matplotlib-users Mailing List for matplotlib

On Tue, Mar 16, 2010 at 4:37 PM, Josh Hemann <jh...@vn...> wrote:

>
> I have an issue with showing more than 81 tick marks on an X axis and I am
> trying to determine a way around it. Background... I am plotting vectors in
> which each element represents a different variable and I really do want to
> see the labels associated with each element. The vectors may be only 8
> elements long, or as much as 110. When there are more than say 40 elements,
> I usually split the plot into two plots contained in a single figure window
> (e.g., plotting elements 0:30 in fig.add_subplot(211) and 30:60 in
> fig.add_subplot(212)).
>
> Here are a couple of examples...
>
> Only 41 variables:
> http://old.nabble.com/file/p27924845/Factor_2_TrainingProfiles.png
>
>
> 71 variables:
> http://old.nabble.com/file/p27924845/Factor_2_TrainingProfiles.jpeg
>
>
> I have a vector with a 105 elements and before I split things into three
> plots I wanted to see what cramming 53 or so variables into a single set of
> axes would look like. But, my code that works for these cases does not show
> enough tickmarks for the 105 element data.
>
> Here is an example that you can copy and paste to see for yourself.
>
> import matplotlib.pyplot as plt
> from matplotlib.ticker import MaxNLocator
> fig = plt.figure(figsize=[12,7])
> ax = fig.add_subplot(111)
> ax.plot(range(110))
> fig.canvas.draw()
> ints = range(1,111)
> ints = [str(num) for num in ints]
> ax.xaxis.set_major_locator(MaxNLocator(110))
> xtickNames = plt.setp(ax, xticklabels=ints)
> plt.setp(xtickNames, rotation=90, fontsize=7);
>
> If you play with the argument to MaxNLocator, you'll see how for smaller
> values (like 40) things work as expected (or at least how I have shown the
> code has worked for the smaller data sets).
>
> I have been poking around trying to see what options I have and have not
> found anything to get past this limit. Before I start diving into source
> code, can anyone suggest
>
> -Is there a limit?
> -Is there an obvious way to accomplish what I need?
>
> Ultimately, I may split large vectors like this into more than two plots
> but
> hitting that limit has made me want to investigate why.
>
> Thanks!
>

Oh these busy chemical compound plots :) Are those results of gas
chromatography analysis?

Something like below produces a nice fully plotted output here. Could you
give it a try?

import matplotlib.pyplot as plt
plt.plot(range(100))
locs, labels = plt.xticks(range(100), range(100))
plt.setp(labels, rotation=90, fontsize=7)
plt.show()


-- 
Gökhan

I have an issue with showing more than 81 tick marks on an X axis and I am
trying to determine a way around it. Background... I am plotting vectors in
which each element represents a different variable and I really do want to
see the labels associated with each element. The vectors may be only 8
elements long, or as much as 110. When there are more than say 40 elements,
I usually split the plot into two plots contained in a single figure window
(e.g., plotting elements 0:30 in fig.add_subplot(211) and 30:60 in
fig.add_subplot(212)).

Here are a couple of examples...

Only 41 variables:
http://old.nabble.com/file/p27924845/Factor_2_TrainingProfiles.png 


71 variables:
http://old.nabble.com/file/p27924845/Factor_2_TrainingProfiles.jpeg 


I have a vector with a 105 elements and before I split things into three
plots I wanted to see what cramming 53 or so variables into a single set of
axes would look like. But, my code that works for these cases does not show
enough tickmarks for the 105 element data.

Here is an example that you can copy and paste to see for yourself.

import matplotlib.pyplot as plt
from matplotlib.ticker import MaxNLocator
fig = plt.figure(figsize=[12,7])
ax = fig.add_subplot(111)
ax.plot(range(110))
fig.canvas.draw()
ints = range(1,111)
ints = [str(num) for num in ints]
ax.xaxis.set_major_locator(MaxNLocator(110))
xtickNames = plt.setp(ax, xticklabels=ints)
plt.setp(xtickNames, rotation=90, fontsize=7);

If you play with the argument to MaxNLocator, you'll see how for smaller
values (like 40) things work as expected (or at least how I have shown the
code has worked for the smaller data sets).

I have been poking around trying to see what options I have and have not
found anything to get past this limit. Before I start diving into source
code, can anyone suggest 

-Is there a limit?
-Is there an obvious way to accomplish what I need? 

Ultimately, I may split large vectors like this into more than two plots but
hitting that limit has made me want to investigate why.

Thanks!


 


-----
Josh Hemann
Statistical Advisor 
http://www.vni.com/ Visual Numerics 
jh...@vn... | P 720.407.4214 | F 720.407.4199 

-- 
View this message in context: http://old.nabble.com/Is-there-a-maximum-number-of-x-tickmarks--tp27924845p27924845.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

For the Windows machine, if you installed with the superpack, you
should find an deinstaller in the Python directory called
"Removematplotlib.exe", I *guess* it only removes in fact the
matplotlib package.  You can also safely rename (or delete) the old
matplotlib directory and the mpl_toolkits directory in
Python-dir\Lib\site-packages\.

Ah I see, you mean how to deinstall on Ubuntu?  I would recommend to
rename the old folders in your site-package directory to, say,
matplotlib_ and mpl_toolkits_, and when you are shure later you could
be able to safely remove.  Does Ubuntu have a package mgr?

Hope I could help,
Friedrich

2010/3/16 Samuel Teixeira Santos <arc...@gm...>:
> Hi all...
> On this afternoon I installed on my ubuntu 8.04 server the python-matplotlib
> package and his dependencies.
> but If I pay attention right that package installed matplotlib 0.91
> on my local desktop wich run windows, I installed the last version...
> Anyone knows how do I de-install only matplotlib pack, let only the
> dependancies
> and install the most recent version?

Dear Michael,

thanks for your input. So far, though, no luck.

I have deleted "SimHei" in matplotlibrc, and I can then continue 
generating CJK characters. The png is generated, the eps is not. Thus, 
no change. The error output is also the same:

RuntimeError: Face has no glyph names
WARNING: Failure executing file: <dea.py>

I have tried

  import matplotlib
  matplotlib.use('GtkCairo')

as you suggested, but they had no effect whatsoever. Even the error 
output is the same.

I attach my code, maybe that gives a hint.

Note: in line 327 and 328 of the matplotlibrc I have added

ps.fonttype=42
pdf.fonttype=42

whereas I have uncommented

pdf.fonttype	: 3

Any ideas? I would most welcome any hint and suggestion!

Many thanks,

David




On 17/03/10 04:15, Michael Droettboom wrote:
> The font you are using (SimHei) does not have any glyph names -- these
> are used in the Postscript backend to refer to glyphs outside of the
> ASCII range. More specifically, it looks like it has at least one
> invalid glyph name (glyph names can only contain ASCII characters) --
> loading it in FontForge complains about this. I haven't come across such
> a font before, but maybe that's common in CJK fonts. I don't know. I'm
> looking through the spec to find a way that glyphs can be referenced
> without a name, but I'm not finding one. Note, the PDF backend has the
> same issue.
>
> Do you have the same problem if you remove "SimHei" from the
> font.sans-serif list and thus use "Adobe Song Std" instead? (I was able
> to find an online download of SimHei to test with, but not Adobe Song Std).
>
> As a workaround, the Cairo backend seems to handle this font just fine.
> You can add
>
> import matplotlib
> matplotlib.use('GtkCairo')
>
> to the top of your script.
>
> Mike
>
> David wrote:
>> Hello everybody,
>>
>> I have a final problem with my graph. As a last step I produce an
>> *.eps file that I use in conjunction with LaTeX.
>>
>> Here is the last part of my code:
>>
>> # plt.title('Title')
>> xlab = plt.xlabel(u'输入 1')
>> #xlab.set_position((0.2, 0.1))
>> ylab = plt.ylabel(u'输入 2')
>> plt.grid(True)
>> plt.subplots_adjust(bottom=0.2)
>> plt.show()
>> plt.savefig('dea.eps')
>>
>>
>> plt.show() produces the correct output,
>>
>> but
>>
>> plt.savefig('dea.eps') produces an error (the error message is attached).
>>
>> The error is clearly linked to the Chinese, as it runs through if I
>> take the Chinese out of the code.
>> Also, plt.savefig('dea.png') works fine.
>>
>> Could anyone indicate where I would have to look for the mistake? The
>> matplotlibrc should be fine, but I am not sure.
>>
>> Your help would be greatly appreciated!
>>
>> Many thanks,
>>
>> David
>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------
>>
>> ------------------------------------------------------------------------------
>>
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

The font you are using (SimHei) does not have any glyph names -- these 
are used in the Postscript backend to refer to glyphs outside of the 
ASCII range.  More specifically, it looks like it has at least one 
invalid glyph name (glyph names can only contain ASCII characters) -- 
loading it in FontForge complains about this. I haven't come across such 
a font before, but maybe that's common in CJK fonts.  I don't know.  I'm 
looking through the spec to find a way that glyphs can be referenced 
without a name, but I'm not finding one.  Note, the PDF backend has the 
same issue.

Do you have the same problem if you remove "SimHei" from the 
font.sans-serif list and thus use "Adobe Song Std" instead?  (I was able 
to find an online download of SimHei to test with, but not Adobe Song Std).

As a workaround, the Cairo backend seems to handle this font just fine.  
You can add

  import matplotlib
  matplotlib.use('GtkCairo')

to the top of your script.

Mike

David wrote:
> Hello everybody,
>
> I have a final problem with my graph. As a last step I produce an 
> *.eps file that I use in conjunction with LaTeX.
>
> Here is the last part of my code:
>
> # plt.title('Title')
> xlab = plt.xlabel(u'输入 1')
> #xlab.set_position((0.2, 0.1))
> ylab = plt.ylabel(u'输入 2')
> plt.grid(True)
> plt.subplots_adjust(bottom=0.2)
> plt.show()
> plt.savefig('dea.eps')
>
>
> plt.show() produces the correct output,
>
> but
>
> plt.savefig('dea.eps') produces an error (the error message is attached).
>
> The error is clearly linked to the Chinese, as it runs through if I 
> take the Chinese out of the code.
> Also, plt.savefig('dea.png') works fine.
>
> Could anyone indicate where I would have to look for the mistake? The 
> matplotlibrc should be fine, but I am not sure.
>
> Your help would be greatly appreciated!
>
> Many thanks,
>
> David
>
>
>
>
>
> ------------------------------------------------------------------------
>
> ------------------------------------------------------------------------------
> Download Intel® Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> ------------------------------------------------------------------------
>
> _______________________________________________
> 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

Hello everybody,

I have a final problem with my graph. As a last step I produce an *.eps 
file that I use in conjunction with LaTeX.

Here is the last part of my code:

# plt.title('Title')
xlab = plt.xlabel(u'输入 1')
#xlab.set_position((0.2, 0.1))
ylab = plt.ylabel(u'输入 2')
plt.grid(True)
plt.subplots_adjust(bottom=0.2)
plt.show()
plt.savefig('dea.eps')


plt.show() produces the correct output,

but

plt.savefig('dea.eps') produces an error (the error message is attached).

The error is clearly linked to the Chinese, as it runs through if I take 
the Chinese out of the code.
Also, plt.savefig('dea.png') works fine.

Could anyone indicate where I would have to look for the mistake? The 
matplotlibrc should be fine, but I am not sure.

Your help would be greatly appreciated!

Many thanks,

David

Thanks a lot, that seems to work!

Alex


On Tue, Mar 16, 2010 at 18:54, Ben Axelrod <BAx...@co...> wrote:
> There is an uncommented, and therefore undocumented function:
>
> axes3d.view_init(elev, azim)
>
> that you can use to rotate the axes.  If you have not already, I suggest you use the current SVN version of MPL instead of the 0.99.1 version.  Mplot3d has some more features in the trunk, but it is still rough around the edges.
>
> -Ben
>
>
>
> -----Original Message-----
> From: Alexander Dietz [mailto:ale...@go...]
> Sent: Tuesday, March 16, 2010 1:25 PM
> To: Mat...@li...
> Subject: [Matplotlib-users] How to 'rotate' a 3D plot?
>
> Hi,
>
> I have successfully created a 3D scatter plot with mplot3d, but how can I rotate the plot around e.g. the z-axis?
>
> I do not want to use the user interface but I would like to use a command to do that. But I could not find good documentation anywhere and the commands attributed to the Axes3D also do not show anything obvious.
>
> Thanks
>  Alex
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

There is an uncommented, and therefore undocumented function:

axes3d.view_init(elev, azim)

that you can use to rotate the axes.  If you have not already, I suggest you use the current SVN version of MPL instead of the 0.99.1 version.  Mplot3d has some more features in the trunk, but it is still rough around the edges.
 
-Ben



-----Original Message-----
From: Alexander Dietz [mailto:ale...@go...] 
Sent: Tuesday, March 16, 2010 1:25 PM
To: Mat...@li...
Subject: [Matplotlib-users] How to 'rotate' a 3D plot?

Hi,

I have successfully created a 3D scatter plot with mplot3d, but how can I rotate the plot around e.g. the z-axis?

I do not want to use the user interface but I would like to use a command to do that. But I could not find good documentation anywhere and the commands attributed to the Axes3D also do not show anything obvious.

Thanks
  Alex

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Matplotlib-users mailing list
Mat...@li...
https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Samuel Teixeira Santos wrote:
> I fix it.
> 
> It was a dumb error
> 
> I using '\' on windows
> and on ubuntu-linux I must use '/'...


note that '\' works in Windows for the most part. Or, better yet, use 
os.path.join() and friends.

-Chris

-- 
Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R            (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chr...@no...

I fix it.

It was a dumb error

I using '\' on windows
and on ubuntu-linux I must use '/'...

sorry...
and thanks

2010/3/16 Michael Droettboom <md...@st...>

> Can you please post the entire traceback?
>
> Mike
>
> Samuel Teixeira Santos wrote:
>
>> Hi folks
>>
>> I'm using ubuntu 8.04 lts and matplotlib 0.91
>>
>> I cannot upgrade in this moment.
>>
>> On my app (for web) I fix several errors (because I did her in 0.99)
>>
>> My last error (I think it is) is on savefig
>>
>> It tells me that cannot open file
>>
>> on log error, appears on write_png method.
>>
>> Is this permission on directory? or a bug?
>>
>>
>> thanks in advanced
>> ------------------------------------------------------------------------
>>
>>
>> ------------------------------------------------------------------------------
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> 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
>
>

Hi,

I have successfully created a 3D scatter plot with mplot3d, but how
can I rotate the plot around e.g. the z-axis?

I do not want to use the user interface but I would like to use a
command to do that. But I could not find good documentation anywhere
and the commands attributed to the Axes3D also do not show anything
obvious.

Thanks
  Alex

On Tue, Mar 16, 2010 at 12:10 PM, PaterMaximus
<pat...@go...> wrote:
> Another way would be to use a font with a dark edge and light interior (or vice
> versa) but I know of know such font for matplotlib.
>

FYI, the svn version of matplotlib supports this.

http://matplotlib.sourceforge.net/trunk-docs/examples/pylab_examples/patheffect_demo.html

Regards,

-JJ

Can you please post the entire traceback?

Mike

Samuel Teixeira Santos wrote:
> Hi folks
>
> I'm using ubuntu 8.04 lts and matplotlib 0.91
>
> I cannot upgrade in this moment.
>
> On my app (for web) I fix several errors (because I did her in 0.99)
>
> My last error (I think it is) is on savefig
>
> It tells me that cannot open file
>
> on log error, appears on write_png method.
>
> Is this permission on directory? or a bug?
>
>
> thanks in advanced
> ------------------------------------------------------------------------
>
> ------------------------------------------------------------------------------
> Download Intel® Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> ------------------------------------------------------------------------
>
> _______________________________________________
> 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

On 3/16/10 10:10 AM, PaterMaximus wrote:
> I have I have an image with both light and dark regions. I want to write text on
> it in a color that contrasts with the underlying image color.
>
> Right now if I make the text black, it is not very legible if the underling
> color is dark. Similarly, if I make the text white, it is not legible if the
> underlying color is light. (using gray text will not work)
>
> One way to solve the problem would be to:
> -produce the image
> 	e.g. im1=plt.imshow (zM,origin='lower',interpolation='hanning',extent=None)
> -get the underlying color from the image where I want to place the text
> 	Need help here. How do I get the color from im1
> -calculate a contrasting color
> 	Need help here
> -plot the text in the contrasting color
>
>
>
> Another way would be to use a font with a dark edge and light interior (or vice
> versa) but I know of know such font for matplotlib.
>
> Any help appreciated...
>    

You could put your text inside a colored box - see e.g. 
http://matplotlib.sourceforge.net/users/text_intro.html 
<http://matplotlib.sourceforge.net/users/text_intro.html?highlight=text%20box>

-Jeff


-- 
Jeffrey S. Whitaker         Phone  : (303)497-6313
Meteorologist               FAX    : (303)497-6449
NOAA/OAR/PSD  R/PSD1        Email  : Jef...@no...
325 Broadway                Office : Skaggs Research Cntr 1D-113
Boulder, CO, USA 80303-3328 Web    : http://tinyurl.com/5telg

On 16/03/10 20:39, Jae-Joon Lee wrote:
 >
 > adjust your subplot parameters.
 >
 > plt.subplots_adjust(bottom=0.2)

Yep, that did the trick! thanks Jae-Joon and Gökhan!

David

I have I have an image with both light and dark regions. I want to write text on
it in a color that contrasts with the underlying image color.

Right now if I make the text black, it is not very legible if the underling
color is dark. Similarly, if I make the text white, it is not legible if the
underlying color is light. (using gray text will not work)

One way to solve the problem would be to: 
-produce the image
	e.g. im1=plt.imshow (zM,origin='lower',interpolation='hanning',extent=None)
-get the underlying color from the image where I want to place the text
	Need help here. How do I get the color from im1
-calculate a contrasting color
	Need help here
-plot the text in the contrasting color



Another way would be to use a font with a dark edge and light interior (or vice
versa) but I know of know such font for matplotlib.

Any help appreciated...

Hi folks

I'm using ubuntu 8.04 lts and matplotlib 0.91

I cannot upgrade in this moment.

On my app (for web) I fix several errors (because I did her in 0.99)

My last error (I think it is) is on savefig

It tells me that cannot open file

on log error, appears on write_png method.

Is this permission on directory? or a bug?


thanks in advanced

Hello,

Unfortunately, the mplot3D page on the SciPy cookbook is not up to date.
The 3D code in mplot3d has been moved to a toolkit and needs to be  
imported as is seen in the examples found here:
http://matplotlib.sourceforge.net/examples/mplot3d/index.html

> So: Any idea how to create a simple 3D scatter plot,seen from any  
> arbitrary angle?

You can rotate the axis in mplot3d :-)

I would say from your error log (the file matplotlib/axes3d.py exists  
whereas it does not in my mpl 0.99.0) that you did not uninstall the  
old matplotlib cleanly. I suggest that you clean the existing  
matplotlib completely and reinstall from the tarball).

Pierre

Le 16 mars 10 à 00:38, Alexander Dietz a écrit :

> Hi,
>
> I would like to create 3D plots. A search revealed the following page:
>
> http://www.scipy.org/Cookbook/Matplotlib/mplot3D
>
> which explains to install  matplotlib version 0.99 to use the 3D code.
> Unfortunately, after installingmatplotlib-0.99.1.1 (from
> matplotlib-0.99.1.2.tar.gz) I am not able to use the 3D code. The
> command that show that the correct version is used together with the
> error is shown below.
>
> Any idea how I can use this code? Do I have to downgrade to the 0.91.x
> maintenance branch? Or do you suggest to use mayavi
> (http://code.enthought.com/projects/mayavi/docs/development/html/ 
> mayavi/mlab.html)?
> I also tried to install the latter, after 1 hour of installing code
> and requirements and cmake and easyinstall.. I got the error that VTK
> is not installed. So: Any idea how to create a simple 3D scatter plot,
> seen from any arbitrary angle?
>
> Thanks
>   Alex
>
>
>
> In [1]: import matplotlib
>
> In [2]: matplotlib.__version__
> Out[2]: '0.99.1.1'
>
> In [3]: import matplotlib.axes3d as p3
> ---------------------------------------------------------------------- 
> -----
> NotImplementedError                       Traceback (most recent  
> call last)
>
> /home/alex/Documents/Job/Travel/2010-03_LSC_Pasadena/3D/<ipython
> console> in <module>()
>
> /usr/lib/python2.6/site-packages/matplotlib/axes3d.py in <module>()
> ----> 1 raise NotImplementedError('axes3d is not supported in
> matplotlib-0.98.  You may want to try the 0.91.x maintenance branch')
>
> NotImplementedError: axes3d is not supported in matplotlib-0.98.  You
> may want to try the 0.91.x maintenance branch
>
> ---------------------------------------------------------------------- 
> --------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On Tue, Mar 16, 2010 at 8:46 AM, Jon Olav Vik <jo...@gm...> wrote:

> Thank you, thank you, thank you.
>
> This is just as convenient, 50% faster even for 1000 series, and runtime does
> indeed scale as O(n) up to 10000 series. The projected speedup for 60000 series
> was 40x. However, in my actual use case it was at least 400x: Finishing in 2
> min 17 sec rather than not getting past halfway in 16 hours.
>
> (The extra difference is probably due to better memory usage. Still,
> LineCollection requires O(n) memory, whereas manually updating a bitmap would
> only use O(1) memory, where 1 = size of bitmap. However, I hope I never have to
> do that...)
>
> May the hours and hours you have saved me be added to your life! 8-)

Since you are granting extra life blessings, I thought I should add
something to the mix.  You should be able to achieve something close
to this using the animation blit API.  There is a little hackery at
the end to use the renderer to directly dump a PNG and thereby
circumvent the normal figure.canvas.draw pipeline, but the advantage
is you render directly to the canvas and save no intermediaries.  See
the examples and tutorial at

  http://matplotlib.sourceforge.net/examples/animation/index.html
  http://www.scipy.org/Cookbook/Matplotlib/Animations

Here's some example code::

import matplotlib._png as _png
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import numpy as np
fig = plt.figure()
ax = fig.add_subplot(111)

n = 10000
line, = ax.plot([],[], alpha=1)

x = np.arange(200)

fig.canvas.draw()
ax.axis([0, 200, -1, 1])
for i in range(n):
    if (i%100)==0: print i
    yy = np.sin(x / (2 * np.pi * x[-1] * i))

    line.set_data(x, yy)
    ax.draw_artist(line)


fig.canvas.blit(ax.bbox)

filename = 'test.png'
renderer = fig.canvas.get_renderer()
_png.write_png(renderer._renderer.buffer_rgba(0, 0),
               renderer.width, renderer.height,
               filename, fig.dpi)



JDH

Jae-Joon Lee <lee.j.joon@...> writes:

> If you're plotting lots of lines, do not use plot but use
> LineCollection instead.
> 
> http://matplotlib.sourceforge.net/examples/api/collections_demo.html
> 
> http://matplotlib.sourceforge.net/api/
collections_api.html#matplotlib.collections.LineCollection
> 
> Here is slightly modified version of your code that uses
> LineCollection (but I haven't check if the code is correct).
> With my not so good macbook, it took me 3 sec for 6000 lines and it
> seems like O(n) to me.

Thank you, thank you, thank you.

This is just as convenient, 50% faster even for 1000 series, and runtime does 
indeed scale as O(n) up to 10000 series. The projected speedup for 60000 series 
was 40x. However, in my actual use case it was at least 400x: Finishing in 2 
min 17 sec rather than not getting past halfway in 16 hours.

(The extra difference is probably due to better memory usage. Still, 
LineCollection requires O(n) memory, whereas manually updating a bitmap would 
only use O(1) memory, where 1 = size of bitmap. However, I hope I never have to 
do that...)

May the hours and hours you have saved me be added to your life! 8-)

Jon Olav

I have a script that calls several subroutines, each of which makes a
different figure. One of these routines makes lots of figures for use
in a webpage, all of which are saved as they are made. When I call
show() at the end of the script it is showing all the figures (as one
might expect), but what I really want is only some of the figures to
be brought up in the GUI. Is there a way of specifying which figures
show() shows (I can't find anything on the webpage)?

-- 
Cheers,

Nick Schurch

Data Analysis Group (The Barton Group),
School of Life Sciences,
University of Dundee,
Dow St,
Dundee,
DD1 5EH,
Scotland,
UK

Tel: +44 1382 388707
Fax: +44 1382 345 893

On Tue, Mar 16, 2010 at 3:30 AM, David <ld...@gm...> wrote:
> This did not yield any results. With the above code, xlab.set_position((0.2,
> 0.1)), I change the position of the xlabel.
> But the problem is that my graph is cut before the xlabel has a chance to
> appear (see dea.png). Basically the graph ends right after the x-axis. What
> I thus need is more whitespace under my x-axis. But how?

adjust your subplot parameters.

plt.subplots_adjust(bottom=0.2)

see

http://matplotlib.sourceforge.net/faq/howto_faq.html#move-the-edge-of-an-axes-to-make-room-for-tick-labels

http://matplotlib.sourceforge.net/faq/howto_faq.html#automatically-make-room-for-tick-labels

-JJ

On Tue, Mar 16, 2010 at 3:30 AM, David <ld...@gm...> wrote:
> On 16/03/10 07:59, Gökhan Sever wrote:
>>
>> Probably you need a unicode font-set that contain all the characters for
>> those alphabets. You can look at this example to see a simple unicode
>> demonstration example.
>
> Yes, I have done that, for example:
>
> ylab = plt.ylabel(u'输入 2')

if you use tex for rendering text, see

http://matplotlib.sourceforge.net/users/usetex.html#usetex-with-unicode

If not, the best option I know of is to set the font name directly, see

http://old.nabble.com/Russian-labels-without-LaTeX-td24538302.html

Matplotlib does not support fontset, so you can only use glyphs in the
font you specified.

Regards,

-JJ

If you're plotting lots of lines, do not use plot but use
LineCollection instead.

http://matplotlib.sourceforge.net/examples/api/collections_demo.html

http://matplotlib.sourceforge.net/api/collections_api.html#matplotlib.collections.LineCollection

Here is slightly modified version of your code that uses
LineCollection (but I haven't check if the code is correct).
With my not so good macbook, it took me 3 sec for 6000 lines and it
seems like O(n) to me.

Regards,

-JJ



    ax = subplot(111)
    x = np.arange(200)
    yy = [np.array((x, np.sin(x / (2 * np.pi * x[-1] * i))))) for i in range(n)]
    yyt = [np.transpose(y1) for y1 in yy]
    from matplotlib.collections import LineCollection

    lc = LineCollection(yyt, colors=[(0, 0, 0, alpha)])
    ax.add_collection(lc)
    ax.autoscale_view()


On Tue, Mar 16, 2010 at 7:26 AM, Jon Olav Vik <jo...@gm...> wrote:
> I want to overlay many line plots using alpha transparency. However, plotting
> them in Matplotlib takes about O(n**2) time, and I think I may be running into
> memory limitations as well.
>
> As a simple benchmark, I used IPython to run alco.ipy (below), which runs
> alco.py for an increasing number of data series. Extrapolating from this,
> plotting 60000 series would take something like 200 minutes. This is similar to
> my actual use case, which takes about 3 hours to finish a plot. Zooming in and
> saving again is much faster, taking only about 30 seconds.
>
> I would appreciate suggestions on how to speed this up. For instance:
>
> Is there a memoryless "canvas" object that I could draw on, just accumulating
> the alpha in each pixel: new_alpha = old_alpha + (1 - old_alpha) * this_alpha.
>
> Failing that, I could do it manually by keeping a Numpy array of the pixels in
> the image. For each series, find the x values corresponding to each column
> index, then interpolate to find the row index corresponding to each y value.
> Finally, use imshow() or something to add axes and annotation.
>
> That you in advance for any help.
>
> Best regards,
> Jon Olav
>
> == Output of alco.ipy ==
>
> The columns are "number of series" and "seconds".
>
> In [8]: run alco.ipy
> 1000 9.07
> 2000 24.8
> 3000 44.73
> 4000 67.85
> 5000 95.67
> 6000 135.1
> 7000 177.82
> 8000 226.03
> 9000 278.32
> 10000 340.81
>
> == alco.ipy ==
>
> n, t = [], []
> for i in range(1000, 10001, 1000):
>    n.append(i)
>    ti = !python alco.py $i
>    t.append(float(ti.s))
>    print n[-1], t[-1]
>
> plot(n, t, '.-')
>
> == alco.py ==
>
> """Alpha compositing of line plots. Usage: python alco.py NSERIES ALPHA"""
> from sys import argv
> import numpy as np
> import matplotlib as mpl
> mpl.use("agg") # noninteractive plotting
> from pylab import *
>
> n = int(argv[1])
> try:
>    alpha = float(argv[2])
> except IndexError:
>    alpha = 0.02
>
> # generate some data
> x = np.arange(200)
> for i in range(n):
>    y = np.sin(x / (2 * np.pi * x[-1] * i))
>    plot(x, y, 'k-', alpha=alpha)
>
> savefig("test.png")
>
>
>
> ------------------------------------------------------------------------------
> Download Intel&#174; Parallel Studio Eval
> Try the new software tools for yourself. Speed compiling, find bugs
> proactively, and fine-tune applications for parallel performance.
> See why Intel Parallel Studio got high marks during beta.
> http://p.sf.net/sfu/intel-sw-dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

I want to overlay many line plots using alpha transparency. However, plotting 
them in Matplotlib takes about O(n**2) time, and I think I may be running into 
memory limitations as well.

As a simple benchmark, I used IPython to run alco.ipy (below), which runs 
alco.py for an increasing number of data series. Extrapolating from this, 
plotting 60000 series would take something like 200 minutes. This is similar to 
my actual use case, which takes about 3 hours to finish a plot. Zooming in and 
saving again is much faster, taking only about 30 seconds.

I would appreciate suggestions on how to speed this up. For instance:

Is there a memoryless "canvas" object that I could draw on, just accumulating 
the alpha in each pixel: new_alpha = old_alpha + (1 - old_alpha) * this_alpha.

Failing that, I could do it manually by keeping a Numpy array of the pixels in 
the image. For each series, find the x values corresponding to each column 
index, then interpolate to find the row index corresponding to each y value. 
Finally, use imshow() or something to add axes and annotation.

That you in advance for any help.

Best regards,
Jon Olav

== Output of alco.ipy ==

The columns are "number of series" and "seconds".

In [8]: run alco.ipy
1000 9.07
2000 24.8
3000 44.73
4000 67.85
5000 95.67
6000 135.1
7000 177.82
8000 226.03
9000 278.32
10000 340.81

== alco.ipy ==

n, t = [], []
for i in range(1000, 10001, 1000):
    n.append(i)
    ti = !python alco.py $i
    t.append(float(ti.s))
    print n[-1], t[-1]

plot(n, t, '.-')

== alco.py ==

"""Alpha compositing of line plots. Usage: python alco.py NSERIES ALPHA"""
from sys import argv
import numpy as np
import matplotlib as mpl
mpl.use("agg") # noninteractive plotting
from pylab import *

n = int(argv[1])
try:
    alpha = float(argv[2])
except IndexError:
    alpha = 0.02

# generate some data
x = np.arange(200)
for i in range(n):
    y = np.sin(x / (2 * np.pi * x[-1] * i))
    plot(x, y, 'k-', alpha=alpha)

savefig("test.png")