matplotlib-users Mailing List for matplotlib

Greetings,

I've noticed an issue when trying to create hatched plots. When I try
to create hatched plots with dashed lines, I get inconsistent behavior
between matplotlib.pyplot.show() and matplotlib.pyplot.savefig(). The
show() command will correctly (in terms of what I want) display the
hatched area as a dashed line, whereas savefig() will save an image
with the edge contour dashed, but the hatching as solid lines.

A self contained example script can be found here:
https://gist.github.com/2378109

For those who just want to see figures as to what I'm talking about
they can be found at the following.
What I want: http://www.patricktmarsh.com/tmp/correct_example.png
What I get: http://www.patricktmarsh.com/tmp/incorrect_example.png

What I want: http://www.patricktmarsh.com/tmp/correct_web.png
What I get: http://www.patricktmarsh.com/wp-content/gallery/2012images/20060406_day2_color.png


Thanks for any help!
Patrick
---
Patrick Marsh
Ph.D. Candidate / Liaison to the HWT
School of Meteorology / University of Oklahoma
Cooperative Institute for Mesoscale Meteorological Studies
National Severe Storms Laboratory
http://www.patricktmarsh.com

Confirmed, when I installed my python as a framework, (With homebrew, `brew
install python --framework`) the focus now works properly.
-E

On Thu, Apr 12, 2012 at 7:57 PM, Chris Laumann <cla...@ph...
> wrote:

> Actually I don't know about the apple supplied python, but I believe
> enthoughts python is installed as a framework.
>
> C
>
>
>
> On Apr 12, 2012, at 8:53 PM, Michiel de Hoon <mjl...@ya...> wrote:
>
> > --- On Wed, 4/11/12, Zachary Pincus <zac...@ya...> wrote:
> >> Hopefully someone who knows more about the OS X backend can
> >> comment here...
> >
> > It sounds like the Python you are using is not installed as a framework.
> Using the --enable-framework flag when compiling Python.
> >
> > -Michiel.
> >
> >
> ------------------------------------------------------------------------------
> > For Developers, A Lot Can Happen In A Second.
> > Boundary is the first to Know...and Tell You.
> > Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
> > http://p.sf.net/sfu/Boundary-d2dvs2
> > _______________________________________________
> > Matplotlib-users mailing list
> > Mat...@li...
> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>
> ------------------------------------------------------------------------------
> For Developers, A Lot Can Happen In A Second.
> Boundary is the first to Know...and Tell You.
> Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
> http://p.sf.net/sfu/Boundary-d2dvs2
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Jouni K. Seppänen <jk...@ik...> writes:

> I have a suggested fix for this at
> https://github.com/matplotlib/matplotlib/pull/817 

This is now merged into the v1.1.x branch, from which the fix should
propagate to the upcoming release.

-- 
Jouni K. Seppänen
http://www.iki.fi/jks

Dear all,

I was wondering if it is possible to use matplotlib from C++ directly
and I have found an example on how to do this on StackOverflow:

http://stackoverflow.com/questions/2509156/anyone-knows-a-matplotlib-equivalent-in-c-or-c

However, my question is about piping. Where could I find any tutorials
on how to pipe data directly to matplotlib so that I would not need to
right it to a file and then read it?

Also, does matplotlib benefit from usage of cython and if it does, are
there any things I should know before just trying to compile a
matplotlib script with cython?

All best,
Ignas A.

On Fri, Apr 13, 2012 at 2:21 AM, Jonathan Bruck <jdt...@gm...> wrote:

> Thanks for a response, but unfortunately it doesn't seem to work. I have
> some sample code on pastebin http://pastebin.com/W6JmbCsz in case the
> following does not email out well
>
> # -*- coding: utf-8 -*-
> """
> Created on Thu Apr 12 11:16:03 2012
>
> Using the current stable version of pythonxy on Windows 7 32bit
> Author: Jonathan
>
>
> Notice how the figure box is not placed correctly, effectively missing the
> legend entirely
> """
>
> import matplotlib.pyplot as plt
> import numpy as np
>
> x = np.arange(-2*np.pi, 2*np.pi, 0.1)
> fig = plt.figure(1, figsize=(8,6))
> ax = fig.add_subplot(111)
> ax.plot(x, np.sin(x), label='Sine')
> ax.plot(x, np.cos(x), label='Cosine')
> ax.plot(x, np.arctan(x), label='Inverse tan')
> handles, labels = ax.get_legend_handles_labels()
> lgd = ax.legend(handles, labels, loc=9, bbox_to_anchor=(0.5,-0.1))
> ax.grid('on')
> #fig.tight_layout()
> fig.savefig('samplefigure', bbox_extra_artists=(lgd,), bbox='tight')
>
>
>
I apologize, it was bbox_inches='tight', not bbox='tight'.  Because savefig
has to be able to take arbitrary kwargs that get passed down to the
backends, it does not check for incorrect kwargs.

Ben Root

Martin Mokrejs wrote:
> Hi,
>   I am trying to improve my code where I cannot find out why matplotlib-1.1.0 does not
> support colors specified as RG tuples. Here is an example.
> 
> 
> 
> import numpy as np
> import matplotlib.pyplot as plt
> 
> _nums = [160.0, 160.0, 160.0, 95.0, 160.0, 160.0]

Grr. This was the trick:

_nums = [[160.0], [160.0], [160.0], [95.0], [160.0], [160.0]]

What a wasteful list creation. :(


> _colors = [(0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255)]
> _legends = ['foo', 'foo', 'foo', 'blah', 'foo', 'foo']
> 
> plt.hist(_nums, histtype='bar', align='mid', color=_colors, log=False, label=_legends)
> 
> plt.show()
> 
> 
> The above code gives me:
> 
>   File "/usr/lib64/python2.7/site-packages/matplotlib/pyplot.py", line 2332, in hist
>     ret = ax.hist(x, bins, range, normed, weights, cumulative, bottom, histtype, align, orientation, rwidth, log, color, label, **kwargs)
>   File "/usr/lib64/python2.7/site-packages/matplotlib/axes.py", line 7598, in hist
>     raise ValueError("color kwarg must have one color per dataset")
> ValueError: color kwarg must have one color per dataset
> 
> 
> Mainly, I am suggesting the error message to be improved and to write out how many
> items were in data, color and legend iterables passed to the function. That would help

Would it tells me my problem is not with color but with data points life would be much easier. ;)


> in some cases albeit not with this example. That needs some other fix. ;)
> 
> I would like that one can also pass in a list of HTML-like colors, e.g. 'F0F8FF' or 0xF0F8FF
> would be valid.

Sorry, meant also '#F0F8FF', but now I have verified that they do work already:

_colors = ['b', 'b', 'b', 'r', 'b', 'b']
_colors = ['#C0C0C0', '#C0C0C0', '#C0C0C0', '#800000', '#C0C0C0', '#C0C0C0']
_colors = [(0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255)]

So this was all about *data* points to be wrapped in iterables. :((

Martin

Hi,
  I am trying to improve my code where I cannot find out why matplotlib-1.1.0 does not
support colors specified as RG tuples. Here is an example.



import numpy as np
import matplotlib.pyplot as plt

_nums = [160.0, 160.0, 160.0, 95.0, 160.0, 160.0]
_colors = [(0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255), (0.5019607843137255, 0.0, 0.5019607843137255)]
_legends = ['foo', 'foo', 'foo', 'blah', 'foo', 'foo']

plt.hist(_nums, histtype='bar', align='mid', color=_colors, log=False, label=_legends)

plt.show()


The above code gives me:

  File "/usr/lib64/python2.7/site-packages/matplotlib/pyplot.py", line 2332, in hist
    ret = ax.hist(x, bins, range, normed, weights, cumulative, bottom, histtype, align, orientation, rwidth, log, color, label, **kwargs)
  File "/usr/lib64/python2.7/site-packages/matplotlib/axes.py", line 7598, in hist
    raise ValueError("color kwarg must have one color per dataset")
ValueError: color kwarg must have one color per dataset


Mainly, I am suggesting the error message to be improved and to write out how many
items were in data, color and legend iterables passed to the function. That would help
in some cases albeit not with this example. That needs some other fix. ;)

I would like that one can also pass in a list of HTML-like colors, e.g. 'F0F8FF' or 0xF0F8FF
would be valid.

Thanks for your comments,
Martin

Thanks for a response, but unfortunately it doesn't seem to work. I have
some sample code on pastebin http://pastebin.com/W6JmbCsz in case the
following does not email out well

# -*- coding: utf-8 -*-
"""
Created on Thu Apr 12 11:16:03 2012

Using the current stable version of pythonxy on Windows 7 32bit
Author: Jonathan


Notice how the figure box is not placed correctly, effectively missing the
legend entirely
"""

import matplotlib.pyplot as plt
import numpy as np

x = np.arange(-2*np.pi, 2*np.pi, 0.1)
fig = plt.figure(1, figsize=(8,6))
ax = fig.add_subplot(111)
ax.plot(x, np.sin(x), label='Sine')
ax.plot(x, np.cos(x), label='Cosine')
ax.plot(x, np.arctan(x), label='Inverse tan')
handles, labels = ax.get_legend_handles_labels()
lgd = ax.legend(handles, labels, loc=9, bbox_to_anchor=(0.5,-0.1))
ax.grid('on')
#fig.tight_layout()
fig.savefig('samplefigure', bbox_extra_artists=(lgd,), bbox='tight')


On Thu, Apr 12, 2012 at 11:51 PM, Benjamin Root <ben...@ou...> wrote:

>
>
> On Thu, Apr 12, 2012 at 2:05 AM, Jonathan Bruck <jdt...@gm...>wrote:
>
>> Hi all,
>>
>> Forgive me as this is the first time I've posted here. I've asked a
>> question on StackOverFlow:
>>
>>
>> http://stackoverflow.com/questions/10101700/moving-matplotlib-legend-outside-of-the-axis-makes-it-cutoff-by-the-figure-box#comment12952803_10101700
>>
>> The question relates to adjusting the size of the figure box to
>> accommodate a large legend when the legend is placed below instead of on
>> top of the axes.
>>
>> I thought I'd post here to see if there are any other answers to avoiding
>> having the figure box cut off the bottom of the legend.
>>
>> Thanks
>>
>> Jonathan
>> --
>> There are no passengers on Spaceship Earth.  We are all crew.
>>
>> E: jdt...@gm...
>>
>>
> If you only care about saving the figure, the savefig() method can take
> bbox='tight' and bbox_extra_artists=[legnd_obj] arguments (assuming you
> save the legend to such a variable.  As for on-screen displays, I have yet
> to find a solution.
>
> Ben Root
>
>


-- 
There are no passengers on Spaceship Earth.  We are all crew.

Jonathan Bruck
E: jdt...@gm...
Mob: 0421188951
Bachelor of Engineering (Mechanical, Biomedical), Bachelor of Medical
Science
--

Hi all ,
Apologies if this was already posted here ..but there is video of John
Hunters advanced tutorial from the pydata conference 2012 courtesy Markana
 at

http://marakana.com/s/advanced_matplotlib_tutorial_with_library_author_john_hunter,1133/index.html#c132880

The link to the tutorial files is
http://bit.ly/Ancxno

Thanks
Hari

Actually I don't know about the apple supplied python, but I believe enthoughts python is installed as a framework. 

C



On Apr 12, 2012, at 8:53 PM, Michiel de Hoon <mjl...@ya...> wrote:

> --- On Wed, 4/11/12, Zachary Pincus <zac...@ya...> wrote:
>> Hopefully someone who knows more about the OS X backend can
>> comment here...
> 
> It sounds like the Python you are using is not installed as a framework. Using the --enable-framework flag when compiling Python.
> 
> -Michiel.
> 
> ------------------------------------------------------------------------------
> For Developers, A Lot Can Happen In A Second.
> Boundary is the first to Know...and Tell You.
> Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
> http://p.sf.net/sfu/Boundary-d2dvs2
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

When the -|> style was added to FancyArrowPatch, the purpose was to add 
an arrow style with a certain style shaft, but a solid head [1]. 
However, since the given linestyle is used for the outline of the head, 
we can have arrowheads that look very odd.  Here is the example input 
and output:

http://aleph.sagemath.org/?q=0b7e7b41-e6cc-4cfe-b176-e42ece3565c9

from matplotlib.pyplot import figure, show
from matplotlib.patches import FancyArrowPatch

fig = figure()
ax = fig.add_subplot(111,autoscale_on=False,)
p = FancyArrowPatch((0,0), (1,1), 
arrowstyle='-|>,head_width=8,head_length=16',lw=3,fc='k',ec='k',linestyle='dashed')
ax.add_patch(p)
show()

Is there an easy fix to make the arrowhead have a solid linestyle, even 
if the shaft is dashed?  Is this desirable to anyone else besides me?

Thanks,

Jason

[1] http://sourceforge.net/mailarchive/message.php?msg_id=21133511

--- On Wed, 4/11/12, Zachary Pincus <zac...@ya...> wrote:
> Hopefully someone who knows more about the OS X backend can
> comment here...

It sounds like the Python you are using is not installed as a framework. Using the --enable-framework flag when compiling Python.

-Michiel.