matplotlib-users Mailing List for matplotlib (Page 4)

Hi all,

I'm pleased to announce here that Spyder version 1.1.0 has been released:
http://packages.python.org/spyder

Spyder (the Scientific PYthon Development EnviRonment) is a free open-source Python development environment providing MATLAB-like features in a simple and light-weighted software, available for Windows XP/Vista/7, GNU/Linux and MacOS X:
    * advanced code editing features (code analysis, ...)
    * interactive console with MATLAB-like workspace (with GUI-based
list, dictionary, tuple, text and array editors -- screenshots:
http://packages.python.org/spyder/console.html#the-workspace) and
integrated matplotlib figures
    * external console to open an interpreter or run a script in a
separate process (with a global variable explorer providing the same
features as the interactive console's workspace)
    * code analysis with pyflakes and pylint
    * search in files features
    * object inspector: automatically retrieves docstrings or
source code of the function/class called in the interactive/external
console
    * online documentation viewer (pydoc)
    * integrated file/directories explorer
    * MATLAB-like path management
    * project management
    ...and more!

Spyder is part of spyderlib, a Python module based on PyQt4 and
QScintilla2 which provides powerful console-related PyQt4 widgets.

Some of the major changes since v1.0.0 (433 commits!):
    * A lot of bugfixes!
    * IPython integration within the external console (still experimental)
    * QScintilla2 is now optional (a whole pure PyQt4 code editor -faster than its QScintilla's counterpart- has been implemented): brings code folding and code completion
    * Improved Matplotlib's figure options feature (added support for image parameters, added an "Apply" button)
    * Added: Project Explorer plugin (Pydev projects may be imported)
    * Added: Online help browser plugin (based on pydoc)
    * Editor new features:
        * Unlimited horizontal/vertical splitting: each new editor panel is a clone of the first panel, allowing comparing two parts of the same file
        * Unlimited independent editor windows creation
        * Flag vertical scrollbar area: shows warnings, TODOs, FIXMEs and occurrence highlighting of the whole file
    * External console: added import/export features to the variable explorer

Cheers,
Pierre

Hi,

Ryan May <rm...@gm...>:
>
> Michael Droettboom <md...@st...> wrote:
> > There is a guide about porting from the old transforms to the new
> > transforms here:
> >
> > http://matplotlib.sourceforge.net/api/api_changes.html#notes-about-the-transforms-refactoring
> 
> It's possible I'm missing something, but I don't see
> matplotlib.transforms.Value() or anything remotely resembling it
> mentioned there.

Same with me. :)  Michael, I've checked the very document and just like
Ryan I couldn't find any hint of what had become of that functionality.

Here is a diff of transforms.py before and after the refactoring:

http://matplotlib.svn.sourceforge.net/viewvc/matplotlib/trunk/matplotlib/lib/matplotlib/transforms.py?r1=3513&r2=4817

On the left side you see that Value() has been featured quite
distinctively. After the refactoring there doesn't seem to be anything
left of it. To a newbie like me this is striking, as it seems to be a
major change in the API. I thus hoped it probably could have been moved
to another module or something like this.

Still glad about any enlightenments. :)

Best,
	Florian

I'm updating my python install but am running into problems with matplotlib, 
as always.  This time around I'd thought I'd use the .dmg installer.  I've 
already installed Python 2.6.5 from python.org and numpy from scipy.org. 
Both of these work.  I noticed the matplotlib installer wants to install to 
/Library/Python/2.6/site-packages, even though the appropriate location for 
the python.org install is in 
/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages. 
As a result, when I run python 2.6.5 from python.org, it doesn't matplotlib.

Two questions: Is there a good reason for this?  And what's the workaround. 
Thanks.

Tim

On Thu, Jun 24, 2010 at 1:11 PM, Michael Droettboom <md...@st...> wrote:
> There is a guide about porting from the old transforms to the new
> transforms here:
>
> http://matplotlib.sourceforge.net/api/api_changes.html#notes-about-the-transforms-refactoring

It's possible I'm missing something, but I don't see
matplotlib.transforms.Value() or anything remotely resembling it
mentioned there. I have no idea what the original method did, so I'm
not sure if this an omission or if I'm just dense.

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma

There is a guide about porting from the old transforms to the new 
transforms here:

http://matplotlib.sourceforge.net/api/api_changes.html#notes-about-the-transforms-refactoring

Mike

On 06/24/2010 01:53 PM, Florian Berger wrote:
> Hi,
>
> I've stumbled across an old application from 2007 which uses the old
> matplotlib.transforms API, namely matplotlib.transforms.Value() which
> obviously disappeared in a great transforms overhaul.
>
> I tried to figure out what has become of these classes and functions
> browsing the changelog and SVN, without success - there doesn't seem to
> be a successor to these classes.
>
> As I would love to tweak that particular application to make it
> work again: could someone enlighten me what has become of
> matplotlib.transforms.Value() and how to replace the functionality?
>
> Thanks a ton!
>
> Florian
>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>    


-- 
Michael Droettboom
Science Software Branch
Space Telescope Science Institute
Baltimore, Maryland, USA

Hi,

I've stumbled across an old application from 2007 which uses the old
matplotlib.transforms API, namely matplotlib.transforms.Value() which
obviously disappeared in a great transforms overhaul.

I tried to figure out what has become of these classes and functions
browsing the changelog and SVN, without success - there doesn't seem to
be a successor to these classes.

As I would love to tweak that particular application to make it
work again: could someone enlighten me what has become of
matplotlib.transforms.Value() and how to replace the functionality?

Thanks a ton!

Florian

Can someone point me to up to date installation instructions for Mac OS (10.6.4, python 2.6.4)?

Every route I pursue hits a dead end.
- if I install from dmg files on the matplotlib, it is not visible to my python installation (I think it installs to the OSX native python).
- if I install via easy_install or pip, then I get an error (previously emailed about) saying that numpy is incorrectly installed (Numpy tests pass)
- if I try to follow the instructions on http://blog.hyperjeff.net/?p=160, I get errors such as:
---
tar: Unrecognized archive format: Inappropriate file type or format
tar: Error exit delayed from previous errors.
make: *** [zlib] Error 1
---

Ages ago I successfully installed on my laptop but I do not seem to be able to repeat the feat on my desktop.
Any help appreciated.

Christopher

On 06/24/2010 04:03 AM, Benjamin Root wrote:
> Ah, I just noticed that.
>
> Actually, I think I just figured out what is happening.  The colorbar
> automatically chooses what values to display, and in your case, the
> values are 0.0, -0.4, -0.8, -1.2, -1.6, ..., -3.6, which when turned
> into integers are 0, 0, 0, -1, -1, ..., -3, which matches what your

Right, good eye!

> first plot had.  All we have to do is specify the values that the
> colorbar should list.  I am not very familiar with this aspect, but it
> has something to do with specifying your colormap and/or the normalizer.

The tick locations can be specified via the ticks kwarg to colorbar. 
 From the docstring:

         *ticks*       [ None | list of ticks | Locator object ]
                       If None, ticks are determined automatically from the
                       input.
         *format*      [ None | format string | Formatter object ]
                       If None, the
                       :class:`~matplotlib.ticker.ScalarFormatter` is used.
                       If a format string is given, e.g. '%.3f', that is
                       used. An alternative
                       :class:`~matplotlib.ticker.Formatter` object may be
                       given instead.

The OP may also wish to specify the range of the colormap via the vmin 
and vmax kwargs to pcolor (note that everything to do with the colormap 
and the norm is specified in pcolor, not in the call to colorbar.

The OP probably does not need to specify both the formatter and the 
ticker.  If it is certain that the range of values will be substantially 
geater than one, and integer tick values are desired, then try this:

int_ticker = MaxNLocator(nbins=6, integer=True)
cbar = colorbar(ticks=int_ticker)

nbins is the maximum number of intervals (one less than the max number 
of ticks).

The default formatter will print integers as integers, so no custom 
formatter is needed.  Specifying a custom format or formatter is risky 
because, as illustrated by your diagnosis of the OP's original problem, 
it can easily lead to labels that are not accurate representations of 
the tick values.


>
> Does anybody know of a good tutorial on creating colormaps and normalizers?

Custom norms are rare, and I don't think there is much in the way of 
documentation or examples.  The code in ticker.py is probably the best 
starting place for learning about customizing norms; it includes 
subclasses of Normalize. Colormaps are illustrated fairly well in the 
examples.  See the second example in 
http://matplotlib.sourceforge.net/examples/pylab_examples/contourf_demo.html 
for ListedColormap, see 
http://matplotlib.sourceforge.net/examples/pylab_examples/contourf_demo.html 
for examples of how to use the somewhat complicated 
LinearSegmentedColormap to generate a custom map, and 
http://matplotlib.sourceforge.net/examples/pylab_examples/show_colormaps.html 
for a view of built-in colormaps.

Eric

>
> Ben Root
>
>
> On Thu, Jun 24, 2010 at 8:53 AM, Benoit Donnet
> <Ben...@uc... <mailto:Ben...@uc...>> wrote:
>
>     Hello Benjamin,
>
>     Thanks for your reply.
>
>>     Is there any particular reason why you can't do a log10() of the
>>     data that is being pcolor()'d and then label the colorbar as
>>     having units of dB?  That would seem to be the most
>>     straight-forward approach to me.
>
>     That's what I first tested.
>
>     In that case, labels on the colorbar are the following (i don't
>     attach the plot to avoid spamming the entire mailing-list): 0.0,
>     -0.4, -0.8, -1.2, -1.6, ..., -3.6).  It does not mean anything :s
>
>     Benoit
>
>>
>>     Ben Root
>>
>>
>>     2010/6/24 Benoit Donnet <ben...@uc...
>>     <mailto:ben...@uc...>>
>>
>>         Hi guys,
>>
>>         I'm struggling with colorbar since this morning.  I'd like the
>>         colorbar being logscale
>>
>>         I'm experimenting some strange behavior with the colorbar as
>>         some 'labels' appear several times.  For instance, 10^0
>>         appears three times, 10^1 appears also three times, and so on.
>>          I believe the exponent is the digit of the float, while I'd
>>         like to see the exponent (of the scientific notation)
>>
>>         I attach a png of the plot
>>
>>         Here is my code:
>>
>>         k,m,fp = np.loadtxt(file, unpack=True)
>>         ki = linspace(k.min(), k.max(), 37)
>>         mi = linspace(m.min(), m.max(), 37)
>>         Z = griddata(k, m, fp, ki, mi)
>>         Z.shape
>>         K, M = meshgrid(ki, mi)
>>
>>         pcolor(K, M, log10(Z))#, cmap=cm.gray)
>>         cbar = colorbar(format=FormatStrFormatter('$10^{%d}$'))
>>
>>         semilogy()
>>         axis([1,20,10000,500000], font2)
>>         xlabel(r'\textrm{\# hash functions ($k$)}', font)
>>         ylabel(r'\textrm{vector size ($m$)}', font)
>>         cbar.ax.set_ylabel(r'$f_p$', font)
>>
>>         I obviously suspect my code is flawed somewhere but I can't
>>         figure out where.  i have tested several format for the
>>         colorbar, like LogFormatterMathText, but it does not solve my
>>         problem.
>>
>>         I would appreciate any kind of help.  Thanks in advance
>>
>>         Keep on Rockin'
>>
>>         Benoit
>>
>>
>>
>>
>>         ---
>>         Dr. Benoit Donnet
>>         Université catholique de Louvain (UCL)
>>         Ecole Polytechnique de Louvain (EPL) - Département
>>         d'Ingénierie Informatique (INGI)
>>         Place Sainte Barbe, 2
>>         B-1348 Louvain-la-Neuve
>>         Phone: +32 10 47 87 18
>>         Fax: +32 10 45 03 45
>>
>>
>>
>>
>>
>>
>>         ------------------------------------------------------------------------------
>>         ThinkGeek and WIRED's GeekDad team up for the Ultimate
>>         GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
>>         lucky parental unit.  See the prize list and enter to win:
>>         http://p.sf.net/sfu/thinkgeek-promo
>>         _______________________________________________
>>         Matplotlib-users mailing list
>>         Mat...@li...
>>         <mailto:Mat...@li...>
>>         https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>>
>
>     ---
>     Dr. Benoit Donnet
>     Université catholique de Louvain (UCL)
>     Ecole Polytechnique de Louvain (EPL) - Département d'Ingénierie
>     Informatique (INGI)
>     Place Sainte Barbe, 2
>     B-1348 Louvain-la-Neuve
>     Phone: +32 10 47 87 18
>     Fax: +32 10 45 03 45
>
>
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
>
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On 6/24/2010 9:49 AM, João Luís Silva wrote:
> Hi,
>
> This simple script will animate correctly on Linux, but will not work on
> Windows (mpl 0.99.3) and at the end will crash with a message box
> (unknown software exception (0x40000015) at the location 0x1e05b62a) and
> prints to the console:
>
> Fatal Python error: PyEval_RestoreThread: NULL tstate
>
> This application has requested the Runtime to terminate it in an unusual
> way.
> Please contact the application's support team for more information.
>


I can not reproduce this bug with matplotlib 1.0dev on Windows. So this 
is apparently fixed in svn trunk, probably by 
<http://matplotlib.svn.sourceforge.net/viewvc/matplotlib/trunk/matplotlib/lib/matplotlib/pyplot.py?r1=8352&r2=8424&pathrev=8424> 
which is related to 
<http://sourceforge.net/tracker/?func=detail&aid=3015013&group_id=80706&atid=560720>

Christoph

The crash happens with the TKAgg backend but not with the GTKAgg 
backend, but the script will still not animate. This one will however:

#---------------------------------------
import matplotlib
matplotlib.use("GTKAgg")
import matplotlib.pyplot as plt
import numpy as np
plt.ion()

x = np.arange(0,2*np.pi,0.01)            # x-array
data = np.array(np.sin(x))
line, = plt.plot(x,data)

for i in range(1,50):
     data[:] = np.sin(x+i/10.0)
     line.set_ydata(data+i-i)
     plt.draw()

plt.ioff()
#---------------------------------------

Note the +i-i on the set_ydata. For some reason set_ydata won't update 
on Windows if the array is the same, even if the data has changed.

João Luís Silva

Hi,

This simple script will animate correctly on Linux, but will not work on 
  Windows (mpl 0.99.3) and at the end will crash with a message box 
(unknown software exception (0x40000015) at the location 0x1e05b62a) and 
prints to the console:

Fatal Python error: PyEval_RestoreThread: NULL tstate

This application has requested the Runtime to terminate it in an unusual 
way.
Please contact the application's support team for more information.

Please check this. Thank you,
João Luís Silva

(script inline and attached)

#-----------------------------------------------------------------------
import matplotlib.pyplot as plt
import numpy as np
plt.ion()

x = np.arange(0,2*np.pi,0.01)            # x-array
data = np.array(np.sin(x))
line, = plt.plot(x,data)

for i in range(1,50):
     data[:] = np.sin(x+i/10.0)
     line.set_ydata(data)
     plt.draw()                        # redraw the canvas
#-----------------------------------------------------------------------

Ah, I just noticed that.

Actually, I think I just figured out what is happening.  The colorbar
automatically chooses what values to display, and in your case, the values
are 0.0, -0.4, -0.8, -1.2, -1.6, ..., -3.6, which when turned into integers
are 0, 0, 0, -1, -1, ..., -3, which matches what your first plot had.  All
we have to do is specify the values that the colorbar should list.  I am not
very familiar with this aspect, but it has something to do with specifying
your colormap and/or the normalizer.

Does anybody know of a good tutorial on creating colormaps and normalizers?

Ben Root


On Thu, Jun 24, 2010 at 8:53 AM, Benoit Donnet
<Ben...@uc...>wrote:

> Hello Benjamin,
>
> Thanks for your reply.
>
> Is there any particular reason why you can't do a log10() of the data that
> is being pcolor()'d and then label the colorbar as having units of dB?  That
> would seem to be the most straight-forward approach to me.
>
>
> That's what I first tested.
>
> In that case, labels on the colorbar are the following (i don't attach the
> plot to avoid spamming the entire mailing-list): 0.0, -0.4, -0.8, -1.2,
> -1.6, ..., -3.6).  It does not mean anything :s
>
> Benoit
>
>
> Ben Root
>
>
> 2010/6/24 Benoit Donnet <ben...@uc...>
>
>> Hi guys,
>>
>> I'm struggling with colorbar since this morning.  I'd like the colorbar
>> being logscale
>>
>> I'm experimenting some strange behavior with the colorbar as some 'labels'
>> appear several times.  For instance, 10^0 appears three times, 10^1 appears
>> also three times, and so on.  I believe the exponent is the digit of the
>> float, while I'd like to see the exponent (of the scientific notation)
>>
>> I attach a png of the plot
>>
>> Here is my code:
>>
>> k,m,fp = np.loadtxt(file, unpack=True)
>> ki = linspace(k.min(), k.max(), 37)
>> mi = linspace(m.min(), m.max(), 37)
>> Z = griddata(k, m, fp, ki, mi)
>> Z.shape
>> K, M = meshgrid(ki, mi)
>>
>> pcolor(K, M, log10(Z))#, cmap=cm.gray)
>> cbar = colorbar(format=FormatStrFormatter('$10^{%d}$'))
>>
>> semilogy()
>> axis([1,20,10000,500000], font2)
>> xlabel(r'\textrm{\# hash functions ($k$)}', font)
>> ylabel(r'\textrm{vector size ($m$)}', font)
>> cbar.ax.set_ylabel(r'$f_p$', font)
>>
>> I obviously suspect my code is flawed somewhere but I can't figure out
>> where.  i have tested several format for the colorbar, like
>> LogFormatterMathText, but it does not solve my problem.
>>
>> I would appreciate any kind of help.  Thanks in advance
>>
>> Keep on Rockin'
>>
>> Benoit
>>
>>
>>
>>
>> ---
>> Dr. Benoit Donnet
>> Université catholique de Louvain (UCL)
>> Ecole Polytechnique de Louvain (EPL) - Département d'Ingénierie
>> Informatique (INGI)
>> Place Sainte Barbe, 2
>> B-1348 Louvain-la-Neuve
>> Phone: +32 10 47 87 18
>> Fax: +32 10 45 03 45
>>
>>
>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> ThinkGeek and WIRED's GeekDad team up for the Ultimate
>> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
>> lucky parental unit.  See the prize list and enter to win:
>> http://p.sf.net/sfu/thinkgeek-promo
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>>
>
>  ---
> Dr. Benoit Donnet
> Université catholique de Louvain (UCL)
> Ecole Polytechnique de Louvain (EPL) - Département d'Ingénierie
> Informatique (INGI)
> Place Sainte Barbe, 2
> B-1348 Louvain-la-Neuve
> Phone: +32 10 47 87 18
> Fax: +32 10 45 03 45
>
>
>
>
>
>

Hello Benjamin,

Thanks for your reply.

> Is there any particular reason why you can't do a log10() of the data that is being pcolor()'d and then label the colorbar as having units of dB?  That would seem to be the most straight-forward approach to me.

That's what I first tested.

In that case, labels on the colorbar are the following (i don't attach the plot to avoid spamming the entire mailing-list): 0.0, -0.4, -0.8, -1.2, -1.6, ..., -3.6).  It does not mean anything :s

Benoit 

> 
> Ben Root
> 
> 
> 2010/6/24 Benoit Donnet <ben...@uc...>
> Hi guys,
> 
> I'm struggling with colorbar since this morning.  I'd like the colorbar being logscale
> 
> I'm experimenting some strange behavior with the colorbar as some 'labels' appear several times.  For instance, 10^0 appears three times, 10^1 appears also three times, and so on.  I believe the exponent is the digit of the float, while I'd like to see the exponent (of the scientific notation)
> 
> I attach a png of the plot
> 
> Here is my code:
> 
> k,m,fp = np.loadtxt(file, unpack=True)
> ki = linspace(k.min(), k.max(), 37)
> mi = linspace(m.min(), m.max(), 37)
> Z = griddata(k, m, fp, ki, mi)
> Z.shape
> K, M = meshgrid(ki, mi)
> 
> pcolor(K, M, log10(Z))#, cmap=cm.gray)
> cbar = colorbar(format=FormatStrFormatter('$10^{%d}$'))
> 
> semilogy()
> axis([1,20,10000,500000], font2)
> xlabel(r'\textrm{\# hash functions ($k$)}', font)
> ylabel(r'\textrm{vector size ($m$)}', font)
> cbar.ax.set_ylabel(r'$f_p$', font)
> 
> I obviously suspect my code is flawed somewhere but I can't figure out where.  i have tested several format for the colorbar, like LogFormatterMathText, but it does not solve my problem.
> 
> I would appreciate any kind of help.  Thanks in advance
> 
> Keep on Rockin'
> 
> Benoit
> 
> 
> 
> 
> ---
> Dr. Benoit Donnet
> Université catholique de Louvain (UCL)
> Ecole Polytechnique de Louvain (EPL) - Département d'Ingénierie Informatique (INGI)
> Place Sainte Barbe, 2
> B-1348 Louvain-la-Neuve
> Phone: +32 10 47 87 18
> Fax: +32 10 45 03 45
> 
> 
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
> 

---
Dr. Benoit Donnet
Université catholique de Louvain (UCL)
Ecole Polytechnique de Louvain (EPL) - Département d'Ingénierie Informatique (INGI)
Place Sainte Barbe, 2
B-1348 Louvain-la-Neuve
Phone: +32 10 47 87 18
Fax: +32 10 45 03 45

Benoit,

Is there any particular reason why you can't do a log10() of the data that
is being pcolor()'d and then label the colorbar as having units of dB?  That
would seem to be the most straight-forward approach to me.

Ben Root


2010/6/24 Benoit Donnet <ben...@uc...>

> Hi guys,
>
> I'm struggling with colorbar since this morning.  I'd like the colorbar
> being logscale
>
> I'm experimenting some strange behavior with the colorbar as some 'labels'
> appear several times.  For instance, 10^0 appears three times, 10^1 appears
> also three times, and so on.  I believe the exponent is the digit of the
> float, while I'd like to see the exponent (of the scientific notation)
>
> I attach a png of the plot
>
> Here is my code:
>
> k,m,fp = np.loadtxt(file, unpack=True)
> ki = linspace(k.min(), k.max(), 37)
> mi = linspace(m.min(), m.max(), 37)
> Z = griddata(k, m, fp, ki, mi)
> Z.shape
> K, M = meshgrid(ki, mi)
>
> pcolor(K, M, log10(Z))#, cmap=cm.gray)
> cbar = colorbar(format=FormatStrFormatter('$10^{%d}$'))
>
> semilogy()
> axis([1,20,10000,500000], font2)
> xlabel(r'\textrm{\# hash functions ($k$)}', font)
> ylabel(r'\textrm{vector size ($m$)}', font)
> cbar.ax.set_ylabel(r'$f_p$', font)
>
> I obviously suspect my code is flawed somewhere but I can't figure out
> where.  i have tested several format for the colorbar, like
> LogFormatterMathText, but it does not solve my problem.
>
> I would appreciate any kind of help.  Thanks in advance
>
> Keep on Rockin'
>
> Benoit
>
>
>
>
> ---
> Dr. Benoit Donnet
> Université catholique de Louvain (UCL)
> Ecole Polytechnique de Louvain (EPL) - Département d'Ingénierie
> Informatique (INGI)
> Place Sainte Barbe, 2
> B-1348 Louvain-la-Neuve
> Phone: +32 10 47 87 18
> Fax: +32 10 45 03 45
>
>
>
>
>
>
>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

Hi guys,

I'm struggling with colorbar since this morning.  I'd like the colorbar being logscale

I'm experimenting some strange behavior with the colorbar as some 'labels' appear several times.  For instance, 10^0 appears three times, 10^1 appears also three times, and so on.  I believe the exponent is the digit of the float, while I'd like to see the exponent (of the scientific notation)

I attach a png of the plot

Here is my code:

k,m,fp = np.loadtxt(file, unpack=True)
ki = linspace(k.min(), k.max(), 37)
mi = linspace(m.min(), m.max(), 37)
Z = griddata(k, m, fp, ki, mi)
Z.shape
K, M = meshgrid(ki, mi)

pcolor(K, M, log10(Z))#, cmap=cm.gray)
cbar = colorbar(format=FormatStrFormatter('$10^{%d}$'))

semilogy()
axis([1,20,10000,500000], font2)
xlabel(r'\textrm{\# hash functions ($k$)}', font)
ylabel(r'\textrm{vector size ($m$)}', font)
cbar.ax.set_ylabel(r'$f_p$', font)

I obviously suspect my code is flawed somewhere but I can't figure out where.  i have tested several format for the colorbar, like LogFormatterMathText, but it does not solve my problem.

I would appreciate any kind of help.  Thanks in advance

Keep on Rockin'

Benoit

I stumbled across a basemap demo a couple of days ago and was immediately
enchanted.

I can't get it to install nicely on a Windoze box with matplotlib 0.99.3.
When I try the famous
  from mpl_toolkits.basemap import Basemap
directive, I get 
  In [1]: from mpl_toolkits.basemap import Basemap
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)

C:\Documents and Settings\Bill.DC1\<ipython console> in <module>()

C:\Python26\lib\site-packages\mpl_toolkits\basemap\__init__.py in <module>()
     36 from matplotlib.lines import Line2D
     37 from matplotlib.transforms import Bbox
---> 38 import pyproj, sys, os, math, dbflib
     39 from proj import Proj
     40 import numpy as np

C:\Python26\lib\site-packages\mpl_toolkits\basemap\pyproj.py in <module>()
     46 CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. """
     47
---> 48 from _proj import Proj as _Proj
     49 from _geod import Geod as _Geod
     50 from _proj import _transform

C:\Documents and Settings\Bill.DC1\c_numpy.pxd in _proj (src/_proj.c:3234)()

ValueError: numpy.dtype does not appear to be the correct type object

I tried manually installing many of the prerequisites (which should be
bundled) and also reinstalling MPL. But still no joy.

On my Linux box running 0.99.0, I was able to install basemap and run one of
the example codes. Works quite nicely. 

Thanks to Benjamin Root and Jae-Joon Lee for their responses.

The solution that I had come up with in the mean time is similar to
Jae-Joon's suggestion.  I did:

c = contour(z,level,colors='k')
xy = c.collections[0].get_paths()[0].vertices  # produces (N,2) array of points
plot(xy[:,0],xy[:,1],'w')

c.collections[0].get_paths() returns a list of Path objects.  These have
the attribute vertices which contains the values used to draw the
contour.  Jae-Joon's method is a bit more straightforward than mine,
though it's nice to know where those contour paths are stored.

Jon

On Tue, 2010-06-22 at 22:40 -0400, Jae-Joon Lee wrote:
> contour creates list of LineCollection objects (per each level I
> suppose) which is stored in "collections" attribute. For example,
> 
> cntr = contour(A, levels)
> 
> then
> 
> cntr.collections[i] is a LineCollection objects that is associated
> with levels[i].
> 
> And you can change colors of each line in the LineCollection object
> with set_edgecolors method.
> 
> So, assuming that there is two contour lines for the first level.
> 
> cntr.collections[0].set_edgecolors(["none", "r"])
> 
> will change the color of the first contour to "none" (i.e., not drawn)
> and the second one to red.
> 
> But you need to figure out which one is the one you want.
> 
> IHTH,
> 
> -JJ
> 
> 
> On Tue, Jun 22, 2010 at 2:46 PM, Jonathan Slavin
> <js...@cf...> wrote:
> > To all:
> >
> > I'm making a plot with an image and a contour on it.  I use only one
> > level in the call to contour, but it results in two distinct contours,
> > an inner closed one and an outer open one.  I want to plot only the
> > outer piece.  How might I go about that?  I've been looking at the
> > properties of the ContourSet object returned by the call to contour but
> > can't find anything useful yet.  Is there an attribute of ContourSet
> > objects that contains the (x,y) values for the contour?  Is there some
> > way to see that a ContourSet object has separate pieces?
> >
> > Any help would be appreciated.
> >
> > Thanks,
> > Jon
> >
> >
> > ------------------------------------------------------------------------------
> > ThinkGeek and WIRED's GeekDad team up for the Ultimate
> > GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> > lucky parental unit.  See the prize list and enter to win:
> > http://p.sf.net/sfu/thinkgeek-promo
> > _______________________________________________
> > Matplotlib-users mailing list
> > Mat...@li...
> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> >
-- 
______________________________________________________________
Jonathan D. Slavin              Harvard-Smithsonian CfA
js...@cf...         60 Garden Street, MS 83
phone: (617) 496-7981           Cambridge, MA 02138-1516
 cell: (781) 363-0035           USA
______________________________________________________________

Hi,

I'm trying to run a demo example,

http://matplotlib.sourceforge.net/plot_directive/mpl_toolkits/axes_grid/figures/demo_axes_grid.py

ran into this problem saying pygtk was needed:

$ curl
http://matplotlib.sourceforge.net/plot_directive/mpl_toolkits/axes_grid/figures/demo_axes_grid.py>
demo_axes_grid.py

$  python demo_axes_grid.py
Traceback (most recent call last):
  File "demo_axes_grid.py", line 1, in <module>
    import matplotlib.pyplot as plt
  File
"/Library/Frameworks/Python.framework/Versions/Current/lib/python2.6/site-packages/matplotlib/pyplot.py",
line 78, in <module>
    new_figure_manager, draw_if_interactive, show = pylab_setup()
  File
"/Library/Frameworks/Python.framework/Versions/Current/lib/python2.6/site-packages/matplotlib/backends/__init__.py",
line 25, in pylab_setup
    globals(),locals(),[backend_name])
  File
"/Library/Frameworks/Python.framework/Versions/Current/lib/python2.6/site-packages/matplotlib/backends/backend_gtkagg.py",
line 10, in <module>
    from matplotlib.backends.backend_gtk import gtk, FigureManagerGTK,
FigureCanvasGTK,\
  File
"/Library/Frameworks/Python.framework/Versions/Current/lib/python2.6/site-packages/matplotlib/backends/backend_gtk.py",
line 11, in <module>
    raise ImportError("Gtk* backend requires pygtk to be installed.")
ImportError: Gtk* backend requires pygtk to be installed.



Then I tried to install pygtk:

$  pip install pygtk
Downloading/unpacking pygtk
  Running setup.py egg_info for package pygtk
    Traceback (most recent call last):
      File "<string>", line 14, in <module>
      File
"/Users/michaelanderson/root/mikeWork/2010June/temp/build/pygtk/setup.py",
line 22, in <module>
        from dsextras import get_m4_define, getoutput, have_pkgconfig, \
    ImportError: No module named dsextras
    Complete output from command python setup.py egg_info:
    Traceback (most recent call last):

  File "<string>", line 14, in <module>

  File
"/Users/michaelanderson/root/mikeWork/2010June/temp/build/pygtk/setup.py",
line 22, in <module>

    from dsextras import get_m4_define, getoutput, have_pkgconfig, \

ImportError: No module named dsextras

----------------------------------------
Command python setup.py egg_info failed with error code 1
Storing complete log in /Users/michaelanderson/.pip/pip.log


Ok, then I tried to install dsextras:
$  pip install dsextras
Downloading/unpacking dsextras
  Could not find any downloads that satisfy the requirement dsextras
No distributions at all found for dsextras
Storing complete log in /Users/michaelanderson/.pip/pip.log


What is this obscure package (dsextras) and is it really necessary to run
the shared axes demo?

Mike

Hello,

I have Matplotlib running embedded in a wxPython app with the ability
to show/hide the toolbar.  When I hide the toolbar the current tool
remains active (i.e. zooming) even though the toolbar is invisible.
Is there a way to turn off the tools when I hide the toolbar?

The toolbar I'm using is
matplotlib.backends.backend_wxagg.NavigationToolbar2WxAgg

Thanks,
David Morton

>> However if I am zoomed in on my graph looking at detail, then click the
>> radio button, the scatter plot forces the graph to resize to once again
>> show the overall intial view (zoomed out).
>>      
> Try using:
>
> axes.set_autoscale_on(False)
>    

Thank you very much, this does exactly what I was after.

I call this BEFORE adding my scatter plot, and the current view (zoom 
in) remains.

Much appreciated, thanks again

- Steve

contour creates list of LineCollection objects (per each level I
suppose) which is stored in "collections" attribute. For example,

cntr = contour(A, levels)

then

cntr.collections[i] is a LineCollection objects that is associated
with levels[i].

And you can change colors of each line in the LineCollection object
with set_edgecolors method.

So, assuming that there is two contour lines for the first level.

cntr.collections[0].set_edgecolors(["none", "r"])

will change the color of the first contour to "none" (i.e., not drawn)
and the second one to red.

But you need to figure out which one is the one you want.

IHTH,

-JJ


On Tue, Jun 22, 2010 at 2:46 PM, Jonathan Slavin
<js...@cf...> wrote:
> To all:
>
> I'm making a plot with an image and a contour on it.  I use only one
> level in the call to contour, but it results in two distinct contours,
> an inner closed one and an outer open one.  I want to plot only the
> outer piece.  How might I go about that?  I've been looking at the
> properties of the ContourSet object returned by the call to contour but
> can't find anything useful yet.  Is there an attribute of ContourSet
> objects that contains the (x,y) values for the contour?  Is there some
> way to see that a ContourSet object has separate pieces?
>
> Any help would be appreciated.
>
> Thanks,
> Jon
>
>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

On Tue, Jun 22, 2010 at 8:56 PM, Stephen George
<ste...@op...> wrote:
> I have an application that draws a line plot of  a spectrum. When the
> spectrum is collected different gains and filters may be used for each
> data point (which I have also collected). I am looking at artefacts in
> the spectrum and trying to correlate them with things such as the gain
> and filter changes etc.
>
> On the application I have a number of radio buttons, when clicked will
> add a scatter plot of the datapoints but color coded by the item of
> interest.
> i.e.
> click the gain btn  I end up with the line plot, and each data point has
> a color coded dot whose color is keyed to the gain the data point was
> taken at.
> click the filter btn  I remove the gain scatter plot, and add a filter
> scatter plot where each data point is color coded with the filter used.
>
> This functionality work fine.
>
> However if I am zoomed in on my graph looking at detail, then click the
> radio button, the scatter plot forces the graph to resize to once again
> show the overall intial view (zoomed out).

Try using:

axes.set_autoscale_on(False)

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma

Stephen,

Most likely, the program isn't really going back to the "original" axes as
much that it is automatically setting the axes to fit all the data from the
new plot (which would likely be the "original axes", but only by
coincidence).  I am sure there is some sort of easy way to do this, but the
brute-force way would be for any action to create a plot to first check and
see what the current x and y lims are and save them to temporary variables.
Then, after creating the plots, call set the x and y lims from the temporary
variables.

Note that there might be an issue with the first graph, because you don't
want to set the axes after creating the graph using the information prior to
the graphing.  I don't know how one would detect that, besides some sort of
counter and an if-statement.

Maybe someone else has a better way?

Ben Root


On Tue, Jun 22, 2010 at 8:56 PM, Stephen George
<ste...@op...>wrote:

>
> Hi,
>
> I have an application that draws a line plot of  a spectrum. When the
> spectrum is collected different gains and filters may be used for each
> data point (which I have also collected). I am looking at artefacts in
> the spectrum and trying to correlate them with things such as the gain
> and filter changes etc.
>
> On the application I have a number of radio buttons, when clicked will
> add a scatter plot of the datapoints but color coded by the item of
> interest.
> i.e.
> click the gain btn  I end up with the line plot, and each data point has
> a color coded dot whose color is keyed to the gain the data point was
> taken at.
> click the filter btn  I remove the gain scatter plot, and add a filter
> scatter plot where each data point is color coded with the filter used.
>
> This functionality work fine.
>
> However if I am zoomed in on my graph looking at detail, then click the
> radio button, the scatter plot forces the graph to resize to once again
> show the overall intial view (zoomed out).
>
> I am wondering how can I add the scatter plot, without changing the
> current view (zoom level) that I am currently using, but still add all
> the scatter plot data?
>
> Any suggestions gratefully accepted.
>
> Steve
>
>
>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hi,

I have an application that draws a line plot of  a spectrum. When the 
spectrum is collected different gains and filters may be used for each 
data point (which I have also collected). I am looking at artefacts in 
the spectrum and trying to correlate them with things such as the gain 
and filter changes etc.

On the application I have a number of radio buttons, when clicked will 
add a scatter plot of the datapoints but color coded by the item of 
interest.
i.e.
click the gain btn  I end up with the line plot, and each data point has 
a color coded dot whose color is keyed to the gain the data point was 
taken at.
click the filter btn  I remove the gain scatter plot, and add a filter 
scatter plot where each data point is color coded with the filter used.

This functionality work fine.

However if I am zoomed in on my graph looking at detail, then click the 
radio button, the scatter plot forces the graph to resize to once again 
show the overall intial view (zoomed out).

I am wondering how can I add the scatter plot, without changing the 
current view (zoom level) that I am currently using, but still add all 
the scatter plot data?

Any suggestions gratefully accepted.

Steve

Actually, I just re-read your original message and noticed that you were
specifying your levels (I believe).  The double set of contours depends on
what your values are.  If you want to make absolutely sure that there aren't
extra lines, you could contour a boolean array:

contour(x, y, z > 4.5, [0, 1])

That should do the trick as well (assuming you know the level that you want
the isopleth for).

Ben Root

On Tue, Jun 22, 2010 at 6:27 PM, Benjamin Root <ben...@ou...> wrote:

> Jon,
>
> One thing you can do is to manually specify the levels to contour for in
> the contour call, or just specify the number of contours (and contour() will
> figure out the levels for you).  The fourth argument to contour() allows you
> to give a sequence of values (or an integer) for the isopleths.  So, if you
> want just one line (but have it chosen automatically):
>
> contour(x, y, z, 1)
>
> If you want a contour to always be for the value of 4.5, for example, then:
>
> contour(x, y, z, [4.5])
>
> Should do the trick.
>
> I hope that helps,
> Ben Root
>
>
>
>
> On Tue, Jun 22, 2010 at 1:46 PM, Jonathan Slavin <js...@cf...>wrote:
>
>> To all:
>>
>> I'm making a plot with an image and a contour on it.  I use only one
>> level in the call to contour, but it results in two distinct contours,
>> an inner closed one and an outer open one.  I want to plot only the
>> outer piece.  How might I go about that?  I've been looking at the
>> properties of the ContourSet object returned by the call to contour but
>> can't find anything useful yet.  Is there an attribute of ContourSet
>> objects that contains the (x,y) values for the contour?  Is there some
>> way to see that a ContourSet object has separate pieces?
>>
>> Any help would be appreciated.
>>
>> Thanks,
>> Jon
>>
>>
>>
>> ------------------------------------------------------------------------------
>> ThinkGeek and WIRED's GeekDad team up for the Ultimate
>> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
>> lucky parental unit.  See the prize list and enter to win:
>> http://p.sf.net/sfu/thinkgeek-promo
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>
>