matplotlib-users Mailing List for matplotlib

ׁHi,
Sorry for the late reply.
Yes the solution suggested in reply to my bug report was indeed helpful !

Thanks for the help !

Oz

On Wed, Aug 4, 2010 at 5:01 PM, Benjamin Root <ben...@ou...> wrote:

> On Wed, Aug 4, 2010 at 12:26 AM, Oz Nahum <na...@gm...> wrote:
>
>> Hi Everyone  !
>> This is urgent, I have to finish some plots by tomorrow, and I totally
>> lost the ability to work with python matplotlib -
>> "import pylab"
>> "from pylab import *"
>>
>> all yield the following error:
>>
>> [code]Traceback (most recent call last):
>>   File "<stdin>", line 1, in <module>
>>   File "/usr/lib/pymodules/python2.6/pylab.py", line 1, in <module>
>>     from matplotlib.pylab import *
>>   File "/usr/lib/pymodules/python2.6/matplotlib/pylab.py", line 206, in
>> <module>
>>     from matplotlib import mpl  # pulls in most modules
>>   File "/usr/lib/pymodules/python2.6/matplotlib/mpl.py", line 3, in
>> <module>
>>     from matplotlib import axes
>>   File "/usr/lib/pymodules/python2.6/matplotlib/axes.py", line 17, in
>> <module>
>>     import matplotlib.dates as mdates
>>   File "/usr/lib/pymodules/python2.6/matplotlib/dates.py", line 87, in
>> <module>
>>     import pytz
>>   File "/usr/lib/python2.6/dist-packages/pytz/__init__.py", line 32, in
>> <module>
>>     from pkg_resources import resource_stream
>> ValueError: bad marshal data
>> [/code]
>>
>>
>> I am using python-matplotlib on Debian Squeeze.
>>
>> Any help would be appreciated !
>> Thanks,
>>
>>
> Oz,
>
> I see you have filed a bug with the Debian people and have gotten a
> response.  Did their suggestion help you?
>
> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=591606
>
> Ben Root
>



-- 
Oz Nahum
Graduate Student
Zentrum für Angewandte Geologie
Universität Tübingen

---

Imagine there's no countries
it isn't hard to do
Nothing to kill or die for
And no religion too
Imagine all the people
Living life in peace

On Wed, Aug 4, 2010 at 3:52 PM, Jorge Scandaliaris
<jor...@ya...>wrote:

> Benjamin Root <ben.root@...> writes:
>
> <snip>
> >
> > Yes, please do, and then mention which bug report you filed to this
> thread.
> Ben Root
>
> Done. Bug number is 3039678
>
> Could this behavior be due to the fact that scatter() accepts sequences for
> setting the color of each point individually? Maybe I am way off, but who
> knows...
>
> Jorge
>
>
>
Probably not directly, but I hadn't thought about that before.  For a set of
scatter points that are colored by values, what should the legend show?  In
other words, what does it *mean* for there to be a legend for points that
are colored in a potentially non-uniform manner?

So, maybe this is desired behavior (but possibly by accident)?

Thanks for your help,
Ben Root

Benjamin Root <ben.root@...> writes:

<snip>
> 
> Yes, please do, and then mention which bug report you filed to this thread.
Ben Root

Done. Bug number is 3039678 

Could this behavior be due to the fact that scatter() accepts sequences for
setting the color of each point individually? Maybe I am way off, but who 
knows...

Jorge

On Wed, Aug 4, 2010 at 3:01 PM, Jorge Scandaliaris
<jor...@ya...>wrote:

> Benjamin Root <ben.root@...> writes:
>
> <snip>
> > Hmm, this definitely looks like a bug.
>
> Should I fill a bug report about this?
>
> Jorge
>
>
Yes, please do, and then mention which bug report you filed to this thread.

Ben Root

Benjamin Root <ben.root@...> writes:

<snip>
> Hmm, this definitely looks like a bug. 

Should I fill a bug report about this?

Jorge

Benjamin Root <ben.root@...> writes:

<snip>
> Hmm, this definitely looks like a bug.  If I explicitly state what color I want
> using the "color" keyword instead of using 'c' and 'norm', then everything
> works properly.

I can't tell if this is a bug or not, but the trick of using a color directly is
an even simpler workaround than the one I used. Thanks.

Jorge

On 08/04/2010 06:19 AM, John Hunter wrote:
> On Wed, Aug 4, 2010 at 11:07 AM, Tommy Grav<tg...@ma...>  wrote:
>> A rather simple question, but I could not find the
>> answer while rummaging around on the matplotlib
>> webpages. Is there a way to increase the size of
>> the tick label sizes from say fontsize 9 to 12?
>
> for label in ax.get_xticklabels() + ax.get_yticklabels():
>      labe.set_fontsize(12)
>
> You may want to take a look at the artist tutorial, which gives an
> overview of all the objects in the matplotlib figure, how to get at
> them and modify them, etc.
>
> http://matplotlib.sourceforge.net/users/artists.html
>
> JDH

Also, if you are running mpl 1.0, you can use the tick_params Axes 
method or pyplot function. E.g.,

ax.tick_params(labelsize=12)

Eric

So I switched to axes_grid1 and got things working.  Thanks.
-- 

R. Padraic Springuel
Research Assistant
Department of Physics and Astronomy
University of Maine
Bennett 309
Office Hours: By Appointment Only

>> Is there a cleaner way to do this?
>
> Use a LineCollection:
>
> http://matplotlib.sourceforge.net/search.html?q=codex+linecollection
>
> JDH

Exactly what I was looking for. Thank you.

/Ulf Larsson
 		 	   		  

On Wed, Aug 4, 2010 at 11:07 AM, Tommy Grav <tg...@ma...> wrote:
> A rather simple question, but I could not find the
> answer while rummaging around on the matplotlib
> webpages. Is there a way to increase the size of
> the tick label sizes from say fontsize 9 to 12?

for label in ax.get_xticklabels() + ax.get_yticklabels():
    labe.set_fontsize(12)

You may want to take a look at the artist tutorial, which gives an
overview of all the objects in the matplotlib figure, how to get at
them and modify them, etc.

http://matplotlib.sourceforge.net/users/artists.html

JDH

A rather simple question, but I could not find the
answer while rummaging around on the matplotlib
webpages. Is there a way to increase the size of
the tick label sizes from say fontsize 9 to 12?

Tommy

On Wed, Aug 4, 2010 at 10:00 AM, Jorge Scandaliaris
<jor...@ya...>wrote:

> Jorge Scandaliaris <jorgesmbox-ml@...> writes:
> >
> > Hi,
> > I am adding several scatter plots to the same axis, each having a
> specific
> > color. When I call legend on the axis it correctly picks all scatter
> plots with
> > their symbols and labels, but it doesn't pick up the color. The example
> below
> > demonstrates this. What I would like to do is to have in the legend the
> marker
> > its color matching that of the set it represents. How would I do this?
>
> Answering my own question, I found a way to achieve what I wanted by using
> a
> proxy artist as suggested in this thread:
>
> http://thread.gmane.org/gmane.comp.python.matplotlib.general/20995/focus=20999
>
> Maybe there's a better approach. I certainly would love legend to pick up
> on the
> color used in the scatter plot by its own.
>
> Jorge
>
>
>
Hmm, this definitely looks like a bug.  If I explicitly state what color I
want using the "color" keyword instead of using 'c' and 'norm', then
everything works properly.

Digging through the code, 'color' gets used right after the creation of the
collection object by calling the .update() function, which in turn calls the
.set_color() function which calls the set_facecolor() and set_edgecolor()
functions.  However, when using 'c' and 'norm', and having 'c' contain an
array that is the same length as the data, causes colors to be set to None
during the creation of the polygon collection.  After the collection
creation, set_array(), set_cmap() and set_norm() are called, which allows
for the scatter graph to be colored correctly upon draw.

However, I suspect that when the legend is being created, the legend uses
whatever the face/edge color that was set for the collection, which probably
isn't set yet and won't be until a draw is performed.

Ben Root

On Wed, Aug 4, 2010 at 9:42 AM, Ulf Larsson <ulf...@ho...> wrote:
>
>
> Hi,
>
> I have some performance problems when plotting several lines and would
> appreciate some comments. My application plots lots of lines (~5000)
> of different sizes. The performance bottleneck lies in the following
> code snippet:
>
> for s in data.layout.segment:
>    x = []
>    y = []
>    for p in s.part:
>        for px, py in p.curve_points():
>            x.append(px)
>            y.append(py)
>    axes.plot(x, y, 'g', label = '_nolegend_')
>
> Profiling showed that half of the time was spent in parsing the plot
> arguments and most of the other half was spent in
> Axes._set_artist_props.
>
> I could speed up the application by using Line2D and
> Axes.add_lines. But the only way to come around the time spent in
> Axes._set_artist_props that I could come up with is this ugly hack
> where I only call Axes.add_line for the first line and after that use
> copies that are added directly to Axes.lines.
>
> org_line = None
> for s in data.layout.segment:
>    x = []
>    y = []
>    for p in s.part:
>        for px, py in p.curve_points():
>            x.append(px)
>            y.append(py)
>    if not org_line:
>        org_line = matplotlib.lines.Line2D(numpy.array(x), numpy.array(y),
>                                           color='green', label = '_nolegend_')
>        axis.add_line(org_line)
>    else:
>        line = copy.copy(org_line)
>        line.set_xdata(numpy.array(x))
>        line.set_ydata(numpy.array(y))
>        axis.lines.append(line)
>
> Is there a cleaner way to do this?

Use a LineCollection:

http://matplotlib.sourceforge.net/search.html?q=codex+linecollection

JDH

On Wed, Aug 4, 2010 at 12:26 AM, Oz Nahum <na...@gm...> wrote:

> Hi Everyone  !
> This is urgent, I have to finish some plots by tomorrow, and I totally lost
> the ability to work with python matplotlib -
> "import pylab"
> "from pylab import *"
>
> all yield the following error:
>
> [code]Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/usr/lib/pymodules/python2.6/pylab.py", line 1, in <module>
>     from matplotlib.pylab import *
>   File "/usr/lib/pymodules/python2.6/matplotlib/pylab.py", line 206, in
> <module>
>     from matplotlib import mpl  # pulls in most modules
>   File "/usr/lib/pymodules/python2.6/matplotlib/mpl.py", line 3, in
> <module>
>     from matplotlib import axes
>   File "/usr/lib/pymodules/python2.6/matplotlib/axes.py", line 17, in
> <module>
>     import matplotlib.dates as mdates
>   File "/usr/lib/pymodules/python2.6/matplotlib/dates.py", line 87, in
> <module>
>     import pytz
>   File "/usr/lib/python2.6/dist-packages/pytz/__init__.py", line 32, in
> <module>
>     from pkg_resources import resource_stream
> ValueError: bad marshal data
> [/code]
>
>
> I am using python-matplotlib on Debian Squeeze.
>
> Any help would be appreciated !
> Thanks,
>
>
Oz,

I see you have filed a bug with the Debian people and have gotten a
response.  Did their suggestion help you?

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=591606

Ben Root

Jorge Scandaliaris <jorgesmbox-ml@...> writes:
> 
> Hi,
> I am adding several scatter plots to the same axis, each having a specific
> color. When I call legend on the axis it correctly picks all scatter plots with
> their symbols and labels, but it doesn't pick up the color. The example below
> demonstrates this. What I would like to do is to have in the legend the marker
> its color matching that of the set it represents. How would I do this? 

Answering my own question, I found a way to achieve what I wanted by using a
proxy artist as suggested in this thread:
http://thread.gmane.org/gmane.comp.python.matplotlib.general/20995/focus=20999

Maybe there's a better approach. I certainly would love legend to pick up on the
color used in the scatter plot by its own.

Jorge

On Tue, Aug 3, 2010 at 11:12 PM, Michael Hannon <jm_...@ya...> wrote:

> Greetings.  I'm unable to get mathtext to work properly on my linux system:
>
>    # cat /etc/redhat-release
>    Fedora release 13 (Goddard)
>
>    # uname -a
>    Linux localhost.localdomain 2.6.33.6-147.2.4.fc13.x86_64 #1 SMP Fri Jul
> 23
>    17:14:44 UTC 2010 x86_64 x86_64 x86_64 GNU/Linux
>
> The problem is essentially identical to one that is described in the thread
> at:
>
>
> http://www.mail-archive.com/mat...@li.../msg09208.html
>
> I didn't see a real resolution to the problem in that thread, at least not
> one
> that applied to me.
>
> In brief, when I specify a math symbol in a text string, say:
>
>    r'$\pi$'
>
> I get some "random" character (capital A in this case) at the place where
> I'm
> supposed to see the Greek letter Pi.
>
> Furthermore, if I try to coerce the use of the Greek letters by setting
> "text.markup" to "tex", either interactively or in the rc file, I get an
> error
> saying that it's not a valid parameter:
>
>    Bad key "text.markup" on line 161 in ... matplotlibrc
>
> for instance.  This is followed by the message:
>
>    You probably need to get an updated matplotlibrc file from
>    http://matplotlib.sf.net/_static/matplotlibrc or from the matplotlib
>    source distribution
>
> But the file at that link still contains:
>
>    #text.markup:'plain'  # Affects how text, such as titles and labels, are
>                          # interpreted by default.
>                          # 'plain': As plain, unformatted text
>                          # 'tex': As TeX-like text.  Text between $'s
>                          # will be
>                          # formatted as a TeX math expression.
>                          # This setting has no effect when text.usetex
>                          # is True.
>                          # In that case, all text will be sent to TeX
>                          # for
>                          # processing.
>
> I don't know what to make of this.  I did the following to try to pin down
> the
> parameters:
>
>    import matplotlib as mpl
>    import pprint
>    x = mpl.rcParams.keys()
>    x.sort()
>    pprint.pprint(x)
>
> This produced:
>
>    .
>    .
>    .
>    'svg.image_noscale',
>    'text.color',
>    'text.dvipnghack',
>    'text.fontangle',
>    'text.fontsize',
>    'text.fontstyle',
>    'text.fontvariant',
>    'text.fontweight',
>    'text.latex.preamble',
>    'text.latex.preview',
>    'text.latex.unicode',
>    'text.usetex',
>    'timezone',
>    .
>    .
>    .
>
> Indeed, there does not seem to be a "text.markup" in this list.
>
> I have tried this both with and without an rc file, and I've tried it after
> deleting completely my .matplotlib directory.  I get the same results in
> all
> cases.
>
> OTOH, if I set:
>
>    text.usetex
>
> to "True", I do get the expected mathematical symbols, albeit after a
> noticeable
> delay.
>
> I've got the following two matplotlib packages installed:
>
>    python-matplotlib-0.99.1.2-4.fc13.x86_64
>    python-matplotlib-tk-0.99.1.2-4.fc13.x86_64
>
> and the system is running Python 2.6.4.
>
> Any suggestions?  Maybe I'm missing a package?
>
> Thanks.
>
> -- Mike
>
>
>
Mike,

Wow, I hadn't realized that Fedora 13 is carrying a rather old version of
matplotlib (~2 years old).  I should double-check what version is in rawhide
and possibly prod some people to get them to choose a more recent version...

I am not exactly sure what the cause of your problem is, but you could try
uninstalling the Fedora packages of matplotlib and install the latest
version from source.  Unless someone else has a better idea about what the
cause of the problem is?

Ben Root

Hi,

I have some performance problems when plotting several lines and would
appreciate some comments. My application plots lots of lines (~5000)
of different sizes. The performance bottleneck lies in the following
code snippet:

for s in data.layout.segment:
    x = []
    y = []
    for p in s.part:
        for px, py in p.curve_points():
            x.append(px)
            y.append(py)
    axes.plot(x, y, 'g', label = '_nolegend_')

Profiling showed that half of the time was spent in parsing the plot
arguments and most of the other half was spent in
Axes._set_artist_props.

I could speed up the application by using Line2D and
Axes.add_lines. But the only way to come around the time spent in
Axes._set_artist_props that I could come up with is this ugly hack
where I only call Axes.add_line for the first line and after that use
copies that are added directly to Axes.lines.

org_line = None
for s in data.layout.segment:
    x = []
    y = []
    for p in s.part:
        for px, py in p.curve_points():
            x.append(px)
            y.append(py)
    if not org_line:
        org_line = matplotlib.lines.Line2D(numpy.array(x), numpy.array(y),
                                           color='green', label = '_nolegend_')
        axis.add_line(org_line)
    else:
        line = copy.copy(org_line)
        line.set_xdata(numpy.array(x))
        line.set_ydata(numpy.array(y))
        axis.lines.append(line)

Is there a cleaner way to do this?

Also, my feelings is that matplotlib 1.0 is slower with my original
code than previous version. But I have no numbers to back it up with.

regards
Ulf Larsson
 		 	   		  

On Wed, Aug 4, 2010 at 6:27 AM, thowa <tho...@fo...> wrote:
> I'm afraid, I made myself not clear enough.
> I want to rotate the numbers on the axis (similar to what autofmt_xdate() is
> doing).
> As I understand, autofmt_xdate() is changing the rotation of the numbers for
> all sub-plots.
> But I want to do it only for selected subplots.
>
> Is that is possible?

Yes, if ax is your subplot instance

for label in ax.get_xticklabels():
    label.set_rotation(45)
    label.set_horizontalalignment('right')

JDH

thowa wrote:
> 
> 
> 
> I'm afraid, I made myself not clear enough.
> I want to rotate the numbers on the axis (similar to what autofmt_xdate()
> is doing).
> As I understand, autofmt_xdate() is changing the rotation of the numbers
> for all sub-plots.
> But I want to do it only for selected subplots.
> 
> Is that is possible?
> 

I have realized, that it is possible.
e.g. by using 

    plot.set_xticklabels(bins, rotation=45)



Thanks a lot!


-- 
View this message in context: http://old.nabble.com/Rotated-text-for-selected-subplots-tp29334275p29344561.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Hi,
I am adding several scatter plots to the same axis, each having a specific
color. When I call legend on the axis it correctly picks all scatter plots with
their symbols and labels, but it doesn't pick up the color. The example below
demonstrates this. What I would like to do is to have in the legend the marker
its color matching that of the set it represents. How would I do this? 

Thanks,

Jorge

PS: I am using matplotlib 1.0

-----------
import numpy as np
import matplotlib as mpl
import matplotlib.pyplot as plt

data0 = np.random.rand(10,2)
data1 = np.random.rand(10,2)
data2 = np.random.rand(10,2)
data = [data0, data1, data2]

fig, ax = plt.subplots(1,1)
norm = mpl.colors.Normalize(0,len(data))

for i,d in enumerate(data):
    ax.scatter(d.T[0], d.T[1], label='data set ' + str(i), c=np.ones(d.shape[0])*i,
            norm=norm)
ax.legend()
plt.show()
-----------

John Hunter-4 wrote:
> 
> On Tue, Aug 3, 2010 at 6:23 AM, thowa <tho...@fo...> wrote:
>>
>> Hi,
>>
>> I'm pretty new to Matplotlib and I'm really impressed about the
>> possibilities !!!
>> GREAT WORK !!!
>>
>> I have a figure with 3 subplots like this.
>>
>> *-------*    *------------------------------*
>> |          |     |                                          |
>> |    A    |     |                                          |
>> |          |     |                                          |
>> *-------*    |                                          |
>>                 |                     C                     |
>> *-------*    |                                          |
>> |          |     |                                          |
>> |    B    |     |                                          |
>> |          |     |                                          |
>> *-------*    *------------------------------*
>>
>> I want to have the text on the x-axes rotated, but only for subplot A and
>> B
>> The text for subplot C should remain unrotated.
> 
> All of the text commands "text", "xlabel", "ylabel", "title" take a
> rotation keyword argument, so you can pass that in and set the angle
> you want.  With an existing text instance, you can call the
> set_rotation method.
> 
> 

I'm afraid, I made myself not clear enough.
I want to rotate the numbers on the axis (similar to what autofmt_xdate() is
doing).
As I understand, autofmt_xdate() is changing the rotation of the numbers for
all sub-plots.
But I want to do it only for selected subplots.

Is that is possible?
-- 
View this message in context: http://old.nabble.com/Rotated-text-for-selected-subplots-tp29334275p29343081.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Hi Everyone  !
This is urgent, I have to finish some plots by tomorrow, and I totally lost
the ability to work with python matplotlib -
"import pylab"
"from pylab import *"

all yield the following error:

[code]Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/pymodules/python2.6/pylab.py", line 1, in <module>
    from matplotlib.pylab import *
  File "/usr/lib/pymodules/python2.6/matplotlib/pylab.py", line 206, in
<module>
    from matplotlib import mpl  # pulls in most modules
  File "/usr/lib/pymodules/python2.6/matplotlib/mpl.py", line 3, in <module>
    from matplotlib import axes
  File "/usr/lib/pymodules/python2.6/matplotlib/axes.py", line 17, in
<module>
    import matplotlib.dates as mdates
  File "/usr/lib/pymodules/python2.6/matplotlib/dates.py", line 87, in
<module>
    import pytz
  File "/usr/lib/python2.6/dist-packages/pytz/__init__.py", line 32, in
<module>
    from pkg_resources import resource_stream
ValueError: bad marshal data
[/code]


I am using python-matplotlib on Debian Squeeze.

Any help would be appreciated !
Thanks,



-- 
Oz Nahum
Graduate Student
Zentrum für Angewandte Geologie
Universität Tübingen

---

Imagine there's no countries
it isn't hard to do
Nothing to kill or die for
And no religion too
Imagine all the people
Living life in peace

Greetings.  I'm unable to get mathtext to work properly on my linux system:

    # cat /etc/redhat-release 
    Fedora release 13 (Goddard)

    # uname -a
    Linux localhost.localdomain 2.6.33.6-147.2.4.fc13.x86_64 #1 SMP Fri Jul 23
    17:14:44 UTC 2010 x86_64 x86_64 x86_64 GNU/Linux

The problem is essentially identical to one that is described in the thread
at:

http://www.mail-archive.com/mat...@li.../msg09208.html

I didn't see a real resolution to the problem in that thread, at least not one
that applied to me.

In brief, when I specify a math symbol in a text string, say:

    r'$\pi$'

I get some "random" character (capital A in this case) at the place where I'm
supposed to see the Greek letter Pi.

Furthermore, if I try to coerce the use of the Greek letters by setting
"text.markup" to "tex", either interactively or in the rc file, I get an error
saying that it's not a valid parameter:

    Bad key "text.markup" on line 161 in ... matplotlibrc

for instance.  This is followed by the message:

    You probably need to get an updated matplotlibrc file from
    http://matplotlib.sf.net/_static/matplotlibrc or from the matplotlib
    source distribution

But the file at that link still contains:

    #text.markup:'plain'  # Affects how text, such as titles and labels, are
                          # interpreted by default.
                          # 'plain': As plain, unformatted text
                          # 'tex': As TeX-like text.  Text between $'s
                          # will be
                          # formatted as a TeX math expression.
                          # This setting has no effect when text.usetex
                          # is True.
                          # In that case, all text will be sent to TeX
                          # for
                          # processing.

I don't know what to make of this.  I did the following to try to pin down the
parameters:

    import matplotlib as mpl
    import pprint
    x = mpl.rcParams.keys()
    x.sort()
    pprint.pprint(x)

This produced:

    .
    .
    .
    'svg.image_noscale',
    'text.color',
    'text.dvipnghack',
    'text.fontangle',
    'text.fontsize',
    'text.fontstyle',
    'text.fontvariant',
    'text.fontweight',
    'text.latex.preamble',
    'text.latex.preview',
    'text.latex.unicode',
    'text.usetex',
    'timezone',
    .
    .
    .

Indeed, there does not seem to be a "text.markup" in this list.

I have tried this both with and without an rc file, and I've tried it after
deleting completely my .matplotlib directory.  I get the same results in all
cases.

OTOH, if I set:

    text.usetex

to "True", I do get the expected mathematical symbols, albeit after a noticeable
delay.

I've got the following two matplotlib packages installed:

    python-matplotlib-0.99.1.2-4.fc13.x86_64
    python-matplotlib-tk-0.99.1.2-4.fc13.x86_64

and the system is running Python 2.6.4.

Any suggestions?  Maybe I'm missing a package?

Thanks.

-- Mike