> axe and  matplotlib.rcParams['xticks.
> labelsize'] = 12.0 ??
>

Wait a minute. What exactly are you trying to do?

-- 
Damon McDougall
http://www.damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

Thanks,

I've found the problem. I use xlabel("name of my axis", size= 30) after
matplotlib.rcParams['xtick.
labelsize'] = 12.0 and it cancel it! Is it possible to have a name on the
axe and  matplotlib.rcParams['xticks.
labelsize'] = 12.0 ??




2012/8/30 Damon McDougall <dam...@gm...>

> On Thu, Aug 30, 2012 at 07:06:14PM +0200, Fabien Lafont wrote:
> > I've tried also but it returns an error:
> >
> >  matplotlib.rcParams['xticks.labelsize'] = 12.0
> >   File "C:\Python27\lib\site-packages\matplotlib\__init__.py", line 653,
> in
> > __setitem__
> >     See rcParams.keys() for a list of valid parameters.' % (key,))
> > KeyError: 'xticks.labelsize is not a valid rc parameter.See
> rcParams.keys()
> > for a list of valid parameters.'
> >
>
> It's actually 'xtick.labelsize'. No 's'.
>
> --
> Damon McDougall
> http://www.damon-is-a-geek.com
> B2.39
> Mathematics Institute
> University of Warwick
> Coventry
> West Midlands
> CV4 7AL
> United Kingdom
>

Hello,

I want to insert an image in a plot, how can I do?

Fabien

On Thu, Aug 30, 2012 at 07:06:14PM +0200, Fabien Lafont wrote:
> I've tried also but it returns an error:
> 
>  matplotlib.rcParams['xticks.labelsize'] = 12.0
>   File "C:\Python27\lib\site-packages\matplotlib\__init__.py", line 653, in
> __setitem__
>     See rcParams.keys() for a list of valid parameters.' % (key,))
> KeyError: 'xticks.labelsize is not a valid rc parameter.See rcParams.keys()
> for a list of valid parameters.'
>

It's actually 'xtick.labelsize'. No 's'.

-- 
Damon McDougall
http://www.damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

I've tried also but it returns an error:

 matplotlib.rcParams['xticks.labelsize'] = 12.0
  File "C:\Python27\lib\site-packages\matplotlib\__init__.py", line 653, in
__setitem__
    See rcParams.keys() for a list of valid parameters.' % (key,))
KeyError: 'xticks.labelsize is not a valid rc parameter.See rcParams.keys()
for a list of valid parameters.'


2012/8/30 Damon McDougall <dam...@gm...>

> On Thu, Aug 30, 2012 at 12:04:48PM -0400, Benjamin Root wrote:
> > On Thu, Aug 30, 2012 at 11:58 AM, Fabien Lafont <laf...@gm...
> >wrote:
> >
> > > There is no effect...
> > >
> > >
> > > 2012/8/30 Damon McDougall <dam...@gm...>
> > >
> > >> On Thu, Aug 30, 2012 at 05:50:18PM +0200, Fabien Lafont wrote:
> > >> > Hello,
> > >> >
> > >> > Do you know to change the size of the numbers under the axis?
> > >> >
> > >>
> > >> import matplotlib
> > >> matplotlib.rcParams['axes.labelsize'] = 12.0
> > >>
> > >> Hope this helps.
> > >>
> > >>
> > You want to modify "xtick.labelsize" or "ytick.labelsize".  By default,
> > they are "medium", but you can use a number for it as well.
> > "axes.labelsize" effects the axes labels, not the labels for the ticks.
> >
>
> Thanks for that, Ben. Apologies if I caused confusion.
>
> --
> Damon McDougall
> http://www.damon-is-a-geek.com
> B2.39
> Mathematics Institute
> University of Warwick
> Coventry
> West Midlands
> CV4 7AL
> United Kingdom
>

On Thu, Aug 30, 2012 at 12:04:48PM -0400, Benjamin Root wrote:
> On Thu, Aug 30, 2012 at 11:58 AM, Fabien Lafont <laf...@gm...>wrote:
> 
> > There is no effect...
> >
> >
> > 2012/8/30 Damon McDougall <dam...@gm...>
> >
> >> On Thu, Aug 30, 2012 at 05:50:18PM +0200, Fabien Lafont wrote:
> >> > Hello,
> >> >
> >> > Do you know to change the size of the numbers under the axis?
> >> >
> >>
> >> import matplotlib
> >> matplotlib.rcParams['axes.labelsize'] = 12.0
> >>
> >> Hope this helps.
> >>
> >>
> You want to modify "xtick.labelsize" or "ytick.labelsize".  By default,
> they are "medium", but you can use a number for it as well.
> "axes.labelsize" effects the axes labels, not the labels for the ticks.
>

Thanks for that, Ben. Apologies if I caused confusion.

-- 
Damon McDougall
http://www.damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

On Thu, Aug 30, 2012 at 12:53 PM, Fabien Lafont <laf...@gm...>wrote:

> Hi,
>
> But how can I do it for every graph and without creating axis?
>
>
>
We just told you:

import matplotlib
matplotlib.rcParams['xticks.labelsize'] = 12.0

Whenever a plot is made, it obtains the defaults from the rcParams.  Note,
if you are changing the rcParams *after* creating a plot, then this would
have no effect.

Cheers!
Ben Root

Hi,

But how can I do it for every graph and without creating axis?



I've seen this example:

import numpy as npimport matplotlib.pyplot as plt
# plt.figure creates a matplotlib.figure.Figure instancefig =
plt.figure()rect = fig.patch # a rectangle
instancerect.set_facecolor('lightgoldenrodyellow')
ax1 = fig.add_axes([0.1, 0.3, 0.4, 0.4])rect =
ax1.patchrect.set_facecolor('lightslategray')

for label in ax1.xaxis.get_ticklabels():
    # label is a Text instance
    label.set_color('red')
    label.set_rotation(45)
    label.set_fontsize(16)


I don't manage to use only  label.set_fontsize(16). It looks so
complicated.

2012/8/30 Benjamin Root <ben...@ou...>

>
>
> On Thu, Aug 30, 2012 at 11:58 AM, Fabien Lafont <laf...@gm...>wrote:
>
>> There is no effect...
>>
>>
>> 2012/8/30 Damon McDougall <dam...@gm...>
>>
>>> On Thu, Aug 30, 2012 at 05:50:18PM +0200, Fabien Lafont wrote:
>>> > Hello,
>>> >
>>> > Do you know to change the size of the numbers under the axis?
>>> >
>>>
>>> import matplotlib
>>> matplotlib.rcParams['axes.labelsize'] = 12.0
>>>
>>> Hope this helps.
>>>
>>>
> You want to modify "xtick.labelsize" or "ytick.labelsize".  By default,
> they are "medium", but you can use a number for it as well.
> "axes.labelsize" effects the axes labels, not the labels for the ticks.
>
> Cheers!
> Ben Root
>
>

On Thu, Aug 30, 2012 at 11:58 AM, Fabien Lafont <laf...@gm...>wrote:

> There is no effect...
>
>
> 2012/8/30 Damon McDougall <dam...@gm...>
>
>> On Thu, Aug 30, 2012 at 05:50:18PM +0200, Fabien Lafont wrote:
>> > Hello,
>> >
>> > Do you know to change the size of the numbers under the axis?
>> >
>>
>> import matplotlib
>> matplotlib.rcParams['axes.labelsize'] = 12.0
>>
>> Hope this helps.
>>
>>
You want to modify "xtick.labelsize" or "ytick.labelsize".  By default,
they are "medium", but you can use a number for it as well.
"axes.labelsize" effects the axes labels, not the labels for the ticks.

Cheers!
Ben Root

There is no effect...

2012/8/30 Damon McDougall <dam...@gm...>

> On Thu, Aug 30, 2012 at 05:50:18PM +0200, Fabien Lafont wrote:
> > Hello,
> >
> > Do you know to change the size of the numbers under the axis?
> >
>
> import matplotlib
> matplotlib.rcParams['axes.labelsize'] = 12.0
>
> Hope this helps.
>
> --
> Damon McDougall
> http://www.damon-is-a-geek.com
> B2.39
> Mathematics Institute
> University of Warwick
> Coventry
> West Midlands
> CV4 7AL
> United Kingdom
>

On Thu, Aug 30, 2012 at 05:50:18PM +0200, Fabien Lafont wrote:
> Hello,
> 
> Do you know to change the size of the numbers under the axis?
>

import matplotlib
matplotlib.rcParams['axes.labelsize'] = 12.0

Hope this helps.

-- 
Damon McDougall
http://www.damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

Hello,

Do you know to change the size of the numbers under the axis?

fabien

On 8/30/2012 2:25 AM, klo uo wrote:
> So I forked then cloned huge Basemap repository with idea to test latest code
>
> README file says it needs geos library, which README suggest:
>
>    nmake /f makefile.vc MSVC_VER=1500
>
> However, error is inevitable almost always and this time Google
> suggests that MS does not support C-99 (can't find inttypes.h).
> Easiest solution is to copy these:
> https://code.google.com/p/msinttypes include files in affected include
> folder and change line `#include <inttypes.h>` in platform.h to
> `#include "inttypes.h"` so it picks it up from current folder
>
> That was it, geos compiled and products ended in src subfolder
> relative to geos folder:
>
>    inlines.obj
>    geos.lib
>    geos_i.exp
>    geos_i.lib
>    geos.dll.manifest
>    geos.pdb
>    geos.ilk
>    geos.dll
>    geos_c_i.exp
>    geos_c_i.lib
>    geos_c.dll.manifest
>    geos_c.pdb
>    geos_c.ilk
>    geos_c.dll
>
> Now, Basemap setup has *nix scenario for GEOS_DIR so I changed setup.py to:
>
>    geos_include_dirs=["C:\\Temp\\basemap\\GEOS-3~1.3\\capi","C:\\Temp\\basemap\\GEOS-3~1.3\\include",numpy.get_include()]
>    geos_library_dirs=["C:\\Temp\\basemap\\GEOS-3~1.3\\src"]
>
> after I learned that `python setup.py build" fails if Basemap source
> is in path with spaces
>
> Still no luck, and thought to ask for kind help:
>
<snip>
>
>    LINK : fatal error LNK1181: cannot open input file 'geos_c.lib'
>
>    error: Command "c:\Program Files\Microsoft Visual Studio
> 9.0\VC\BIN\link.exe /DLL /nologo /INCREMENTAL:NO
>        /LIBPATH:C:\Temp\basemap\GEOS-3~1.3\src
> /LIBPATH:C:\Python27\libs /LIBPATH:C:\Python27\PCbuild geos_c.lib
> geos.lib
>        /EXPORT:init_geoslib
> build\temp.win32-2.7\Release\src/_geoslib.obj
> /OUT:build\lib.win32-2.7\_geoslib.pyd
>        /IMPLIB:build\temp.win32-2.7\Release\src\_geoslib.lib
> /MANIFESTFILE:build\temp.win32-2.7\Release\src\_geoslib.pyd.manifest"
>        failed with exit status 1181
>

Try to remove 'geos_c' from the list of libraries in setup.py. Works for me.

-                                libraries=['geos_c','geos']))
+                                libraries=['geos']))

Christoph

Le 30 août 2012 à 10:04, Fabien Lafont a écrit :

> Hello everyone,
> 
> I'm wondering if it's possible to have a matplotlib graph online ? I mean zoom drag and so on on a plot? 

Hi Fabien, 
http://code.google.com/p/mplh5canvas/
maybe ?
(haven't tried it myself)
-- 
Matthias