matplotlib-users Mailing List for matplotlib

I guess that is the problem. Thanks for the idea.

Regards, Ozgur



On Sat, 13 Aug 2011, Benjamin Root wrote:

> 
> 
> On Saturday, August 13, 2011, Özgür <ozg...@m3...> wrote:
> > Please forget it. For now the real problem is: I am plotting by using "contour" and I get the data values by using get_paths. But
> it does not include some part of the contour. A screenshot is attached. The colored lines are those returned by get_paths. But the
> black line is not included. I guess this is a bug?
> >
> > contour = plt.contour(xi,yi,zi,numberoflevels,linewidths=1,colors='k')
> > lev = contour.levels
> > col = contour.collections
> > polygons = []
> > for i in range(len(lev)):
> >        polygons.append (col[i].get_paths()[0].vertices)
> >
> > ... other code to plot the colored points...
> >
> 
> You are assuming that there are the same number of contours as levels.  In the general case, this can never be assumed.  Why not
> just simply loop over the collections(or, even better, use a list comprehension)?
> 
> Ben Root
>

Hi all,
I'd like to ask a probably basic qustion about backends in matplotlib
- specifically with regard to WXAgg and WX.
What are the differences of the two, and what are the supposed
usecases, and maybe restrictions for them?

Using matplotlib/pylab pylab in a rather basic way I encountered
problems with both backends and would like to clarify the differences
before i try to find a fix.

using wxAgg I am not able to display unicode text (some accents,
non-latin scripts etc. e.g. in pylab.title(...) and otherwhere .
Just changing:
matplotlib.use('WXAgg')
to
matplotlib.use('WX')
seemed to fix this, and the graph gui also looks more consistent with
my wxPython app.

However, using the current wxPython 2.9.2.1. (python 2.7.2, Win 7,
Czech) I see a more serious bug in the wx-backend (basically failing
to display the graph at all
(cf. http://sourceforge.net/mailarchive/message.php?msg_id=27902016 ),
which doesn't apply for WXAgg.

Coud someone maybe provide some hints, where I should look into to
understand the backend differences better?

Thanks in advance,

Regards,
    Vlastimil Brom

On Saturday, August 13, 2011, Özgür <ozg...@m3...> wrote:
> Please forget it. For now the real problem is: I am plotting by using
"contour" and I get the data values by using get_paths. But it does not
include some part of the contour. A screenshot is attached. The colored
lines are those returned by get_paths. But the black line is not included. I
guess this is a bug?
>
> contour = plt.contour(xi,yi,zi,numberoflevels,linewidths=1,colors='k')
> lev = contour.levels
> col = contour.collections
> polygons = []
> for i in range(len(lev)):
>        polygons.append (col[i].get_paths()[0].vertices)
>
> ... other code to plot the colored points...
>

You are assuming that there are the same number of contours as levels.  In
the general case, this can never be assumed.  Why not just simply loop over
the collections(or, even better, use a list comprehension)?

Ben Root

Please forget it. For now the real problem is: I am plotting by using 
"contour" and I get the data values by using get_paths. But it does not 
include some part of the contour. A screenshot is attached. The colored 
lines are those returned by get_paths. But the black line is not included. 
I guess this is a bug?

contour = plt.contour(xi,yi,zi,numberoflevels,linewidths=1,colors='k')
lev = contour.levels
col = contour.collections
polygons = []
for i in range(len(lev)):
 	polygons.append (col[i].get_paths()[0].vertices)

... other code to plot the colored points...



On Sat, 13 Aug 2011, Özgür wrote:

> Hi, when I use the function "contour" sometimes I get a broken contour
> line. This should be continuous. Probably this is a bug. Does anyone know
> anything about this issue?
>
> Regards,
> Ozgur
>
> ------------------------------------------------------------------------------
> FREE DOWNLOAD - uberSVN with Social Coding for Subversion.
> Subversion made easy with a complete admin console. Easy
> to use, easy to manage, easy to install, easy to extend.
> Get a Free download of the new open ALM Subversion platform now.
> http://p.sf.net/sfu/wandisco-dev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

On Saturday, August 13, 2011, Özgür <ozg...@m3...> wrote:
> Hi, when I use the function "contour" sometimes I get a broken contour
> line. This should be continuous. Probably this is a bug. Does anyone know
> anything about this issue?
>
> Regards,
> Ozgur
>

Could you please post an image of what you are seeing, and possibly the data
used to create this?  Also, which version of mpl are you using?

Ben Root

Hi, when I use the function "contour" sometimes I get a broken contour 
line. This should be continuous. Probably this is a bug. Does anyone know 
anything about this issue?

Regards,
Ozgur

Hi,

I am trying to figure out how to produce a mixed-mode rendering PDF of a
contourf plot.

I tried something like this:

 >>> from pylab import meshgrid, sin, cos, linspace, contourf,  
savefig, clf
 >>> x, y = meshgrid(*(linspace(-1,1,500),)*2)
 >>> z = sin(20*x**2)*cos(30*y)
 >>> c = contourf(x,y,z,30,rasterized=True)
 >>> savefig('tst0.pdf')

but this does not work.  (The QuadContourSet c does not support  
set_rasterized()
so the rasterized argument is just ignored.  Is the ignoring of the  
argument a
bug or a feature?)

I tried calling set_rasterized() on all of the of PathCollection  
objects in
c.collections but this did not help:

 >>> for pc in c.collections:
 >>>    pc.set_rasterized(True)
 >>> savefig('tst1.pdf')

While this helps, there are still 30 rasterized patches which renders  
slowly as
a PDF.

I found one possible solution: make a custom Collection subclass as  
shown below,
then remove all of the PathCollection instances from the figure,  
inserting
instead this new collection as a single rasterized entity:

 >>> insert(c)            # See code below for definition of insert()
 >>> savefig('tst2.pdf')

The file-sizes are:

$ ls -lah *.pdf
... 4.2M Aug 13 02:38 tst0.pdf
... 1.8M Aug 13 02:38 tst1.pdf
... 629K Aug 13 02:40 tst2.pdf

but tst2.pdf looks as good as tst1.pdf, and is much faster to load.   
(There are
some faint white lines between the patches similar to those discussed  
before:

http://thread.gmane.org/gmane.comp.python.matplotlib.devel/8540/focus=8590

but the rasterizer here is in the PDF backend, so maybe something can  
be done?
Should I look into this as a potential bug?  I am using the version of
matplotlib included with the latest EPD on a 32 bit Mac.)

Question: Is there some way of doing this with current matplotlib API,  
or is
some sort of extension like this required?  If the latter, is there  
anything I
should be careful about? (The ListCollection should probably be  
inserted in
contourf rather than retroactively like this.)

Thanks,
Michael.

-----------------------------------------------
from matplotlib.collections import Collection
from matplotlib.artist import allow_rasterization
from matplotlib import pyplot as plt

class ListCollection(Collection):
     def __init__(self, collections, **kwargs):
         Collection.__init__(self, **kwargs)
         self.set_collections(collections)
     def set_collections(self, collections):
         self._collections = collections
     def get_collections(self):
         return self._collections
     @allow_rasterization
     def draw(self, renderer):
         for _c in self._collections:
             _c.draw(renderer)

def insert(c):
     collections = c.collections
     for _c in collections:
         _c.remove()
     cc = ListCollection(collections, rasterized=True)
     ax = plt.gca()
     ax.add_artist(cc)
     return cc

Hello!

I'm using matplotlib in my wx.Frame to draw some arrows and text at a 
certain position. The number and position of the arrows is based on 
previously created data. I do some calculations but basically it's like 
this:

arrow = primer
for each primer:
draw one arrow at position y - 0.05 to the previos arrow (primer_y - 0.05).
The x position of the arrow comes from the data, which is calculated 
and scaled (not so important know, it's just to now where the arrow 
should be at primer_x).

Everything works fine until I have many arrows to draw. E.g. in my code 
example the data contain 62 arrows, where 18 are drawn correctly, the 
next 3 are missing but the text is there and the rest is completly 
missing.

Does anyone knows what could be the problem? I tried allready to change 
the FigureSize but it's only stretch the arrows.

Here's a quick and dirty working example, data is included in the code:
http://pastebin.com/7mQmZm2c

Any help is highly appreciated!
Thanks in advance!
Stefanie

Hi Ben, Eric and Paul,

Paul's partial solution is actually exactly what I need. Thanks!

-Chris

On Friday, August 12, 2011 07:26:22 PM Benjamin Root wrote:
> On Friday, August 12, 2011, Christopher Brown <c-...@as...> wrote:
> > Thanks Ben,
> > 
> > I should have been more clear. In fact I do want to create an entirely
> > new figure containing the already created axes. This might not be the
> > best way
> 
> to
> 
> > do what I want to do, but the docs for add_axes states that passing an
> 
> axes
> 
> > instance will add that instance to the figure. I must be misunderstanding
> > something.
> > 
> > -Chris
> 
> So, are you specifically wanting the same axes object in two figures?  If
> so, as Paul said, it is technically possible (using the method that Eric
> mentioned), but that mode is currently unsupported and we can not guarantee
> that things will work as expected.  Matplotlib assumes that an axes object
> has a single parent, and breaking that assumption can cause issues down the
> line.
> 
> However, if you don't intend to have the same axes appear twice, then the
> question becomes why do you need to create the figure after the axes?
> 
> Ben Root

On 08/12/2011 03:58 PM, Christopher Brown wrote:
> Thanks Ben,
>
> I should have been more clear. In fact I do want to create an entirely new
> figure containing the already created axes. This might not be the best way to
> do what I want to do, but the docs for add_axes states that passing an axes
> instance will add that instance to the figure. I must be misunderstanding
> something.

The docstring is seriously misleading; I will fix it.

Eric

>
> -Chris
>
> On Friday, August 12, 2011 05:57:17 PM Benjamin Root wrote:
>> Yes, you are making it harder on yourself...
>>
>> On Friday, August 12, 2011, Christopher Brown<c-...@as...>  wrote:
>>> I feel like I'm doing this right, but it doesn't work. Any clues?
>>>
>>> from matplotlib import pyplot as pp
>>>
>>> pp.plot((1,2,3))
>>> ax = pp.gca()
>>
>> At this point, a figure is implicitly created because none exists at this
>> point.  The ax object is implicitly added to that.
>>
>>> f = pp.figure(num=2)
>>
>> This will be a completely new figure.
>>
>>> print 'first: %i' % ax.figure.number
>>> print 'second: %i' % f.number
>>> f.add_axes(ax)
>>
>> Since ax was already attached, it can't be in two figures at once, so the
>> assertion fails.
>>
>>> yields:
>>>
>>> Traceback (most recent call last):
>>>    File "<stdin>", line 1, in<module>
>>>    File "C:\Python27\lib\site-packages\matplotlib\figure.py", line 606,
>>>
>>> in add_axes
>>>
>>>      assert(a.get_figure() is self)
>>>
>>> AssertionError
>>
>> So, just simply create your figure first, and call ax = f.gca() after.
>>   There should be no need to call add_axes except in very special
>> situations.
>>
>> I hope that helps!
>>
>> Ben Root
>
> ------------------------------------------------------------------------------
> FREE DOWNLOAD - uberSVN with Social Coding for Subversion.
> Subversion made easy with a complete admin console. Easy
> to use, easy to manage, easy to install, easy to extend.
> Get a Free download of the new open ALM Subversion platform now.
> http://p.sf.net/sfu/wandisco-dev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On 08/12/2011 03:00 PM, Paul Ivanov wrote:
> Christopher Brown, on 2011-08-12 16:56,  wrote:
>> I feel like I'm doing this right, but it doesn't work. Any clues?
>>
>> from matplotlib import pyplot as pp
>>
>> pp.plot((1,2,3))
>> ax = pp.gca()
>> f = pp.figure(num=2)
>> print 'first: %i' % ax.figure.number
>> print 'second: %i' % f.number
>> f.add_axes(ax)
>>
>> yields:
>>
>> Traceback (most recent call last):
>>     File "<stdin>", line 1, in<module>
>>     File "C:\Python27\lib\site-packages\matplotlib\figure.py", line 606,
>> in add_axes
>>       assert(a.get_figure() is self)
>> AssertionError
>
> Right now, matplotlib allows an axes to participate in only one
> figure. You can get part-way to what you want by setting
> ax.figure=f before calling f.add_axes(ax). Then, you'll have ax
> in both figures, but its transforms are still rooted in the first
> figure, so resizing the first figure, you'll see changes to the
> size of the plot in the second figure.

It's worse than that; it really just doesn't work in any useful way, and 
it was never intended to do so.  Try it:

In [1]: fig = gcf()

In [2]: fig.add_axes?

In [3]: plot([1,2,3])
Out[3]: [<matplotlib.lines.Line2D at 0x335cf10>]

In [4]: ax = gca()

In [5]: fig2 = figure()

In [6]: ax.set_figure(fig2)

In [7]: fig2.add_axes(ax)
Out[7]: <matplotlib.axes.AxesSubplot at 0x333b250>

In [8]: ax.plot([3,4,5])
Out[8]: [<matplotlib.lines.Line2D at 0x3368890>]

In [9]: draw()

Eric

>
> best,
>
>
>
> ------------------------------------------------------------------------------
> FREE DOWNLOAD - uberSVN with Social Coding for Subversion.
> Subversion made easy with a complete admin console. Easy
> to use, easy to manage, easy to install, easy to extend.
> Get a Free download of the new open ALM Subversion platform now.
> http://p.sf.net/sfu/wandisco-dev2dev
>
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On Friday, August 12, 2011, Christopher Brown <c-...@as...> wrote:
> Thanks Ben,
>
> I should have been more clear. In fact I do want to create an entirely new
> figure containing the already created axes. This might not be the best way
to
> do what I want to do, but the docs for add_axes states that passing an
axes
> instance will add that instance to the figure. I must be misunderstanding
> something.
>
> -Chris
>

So, are you specifically wanting the same axes object in two figures?  If
so, as Paul said, it is technically possible (using the method that Eric
mentioned), but that mode is currently unsupported and we can not guarantee
that things will work as expected.  Matplotlib assumes that an axes object
has a single parent, and breaking that assumption can cause issues down the
line.

However, if you don't intend to have the same axes appear twice, then the
question becomes why do you need to create the figure after the axes?

Ben Root

Thanks Ben,

I should have been more clear. In fact I do want to create an entirely new 
figure containing the already created axes. This might not be the best way to 
do what I want to do, but the docs for add_axes states that passing an axes 
instance will add that instance to the figure. I must be misunderstanding 
something. 

-Chris

On Friday, August 12, 2011 05:57:17 PM Benjamin Root wrote:
> Yes, you are making it harder on yourself...
> 
> On Friday, August 12, 2011, Christopher Brown <c-...@as...> wrote:
> > I feel like I'm doing this right, but it doesn't work. Any clues?
> > 
> > from matplotlib import pyplot as pp
> > 
> > pp.plot((1,2,3))
> > ax = pp.gca()
> 
> At this point, a figure is implicitly created because none exists at this
> point.  The ax object is implicitly added to that.
> 
> > f = pp.figure(num=2)
> 
> This will be a completely new figure.
> 
> > print 'first: %i' % ax.figure.number
> > print 'second: %i' % f.number
> > f.add_axes(ax)
> 
> Since ax was already attached, it can't be in two figures at once, so the
> assertion fails.
> 
> > yields:
> > 
> > Traceback (most recent call last):
> >   File "<stdin>", line 1, in <module>
> >   File "C:\Python27\lib\site-packages\matplotlib\figure.py", line 606,
> > 
> > in add_axes
> > 
> >     assert(a.get_figure() is self)
> > 
> > AssertionError
> 
> So, just simply create your figure first, and call ax = f.gca() after.
>  There should be no need to call add_axes except in very special
> situations.
> 
> I hope that helps!
> 
> Ben Root

On 08/12/2011 01:56 PM, Christopher Brown wrote:
> I feel like I'm doing this right, but it doesn't work. Any clues?
>
> from matplotlib import pyplot as pp
>
> pp.plot((1,2,3))
> ax = pp.gca()
> f = pp.figure(num=2)
> print 'first: %i' % ax.figure.number
> print 'second: %i' % f.number
> f.add_axes(ax)


That looks pretty convoluted.  What are you really trying to do, and 
why?  It looks like you are trying to transplant an Axes from one figure 
to another; add_axes doesn't do that.  I'm not sure add_axes should even 
include a signature in which an axes instance is the argument; it is 
part of the behavior in which add_axes returns an existing Axes if it 
can find one that matches the requested args and kwargs.  There was 
probably a good reason for this, but it looks to me like excessive 
overloading of add_axes--making it do something other than what its name 
clearly states it should do.

The Figure instance is a required argument when creating an Axes 
instance.  Axes are not designed to be copied or moved between figures, 
as far as I can see.

Eric

>
> yields:
>
> Traceback (most recent call last):
>     File "<stdin>", line 1, in<module>
>     File "C:\Python27\lib\site-packages\matplotlib\figure.py", line 606,
> in add_axes
>       assert(a.get_figure() is self)
> AssertionError
>
>

Christopher Brown, on 2011-08-12 16:56,  wrote:
> I feel like I'm doing this right, but it doesn't work. Any clues?
> 
> from matplotlib import pyplot as pp
> 
> pp.plot((1,2,3))
> ax = pp.gca()
> f = pp.figure(num=2)
> print 'first: %i' % ax.figure.number
> print 'second: %i' % f.number
> f.add_axes(ax)
> 
> yields:
> 
> Traceback (most recent call last):
>    File "<stdin>", line 1, in <module>
>    File "C:\Python27\lib\site-packages\matplotlib\figure.py", line 606, 
> in add_axes
>      assert(a.get_figure() is self)
> AssertionError

Right now, matplotlib allows an axes to participate in only one
figure. You can get part-way to what you want by setting
ax.figure=f before calling f.add_axes(ax). Then, you'll have ax
in both figures, but its transforms are still rooted in the first
figure, so resizing the first figure, you'll see changes to the
size of the plot in the second figure.

best,
-- 
Paul Ivanov
314 address only used for lists,  off-list direct email at:
http://pirsquared.org | GPG/PGP key id: 0x0F3E28F7 

Yes, you are making it harder on yourself...

On Friday, August 12, 2011, Christopher Brown <c-...@as...> wrote:
> I feel like I'm doing this right, but it doesn't work. Any clues?
>
> from matplotlib import pyplot as pp
>
> pp.plot((1,2,3))
> ax = pp.gca()

At this point, a figure is implicitly created because none exists at this
point.  The ax object is implicitly added to that.

> f = pp.figure(num=2)

This will be a completely new figure.

> print 'first: %i' % ax.figure.number
> print 'second: %i' % f.number
> f.add_axes(ax)
>

Since ax was already attached, it can't be in two figures at once, so the
assertion fails.

> yields:
>
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "C:\Python27\lib\site-packages\matplotlib\figure.py", line 606,
> in add_axes
>     assert(a.get_figure() is self)
> AssertionError
>
>

So, just simply create your figure first, and call ax = f.gca() after.
 There should be no need to call add_axes except in very special situations.

I hope that helps!

Ben Root