matplotlib-users Mailing List for matplotlib

Tom Aldcroft wrote:
> I've run into a case where the rendering in a line plot is incomplete
> and some lines are not drawn at all.  Basically I have a dataset (see
> below) where I know two points go to a value of zero.   When I plot
> the points and do interactive pan/zoom sometimes the line going down
> to zero is there and sometimes not.  After a bunch of playing around
> trying to reduce the problem to a minimum the script below is what I
> got.  If I tried to make the 'y' array much smaller then the problem
> went away.  I played with antialias and interactive settings with no
> benefit.  Changing the window size can also produce the same effect
> I'm describing.  Finally, when I use savefig to save in various
> formats the results varied, perhaps just a side-effect of the size of
> the saved figure.
> 
> I'm using Matplotlib 0.99.1.1 built from source with the TkAgg backend
> on CentOS-5 with python 2.6.  This same problem was also evident using
> GtkAgg and MacOSX backends so I don't think the details of my build
> are relevant (but I can supply if needed).
> 
> Thanks, Tom

It works for me in svn, so I am pretty sure you have run into the 
infamous path simplification bug.  If you are building from source 
anyway, then just build from svn.  Alternatively, you can put

path.simplify : False

in your matplotlibrc file to disable path simplification.

Eric

> 
> import numpy
> import matplotlib.pyplot as plt
> 
> y = numpy.array([
>  4., 2., 2., 3., 3., 2., 2., 6., 6., 5., 5., 4., 4.,
>  7., 7., 2., 2., 4., 4., 2., 2., 2., 2., 4., 4., 4.,
>  4., 4., 4., 7., 7., 3., 3., 5., 5., 4., 4., 5., 5.,
>  4., 4., 7., 7., 6., 6., 2., 2., 2., 2., 5., 5., 4.,
>  4., 4., 4., 6., 6., 3., 3., 4., 4., 3., 3., 2., 2.,
>  3., 3., 4., 4., 4., 4., 4., 4., 6., 6., 5., 5., 4.,
>  4., 7., 7., 3., 3., 4., 4., 4., 4., 5., 5., 4., 4.,
>  7., 7., 3., 3., 4., 4., 4., 4., 6., 6., 4., 4., 4.,
>  4., 4., 4., 2., 2., 5., 5., 6., 6., 3., 3., 5., 5.,
>  4., 4., 0., 0., 5., 5., 1., 1., 4., 4., 5., 5., 4.])
> 
> plt.figure()
> plt.plot(y)
> 
> plt.figure()
> plt.plot(y)
> plt.xlim(-7200, 6500)
> # Does it go down to 0 now?
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

I've run into a case where the rendering in a line plot is incomplete
and some lines are not drawn at all.  Basically I have a dataset (see
below) where I know two points go to a value of zero.   When I plot
the points and do interactive pan/zoom sometimes the line going down
to zero is there and sometimes not.  After a bunch of playing around
trying to reduce the problem to a minimum the script below is what I
got.  If I tried to make the 'y' array much smaller then the problem
went away.  I played with antialias and interactive settings with no
benefit.  Changing the window size can also produce the same effect
I'm describing.  Finally, when I use savefig to save in various
formats the results varied, perhaps just a side-effect of the size of
the saved figure.

I'm using Matplotlib 0.99.1.1 built from source with the TkAgg backend
on CentOS-5 with python 2.6.  This same problem was also evident using
GtkAgg and MacOSX backends so I don't think the details of my build
are relevant (but I can supply if needed).

Thanks, Tom

import numpy
import matplotlib.pyplot as plt

y = numpy.array([
 4., 2., 2., 3., 3., 2., 2., 6., 6., 5., 5., 4., 4.,
 7., 7., 2., 2., 4., 4., 2., 2., 2., 2., 4., 4., 4.,
 4., 4., 4., 7., 7., 3., 3., 5., 5., 4., 4., 5., 5.,
 4., 4., 7., 7., 6., 6., 2., 2., 2., 2., 5., 5., 4.,
 4., 4., 4., 6., 6., 3., 3., 4., 4., 3., 3., 2., 2.,
 3., 3., 4., 4., 4., 4., 4., 4., 6., 6., 5., 5., 4.,
 4., 7., 7., 3., 3., 4., 4., 4., 4., 5., 5., 4., 4.,
 7., 7., 3., 3., 4., 4., 4., 4., 6., 6., 4., 4., 4.,
 4., 4., 4., 2., 2., 5., 5., 6., 6., 3., 3., 5., 5.,
 4., 4., 0., 0., 5., 5., 1., 1., 4., 4., 5., 5., 4.])

plt.figure()
plt.plot(y)

plt.figure()
plt.plot(y)
plt.xlim(-7200, 6500)
# Does it go down to 0 now?

imshow has a number of different interpolation methods you can try:

    
http://matplotlib.sourceforge.net/api/pyplot_api.html#matplotlib.pyplot.imshow

Mike

John wrote:
> Hello,
>
> I'm trying to plot 0.5x0.5 global data. I'm currently using
> pcolormesh. I have a log color scale and haven't been able to use
> contourf succesfully for some reason. If I use imshow, it seems to be
> too 'bleedy'.
>
> Is there a method for plotting that does some more 'smoothing'
> interpolation of the data than pcolormesh? Is there something I'm
> doing wrong that contourf doesn't work with log data?
>
> Thanks,
> john
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>   

-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Many thanks!

carlos

On Thu, Apr 22, 2010 at 18:54, Gökhan Sever <gok...@gm...> wrote:
>
>
> On Thu, Apr 22, 2010 at 4:44 PM, Carlos Grohmann <car...@gm...>
> wrote:
>>
>> Hi all,
>>
>> I'm getting this weird behavior of legend. I always get two symbols in
>> front of the labels, instead of one.
>> Did anyone experienced this before?
>>
>> I'm working with MPL in a WxPython app.
>>
>> code snippet:
>>
>>
>> [code]
>>            for i in range(len(self.idxPlan)):
>>                x = self.PeigenList[i][14] + (self.PeigenList[i][15] / 2)
>>                y = self.PeigenList[i][15] * sqrt3_2
>>                axes.plot(x,y, self.PProps[i][3], c=self.PProps[i][2],
>> ms=self.PProps[i][4],label=self.Pname[i])
>>
>>            axes.legend(bbox_to_anchor=(0.85, 0.75), loc=2,
>> prop=FontProperties(size='small'))
>>
>>            axes.set_xlim(-0.1,1.2)
>>            axes.set_ylim(-0.1,1.2)
>>            self.dataCanvas.draw()
>>
>> [\code]
>>
>>
>> tks
>> --
>> Prof. Carlos Henrique Grohmann - Geologist D.Sc.
>> Institute of Geosciences - Univ. of São Paulo, Brazil
>> http://www.igc.usp.br/pessoais/guano
>> Linux User #89721
>> ________________
>> Can’t stop the signal.
>
> Hi,
>
> Try using numpoints keyword:
>
> plt.legend(numpoints=1)
>
>
> --
> Gökhan
>



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

Hi,

I'm having issues with translucent patches. The following code

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as mpl
from matplotlib.patches import Circle

fig = mpl.figure()
ax = fig.add_subplot(111)
c = Circle((0.25,0.5), radius=0.2, facecolor='none', edgecolor='green', alpha=0.2)
ax.add_patch(c)
c = Circle((0.75,0.5), radius=0.2, facecolor='blue', edgecolor='green', alpha=0.2)
ax.add_patch(c)
fig.savefig('test_alpha.png')

Produces a plot where the two circles have different borders - the left circle has a border with alpha=1.0 and the right circle has a border with alpha=0.2. Should I submit this as a bug report?

I am also having issues with the following script:

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as mpl
from matplotlib.patches import Circle

import time

fig = mpl.figure()
ax = fig.add_subplot(111)
ax.add_patch(Circle((1.e10,1.e10), radius=1.e9))

time1 = time.time()

fig.savefig('test_perf.png')

print "Time to save: %.3f" % (time.time() - time1)

The issue is that the time to save is typically 3 seconds or so, up to 5 seconds in some caes (way more than normal), and the axes do not autoscale to show the patch (I imagine the two issues are related). Is this a bug, or just due to the way matplotlib works?

Thanks for any help,

Thomas

On Thu, Apr 22, 2010 at 7:38 PM, John Hunter <jd...@gm...> wrote:

>
> The standard trick is to make the ticklabels on one axes invisible
>
> for label in ax.get_xticklabels():
>    label.set_visible(False)
>
>
> JDH
>

Thanks John, This answer seems familiar to me :)  You have caught my
intention correctly --not completely clearing the x-axis but rather
turning-off the labels.

Your trick works independent from sharing state of an x-axis.

Alternatively, I use ax.xaxis.set_ticklabels ("")  to clear tick-labels when
I don't share the axis.

-- 
Gökhan

On 23/04/2010 10:30 AM, wil...@co... wrote:
> I use pylab.plot(x,y) for exploring or debugging some functions or 
> subroutines.
> I would like to cycle through 2 or more plot windows, in a simple way 
> that won't force me to entangle the code in bothersome ways.  But I 
> can't seem to get more than one plot window to open per launch.
>
> Here is an example of what I want to have happen:
>
>   The program calls subroutine A, and creates a plot so that  I can 
> see what it has done.
>   The plot opens, and I see that A is doing what I want, then I close 
> the plot window (I interactively click the close button).
>   Then the program calls subroutine B, and creates another plot so I 
> can check its results.
>   The plot opens, I see what I want and close it.
>   And so on.
>
> But I can't figure out how to make that work.  Only one plot ever shows.
>
> Here is a very minimal example of what I have tried:
>
>         import numpy as N
>         import pylab as P
>
>         x = N.linspace(0,6,100)
>         y = N.sin(x)
>
>         P.plot(x,y)
>         P.show()     # At this point, I want to see what I've got, and 
> then move on.
>
>         y = -y
>         P.plot(x,y)
>         P.show()     # This plot will not show.
>
> The answer for me is not to use "subplot()".  I already use that 
> often.  But I want these plots to stand on their own, so that I don't 
> have to entangle the different pieces of code just for some 
> exploratory plotting.
>
> I hope someone has an easy solution!
> Thanks,
> Bill

Hi,

I gave a solution just a couple of days ago, I believe the similar 
approach could solve this problem also
http://old.nabble.com/plotting-in-a-loop-td28306656.html

Essentially, . each time you press a button on graph (not the close 
button) you generate an event which would call an update function, you 
could farm out work to perform subroutine B here, then replot the result.

Steve

Hello,

I'm trying to plot 0.5x0.5 global data. I'm currently using
pcolormesh. I have a log color scale and haven't been able to use
contourf succesfully for some reason. If I use imshow, it seems to be
too 'bleedy'.

Is there a method for plotting that does some more 'smoothing'
interpolation of the data than pcolormesh? Is there something I'm
doing wrong that contourf doesn't work with log data?

Thanks,
john

On Apr 22, 2010, at 12:28 PM, Gökhan Sever <gok...@gm...>  
wrote:

> Hello,
>
> Consider my simple test case:
>
> import matplotlib.pyplot as plt
>
> fp = plt.figure()
> ax1 = fp.add_subplot(4,1,1)
> ax1.plot(range(10))
> ax2 = fp.add_subplot(4,1,2)
> ax2.plot(range(10))
> ax1.xaxis.set_major_locator (plt.NullLocator ())
>
> ax3 = fp.add_subplot(4,1,3)
> ax3.plot(range(10))
> # Can't turn off
> ax3.xaxis.set_major_locator (plt.NullLocator ())
> ax4 = fp.add_subplot(4,1,4, sharex=ax3)
> ax4.plot(range(10))
> # Turn-off both ax3 and ax4
> #ax3.xaxis.set_major_locator (plt.NullLocator ())
>
> plt.show()
>
> When I share x-axis in between two plots I can't turn-off one axis'  
> major ticks. Depends on the location of the function call it either  
> turn nothing or all. Is it intended or a mal-functioning? (Shared  
> axis is for my screen view, I can live without it by turning off  
> sharex option since it won't make any difference for the final saved  
> figures.)
>

The standard trick is to make the ticklabels on one axes invisible

for label in ax.get_xticklabels():
     label.set_visible(False)


JDH

I use pylab.plot(x,y) for exploring or debugging some functions or subroutines. 
I would like to cycle through 2 or more plot windows, in a simple way that won't force me to entangle the code in bothersome ways. But I can't seem to get more than one plot window to open per launch. 

Here is an example of what I want to have happen: 

The program calls subroutine A, and creates a plot so that I can see what it has done. 
The plot opens, and I see that A is doing what I want, then I close the plot window (I interactively click the close button). 
Then the program calls subroutine B, and creates another plot so I can check its results. 
The plot opens, I see what I want and close it. 
And so on. 

But I can't figure out how to make that work. Only one plot ever shows. 

Here is a very minimal example of what I have tried: 

import numpy as N 
import pylab as P 

x = N.linspace(0,6,100) 
y = N.sin(x) 

P.plot(x,y) 
P.show() # At this point, I want to see what I've got, and then move on. 

y = -y 
P.plot(x,y) 
P.show() # This plot will not show. 

The answer for me is not to use "subplot()". I already use that often. But I want these plots to stand on their own, so that I don't have to entangle the different pieces of code just for some exploratory plotting. 

I hope someone has an easy solution! 
Thanks, 
Bill