matplotlib-users Mailing List for matplotlib

On Thu, 2004-04-22 at 06:11, Gary Ruben wrote:
> I've installed 0.53 on my Win98 and Win2000 PCs and both suffer a problem which seems specific to the default Agg backend. It's an interoperability problem with trying to use TkAgg from the SciTE editor IDE or from an IDLE module window.

As JDH suggested, TkAgg works (as far as I've tested anyway) with IDLE
by using the -n flag when starting IDLE;  I build -n into the shortcut
for IDLE on my desktop.

In general, TkAgg is known to work with:

python
idle -n
IPython

TkAgg is known not work with:

SciTE 
pythonw
Pythonwin
idle

Both of the latter shells fail with a RuntimeError "abnormal program
termination".

I checked on www.python.org about Tkinter and Pythonwin and they're
known not to work together so that explains TkAgg on Pythonwin.
http://www.python.org/topics/tkinter/trouble.html

I also looked into SciTE a little and discovered that it is related to
Scintilla which in turn was derived from Pythonwin.
http://www.scintilla.org/
This indicates to me that the same problem with Tkinter may be affecting
both (SciTE and Pythonwin)... but I am out on a limb.

Regards,
Todd
-- 
Todd Miller <jm...@st...>

On Thu, 2004-04-22 at 08:48, John Hunter wrote:
> >>>>> "Kenneth" == Kenneth McDonald <kmm...@wi...> writes:
> 
>     Kenneth> 1) (Simple) Is there a defined behavior for matplotlib
>     Kenneth> when it attempts to graph data containing NaN values?
>     Kenneth> (OK, I admit-- it's really, really late, and I have tried
>     Kenneth> it to see what happens.  But even that wouldn't tell me
>     Kenneth> if that was the _defined_ behavior :-))
> 
> No, it's not defined.  I don't know that NaN is defined across
> platforms in python.  See my recent question on comp.lang.python
> 
> http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=mailman.141.1080681106.20120.python-list%40python.org&rnum=3&prev=/groups%3Fq%3Dtest%2Bnan%2Bgroup%253A*python*%26ie%3DUTF-8%26oe%3DUTF-8%26hl%3Den%26btnG%3DGoogle%2BSearch
> 
> Perhaps Todd or Perry can comment on what the status of NaN vis-a-vis
> Numeric and numarray.  

I'm not aware of any functionality in Numeric for dealing with IEEE
special values.  (Looking in the Numeric manual suggests that problems
with these values lead to the development of the Masked Array capability
as an alternative.)  Lack of support for IEEE special values in Numeric
makes it difficult to provide a unified approach in matplotlib.

numarray has a module (ieeespecial) for dealing with different IEEE
special values, including NaN.  In numarray.ieeespecial are functions
for identifying the locations of and setting IEEE special values to some
other value.  My thought was that you could use these functions in your
own code to define whatever behavior you want.  This is perhaps more
work than is convenient but has the advantage that you can do it now
yourself.  

Regards,
Todd

-- 
Todd Miller <jm...@st...>

>>>>> "Yann" == Yann Le Du <yan...@no...> writes:

    Yann> So it looks as though you check for change in content, but
    Yann> not for change in size.

Yep, you're right.  mathtext does cache for efficiency and the
fontsize key was not being used in the cache.  This does not affect
normal text, which does the caching properly.  The good news it's a
one-line-fix.  In matplotlib.mathtext.math_parse_s, change the
cacheKey to

    cacheKey = (s, dpi, fontsize)

Thanks for the detailed report.

JDH

>>>>> "Flavio" == Flavio Codeco Coelho <fcc...@fi...> writes:

    Flavio> it worked in 0.5.2

    Flavio> What is the problem?

I implemented new tick locator in 0.53 which should provide nicer
default tick locations as well as better user customization.  I failed
to test for the equal data limits case, which several people have
already reported.  This is fixed in CVS.  

Early next week, probably Monday, I will accumulate all the reported
bugs and fixes into a 0.53.1 release, so please continue to let me
know as you find them.

Note if you actually want to see the result of your plot, you'll need
to add a marker

  from matplotlib.matlab import *
  plot ([1], 'o')

since otherwise you have a zero length line.


JDH

On Wed, 21 Apr 2004, John Hunter wrote:

> >>>>> "Yann" == Yann Le Du <yan...@no...> writes:
> 
>     Yann> puts the text horizontally instead of vertically. 
> 
> See http://matplotlib.sourceforge.net/matplotlib.mathtext.html for a
> list of known issues with mathtext.  Short answer, it's on the list of
> things to do.

Ok, thanks


>     Yann> don't get any effect using fontsize :
> 
> Fixed in matplotlib-0.53.  What, you mean you haven't upgraded yet?
> :-)

I have indeed ! Always keen to upgrade fine software !

And here's what I found out (with probable solution at the end). I run 
python shell, then I do :

from matplotlib.matlab import *
plot([1,2,3,4])
xlabel(r'$\alpha > \beta$',fontsize=5)
show()

and then this figure pops up, and it works fine. Now I close the window
and I do :

xlabel(r'$\alpha > \beta$',fontsize=20)
show()

but nothing changes in the size of the x label.

Now if I change the size AND the content, then it works fine :

xlabel(r'$\alpha$',fontsize=20)
show()

This gives me a size 20 alpha.

So it looks as though you check for change in content, but not for change
in size.

YLD

I'm trying to run the first tutorial example in matplotlib-0.53 and am 
running into an ImportError. Here's the python script I'm trying to run:

import matplotlib
matplotlib.use('TkAgg')
from matplotlib.matlab import *
plot([1,2,3,4])
show()

Here's the output I'm getting:


light:gerry> python matplotlibtest.py
Traceback (most recent call last):
  File "matplotlibtest.py", line 15, in ?
    from matplotlib.matlab import *
  File 
"/d2/gerry/local/ActivePython-2.3/lib/python2.3/site-packages/matplotlib/matlab.py", 
line 129, in ?
    from backends import new_figure_manager, error_msg, \
  File 
"/d2/gerry/local/ActivePython-2.3/lib/python2.3/site-packages/matplotlib/backends/__init__.py", 
line 29, in ?
    from backend_tkagg import error_msg, draw_if_interactive, show, 
new_figure_manager
  File 
"/d2/gerry/local/ActivePython-2.3/lib/python2.3/site-packages/matplotlib/backends/backend_tkagg.py", 
line 8, in ?
    import tkagg                 # Paint image to Tk photo blitter extension
  File 
"/d2/gerry/local/ActivePython-2.3/lib/python2.3/site-packages/matplotlib/backends/tkagg.py", 
line 1, in ?
    import _tkagg
ImportError: 
/d2/gerry/local/ActivePython-2.3/lib/python2.3/site-packages/matplotlib/backends/_tkagg.so: 
undefined symbol: XFreePixmap


Thanks for any help.

--Gerry Wiener

I had been using multi-line ticklabels in my plots. They worked fine in
in 0.52 (except for postscript output).  Much to my dismay, I find that
in 0.53 multi-line ticklabels do not work at all. The embedded '\n'
shows up as a small box, and the lines run together.

I hope that this will be fixed soon.

Thanks.

	-Al Schapira

Thanks for all the great features in matplotlib 0.53.

date_demo2.py with 0.53 looks normal initially. However, if you compress
the scale by clicking the (-) Horizontal magnifier, the new date
ticklabels that shift in from the right are horizontal, not vertical.

I found that if you expand and then compress, the original ticklabels
remain vertical, but the new ones (those not present in the original
plot) shift in horizontal. Looks like the 'rotation' attribute of the
additional ticklabels should have been copied from the original ones.

	-Al Schapira

I shouldn't have jumped to conclusions. Agg seems OK. GTKAgg works from SciTE and IDLE just using show().
If I substitute this:

manager = get_current_fig_manager()
manager.window.show()
raw_input('paused')

then a window is opened but nothing is painted onto the canvas area from either IDLE or SciTE.
It looks like GTKAgg might be the backend of choice for these environments.

Gary

----- Original Message -----
From: John Hunter <jdh...@ac...>
Date: Thu, 22 Apr 2004 08:09:52 -0500
To: "Gary Ruben" <ga...@em...>
Subject: Re: [Matplotlib-users] Agg backend problem with 0.53

> >>>>> "Gary" == Gary Ruben <ga...@em...> writes:
> 
>     Gary> It's evidently something to do with threads and is probably
>     Gary> an inherent problem with Agg. I'm just bringing this to the
>     Gary> attention of others here who may be having similar
>     Gary> problems. Perhaps John knows what's going on and whether
>     Gary> anything can be done about it. The convenience of working in
>     Gary> SciTE For the moment,
> 
> I would be very surprised if it's agg - I think it's a GUI conflict on
> the Tk side.  Todd is this related to the idle -n thingie?  If so,
> perhaps a FAQ and status report are warranted.
> 
> Gary, since you say GTK is working, you may also want to check GTKAgg
> which will help up narrow it down to Tk, and would rule out agg
> problems.  But since SciTE is GTK based, you probably need make sure
> that matplotlib is not firing up it's own gtk mainloop, by not calling
> 'show'.  You can show your GTK figures manually if necessary by doing
> 
>    manager = get_current_fig_manager()
>    manager.window.show()
> 
> This is a GTK specific call applicable only if the GTK mainloop is
> already running (as I suspect it is in SciTE, which I haven't used),
> It is probably a good idea to standardize the API here across GUIs to
> handle exactly this case.
> 
> The larger question is why would there be a GTK/Tk conflict?  In any
> case, let us know if you learn anything else.
> 
> JDH
> 

-- 
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://promo.mail.com/adsfreejump.htm

When compiling matplotlib on debian with :
python setup.py install --home=/home/jonas/python

I get an error message about not being able to find tk.h (appended
below). This is with (I believe) all necessary tk libraries installed.
When I set CPATH=/usr/include/tcl8.4 and then try, it compiles just
fine. Is there any way we could make the tk-finding code a bit smarter
so it would work out of the box on debian? 

Thanks for such an amazing plotting package -- at last, I might be able
to ditch matlab!
		...Eric

[full error below]

running install
running build
running build_py
running build_ext
building 'matplotlib.backends._tkagg' extension
gcc -pthread -fno-strict-aliasing -DNDEBUG -g -O3 -Wall
-Wstrict-prototypes -fPI
C -I/usr/lib/tcl8.4/../../include -I/usr/include -Isrc -Iagg2/include
-I/usr/inc
lude -I/usr/lib/tcl8.4/../../include/freetype2 -I/usr/include/freetype2
-Isrc/fr
eetype2 -Iagg2/include/freetype2 -I/usr/include/freetype2
-I/usr/include/python2
.3 -c src/_tkagg.cpp -o build/temp.linux-i686-2.3/src/_tkagg.o
src/_tkagg.cpp:20:19: tk.h: No such file or directory
src/_tkagg.cpp:28: error: syntax error before `*' token
src/_tkagg.cpp:32: error: `ClientData' was not declared in this scope
src/_tkagg.cpp:32: error: parse error before `,' token
src/_tkagg.cpp: In function `int PyAggImagePhoto(...)':
src/_tkagg.cpp:35: error: `Tk_PhotoHandle' undeclared (first use this
function)
src/_tkagg.cpp:35: error: (Each undeclared identifier is reported only
once for 
   each function it appears in.)
src/_tkagg.cpp:35: error: parse error before `;' token
src/_tkagg.cpp:36: error: `Tk_PhotoImageBlock' undeclared (first use
this 
   function)
src/_tkagg.cpp:40: error: `argc' undeclared (first use this function)
src/_tkagg.cpp:41: error: `interp' undeclared (first use this function)
src/_tkagg.cpp:41: error: `argv' undeclared (first use this function)
src/_tkagg.cpp:42: error: `Tcl_AppendResult' undeclared (first use this 
   function)
src/_tkagg.cpp:43: error: `TCL_ERROR' undeclared (first use this
function)
src/_tkagg.cpp:47: error: `photo' undeclared (first use this function)
src/_tkagg.cpp:47: error: `Tk_FindPhoto' undeclared (first use this
function)
src/_tkagg.cpp:65: error: `block' undeclared (first use this function)
src/_tkagg.cpp:88: error: `Tk_PhotoBlank' undeclared (first use this
function)
src/_tkagg.cpp:90: error: `Tk_PhotoPutBlock' undeclared (first use this 
   function)
src/_tkagg.cpp:92: error: `TCL_OK' undeclared (first use this function)
src/_tkagg.cpp: In function `PyObject* _tkinit(PyObject*, PyObject*)':
src/_tkagg.cpp:108: error: `Tcl_Interp' undeclared (first use this
function)
src/_tkagg.cpp:117: error: parse error before `)' token
src/_tkagg.cpp:122: error: 'struct TkappObject' has no member named
'interp'
src/_tkagg.cpp:128: error: `Tcl_CmdProc' undeclared (first use this
function)
src/_tkagg.cpp:128: error: parse error before `)' token
error: command 'gcc' failed with exit status 1

>>>>> "Gary" == Gary Ruben <ga...@em...> writes:

    Gary> It's evidently something to do with threads and is probably
    Gary> an inherent problem with Agg. I'm just bringing this to the
    Gary> attention of others here who may be having similar
    Gary> problems. Perhaps John knows what's going on and whether
    Gary> anything can be done about it. The convenience of working in
    Gary> SciTE For the moment,

I would be very surprised if it's agg - I think it's a GUI conflict on
the Tk side.  Todd is this related to the idle -n thingie?  If so,
perhaps a FAQ and status report are warranted.

Gary, since you say GTK is working, you may also want to check GTKAgg
which will help up narrow it down to Tk, and would rule out agg
problems.  But since SciTE is GTK based, you probably need make sure
that matplotlib is not firing up it's own gtk mainloop, by not calling
'show'.  You can show your GTK figures manually if necessary by doing

   manager = get_current_fig_manager()
   manager.window.show()

This is a GTK specific call applicable only if the GTK mainloop is
already running (as I suspect it is in SciTE, which I haven't used),
It is probably a good idea to standardize the API here across GUIs to
handle exactly this case.

The larger question is why would there be a GTK/Tk conflict?  In any
case, let us know if you learn anything else.

JDH

>>>>> "Kenneth" == Kenneth McDonald <kmm...@wi...> writes:

    Kenneth> 1) (Simple) Is there a defined behavior for matplotlib
    Kenneth> when it attempts to graph data containing NaN values?
    Kenneth> (OK, I admit-- it's really, really late, and I have tried
    Kenneth> it to see what happens.  But even that wouldn't tell me
    Kenneth> if that was the _defined_ behavior :-))

No, it's not defined.  I don't know that NaN is defined across
platforms in python.  See my recent question on comp.lang.python

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=mailman.141.1080681106.20120.python-list%40python.org&rnum=3&prev=/groups%3Fq%3Dtest%2Bnan%2Bgroup%253A*python*%26ie%3DUTF-8%26oe%3DUTF-8%26hl%3Den%26btnG%3DGoogle%2BSearch

Perhaps Todd or Perry can comment on what the status of NaN vis-a-vis
Numeric and numarray.  This has come up a number of times before, and
would be nice to be able to handle it.  As always, these ease of use
features imply a performance cost that the typical user may not want
to pay....

    Kenneth> 2) I expect to be using matplotlib quite a bit in the
    Kenneth> future, and will likely be using it in a more
    Kenneth> object-oriented mode (i.e. not through the matlab-style
    Kenneth> interface), since that's what I like better. I haven't
    Kenneth> seen any "howto" documents for using the OO API. Do any
    Kenneth> exist? I'd be happy to make notes and write a tutorial as
    Kenneth> I learn, but don't want to duplicate work that already
    Kenneth> exists.

I wrote a little example which I'll include below.  The pure OO API is
not really designed to be too user friendly.  Eg to instantiate a
line, you do

        vline = Line2D(
            dpi, bbox,
            xdata=x, ydata=y,
            color=color,
            antialiased=False,   # no need to antialias vert lines
            transx = ax.xaxis.transData,
            transy = ax.yaxis.transData)

It's useful if you want to have total control of the lines
transformations, bounding boxes and so on, but is overkill for making
most plots.  Likewise, instantiating your own Axes and Figures
requires a extra overhead.  This is addressed more in the example
below, which recommends a hybrid approach.

If you want to take some of this and extend it into a guide of sorts,
that would be great.  If what you are looking for is a developer's
guide which describes the process of Figure, Axes, Line2D, Text, etc,
creation and how to use them together, that doesn't exist yet.

However, matplotlib is undergoing rapid development and changes.  One
nice thing about having people use the matlab interface is that it
frees me to refactor the OO API.  There have been several major
refactorings to date.  If a lot of people are using this API, and it's
documented, it is more difficult to change.  I am not totally happy
with the current design (eg it us cumbersome to have to pass all those
objects just to create a line) so for now I prefer not to have too
many people creating lots of code with the OO API.  I think the
overall design is stable (Figures contain Axes which contain Text,
Axis and Lines, etc) but some of the constructor signatures may
change.

The hybrid approach I recommend below keeps you safely at the
interface level and insulated from any API changes, which can be
painful for application developers.  It does, however, enable you to
write more pythonic code.

Here is the new examples/pythonic_matplotlib.py:

"""
Some people prefer to use the python object oriented face rather than
the matlab interface to matplotlib.  This example show you how.

Unless you are an application developer, I recommend using part of the
matlab interface, particularly the figure, close, subplot, axes, and
show commands.  These hide a lot of complexity from you that you don't
need to see in normal figure creation, like instantiating DPI
instances, managing the bounding boxes of the figure elements,
creating and reaslizing GUI windows and embedding figures in them.

If you are an application developer and want to embed matplotlib in
your application, follow the lead of examples/embedding_in_wx.py,
examples/embedding_in_gtk.py or examples/embedding_in_tk.py.  In this
case you will want to control the creation of all your figures,
embedding them in application windows, etc.

If you seen an example in the examples dir written in matlab
interface, and you want to emulate that using the true python method
calls, there is an easy mapping.  Many of those examples use 'set' to
control figure properties.  Here's how to map those commands onto
instance methods

The syntax of set is

  set(object or sequence, somestring, attribute)

if called with an object, set calls

  object.set_somestring(attribute)

if called with a sequence, set does

  for object in sequence:
       object.set_somestring(attribute)

So for your example, if a is your axes object, you can do

  a.set_xticklabels([])
  a.set_yticklabels([])
  a.set_xticks([])
  a.set_yticks([])
"""


from matplotlib.matlab import figure, close, axes, subplot, show
from matplotlib.numerix import arange, sin, pi

t = arange(0.0, 1.0, 0.01)

fig = figure(1)

ax1 = subplot(211)
ax1.plot(t, sin(2*pi*t))
ax1.grid(True)
ax1.set_ylim( (-2,2) )
ax1.set_ylabel('1 Hz')
ax1.set_title('A sine wave or two')

for label in ax1.get_xticklabels():
    label.set_color('r')


ax2 = subplot(212)
ax2.plot(t, sin(2*2*pi*t))
ax2.grid(True)
ax2.set_ylim( (-2,2) )
l = ax2.set_xlabel('Hi mom')
l.set_color('g')
l.set_fontsize(15)

show()        

Just a little follow-up point. The normal axis autoscaling, with "plot(t, s, 'b-')" commented out, doesn't take account of the errorbar extrema. That is, the errorbars extend beyond the autoscaled window, whereas the window should be scaled to fully contain the errorbars.

----- Original Message -----
From: "Gary Ruben" <ga...@em...>
Date: Thu, 22 Apr 2004 20:19:10 +1000
To: mat...@li...
Subject: [Matplotlib-users] axis scaling anomaly

> Here's an example of my attempt to plot errorbar data with a series line of a different colour. This example reveals an anomaly with the axis scaling which may the sign of an underlying bug. If you comment out the "plot(t, s, 'b-')" line and run the example, the axes autoscale. If you run the example as is, the scaling changes for some reason.
> Gary Ruben
> :
> 
> 
> from matplotlib.matlab import *
> 
> t = arange(0.1, 4, 0.1)
> s = exp(-t)
> e = 0.1*randn(len(s))
> f = 0.1*randn(len(s))
> 
> grid(1)
> lines = errorbar(t, s, e, f, '.', capsize=4)
> set(lines, 'color', 'g')
> plot(t, s, 'b-')
> xlabel('Distance (m)')
> ylabel('Height (m)')
> title('Mean and standard error as a function of distance')
> show()



-- 
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://promo.mail.com/adsfreejump.htm

Here's an example of my attempt to plot errorbar data with a series line of a different colour. This example reveals an anomaly with the axis scaling which may the sign of an underlying bug. If you comment out the "plot(t, s, 'b-')" line and run the example, the axes autoscale. If you run the example as is, the scaling changes for some reason.
Gary Ruben
:


from matplotlib.matlab import *

t = arange(0.1, 4, 0.1)
s = exp(-t)
e = 0.1*randn(len(s))
f = 0.1*randn(len(s))

grid(1)
lines = errorbar(t, s, e, f, '.', capsize=4)
set(lines, 'color', 'g')
plot(t, s, 'b-')
xlabel('Distance (m)')
ylabel('Height (m)')
title('Mean and standard error as a function of distance')
show()
-- 
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://promo.mail.com/adsfreejump.htm

I've installed 0.53 on my Win98 and Win2000 PCs and both suffer a problem which seems specific to the default Agg backend. It's an interoperability problem with trying to use TkAgg from the SciTE editor IDE or from an IDLE module window. If I use the TkAgg backend from a DOS window, it works OK. Also, if I change the backend to GTK, all is well from all environments.

Here is the output if I run an example from SciTE under Win98. Win2000 fails similarly.

>pythonw -u errorbar_demo.pyw
Fatal Python error: PyEval_RestoreThread: NULL tstate

>Process failed to respond; forcing abrupt termination...>Exit code: 0

It's evidently something to do with threads and is probably an inherent problem with Agg. I'm just bringing this to the attention of others here who may be having similar problems. Perhaps John knows what's going on and whether anything can be done about it. The convenience of working in SciTE For the moment, 

Gary Ruben
-- 
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://promo.mail.com/adsfreejump.htm

1) (Simple) Is there a defined behavior for matplotlib when
it attempts to graph data containing NaN values? (OK, I admit--
it's really, really late, and I have tried it to see what happens.
But even that wouldn't tell me if that was the _defined_
behavior :-))

2) I expect to be using matplotlib quite a bit in the future,
and will likely be using it in a more object-oriented mode
(i.e. not through the matlab-style interface), since that's
what I like better. I haven't seen any "howto" documents
for using the OO API. Do any exist? I'd be happy to make
notes and write a tutorial as I learn, but don't want to
duplicate work that already exists.

Thanks,
Ken