matplotlib-users Mailing List for matplotlib

Tim Michelsen wrote:
> Hello users and developers of matplot,
> I would like to ask what kind of interpoaltion method is used by
> pylab.countourf()?
> 
> I suppose that it uses linear interpolation to derive the areas between 
> the contour lines.
> 
> Is that true?
> 
> Kind regards,
> Timmie
Timmie,
I am not sure exactly what you are asking. General contouring routines, 
such as the one in mpl, work by subdividing the domain into a set of 
contiguous polygons, calculating the points at which contours cross 
boundaries, and then connecting these points, usually with straight 
lines. Routines differ in the methods used to subdivide the domain and 
in the methods used to connect the points, since this can be ambiguous. 
 The cntr.c code in mpl subdivides the domain into the rectangles given 
by the quadrilateral grid on which the data are provided. In contrast, 
some other routines divide each such rectangle into 4 triangles by 
interpolating a point in the middle.
If you want to contour data that are not on a regular grid, or that are 
on a regular but coarse grid, or that are too noisy to be contoured 
nicely (assuming the noise has a small scale relative to the scale of 
the signal), then you need to use a gridding routine of your choice 
before contouring. It is the job of the gridding routine to do any 
necessary interpolation and/or smoothing.
Eric

Hello users and developers of matplot,
I would like to ask what kind of interpoaltion method is used by
pylab.countourf()?
I suppose that it uses linear interpolation to derive the areas between 
the contour lines.
Is that true?
Kind regards,
Timmie

Hi everybody,
Below you will find a small script that plots the graph of a simple function. 
This code is aimed to be embedded in a GUI application.
I set the subplot dimension with the b.set_position([]) command. The floats 
that are given to set_position() are percents relative to the Figure 
dimensions:
[ left corner x, left corner y, width, height ]
So, here I have put [0.1, 0.1, 0.8, 0.8 ] to have a small space between the 
subplot and the Figure border.
But when I resize the figure with the mouse, the space between the subplot and 
the Figure border is growing or decreasing, due to the relative definition of 
the dimensions.
I would like to be able to define subplot size and position in an absolute 
way. For example, I would like to have a space between the left border of the 
figure and the left border of the subplot being constant equal to 10px.
Is this possible?
Thanks a lot
Julien
####################
from pylab import *
import Tkinter as Tk
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
root = Tk.Tk()
t = arange(0.01, 5.0, 0.01)
s1 = sin(2*pi*t)
f = Figure( figsize = (8,7) )
veryplot = FigureCanvasTkAgg( f
 , master = root )
veryplot.get_tk_widget().pack( side = Tk.LEFT
 , expand = Tk.YES
 , fill = Tk.BOTH )
b = f.add_subplot( 211 )
b.plot( t, s1 )
# [ left corner x, left corner y, width, height ]
b.set_position( [ 0.1, 0.1, 0.8, 0.8 ] )
show()
root.mainloop()
####################

Steffen Wischmann <ste...@un...>
writes:
> I searched the forum and the net but I cannot figure out how I can 
> change the line colors for a boxplot (i.e., making all lines black such 
> as medians, whiskers, boxlines, black).
Change them after plotting, using the returned dictionary:
In [1]: data = random((10,10))
In [2]: r = boxplot(data)
In [3]: r.keys()
Out[3]: ['medians', 'fliers', 'whiskers', 'boxes', 'caps']
In [4]: setp(r['medians'], color='black')
Out[4]: [None, None, None, None, None, None, None, None, None, None]
This allows you to set the various drawing parameters of the different
parts separately, e.g.:
In [5]: setp(r['whiskers'], color='black', lw=2)
-- 
Jouni K. Seppänen
http://www.iki.fi/jks

Dear All,
Does anyone know of a citation that I can use for Numpy?
Thanks! and all the best,
--Buz

Hi,
I searched the forum and the net but I cannot figure out how I can 
change the line colors for a boxplot (i.e., making all lines black such 
as medians, whiskers, boxlines, black).
the only thing I accomplished is changing the color of all outliers, 
e.g., via:
boxplot(X, sym='r+')
Even setting the general line color like
rc('lines', color='k')
doesn't change anything.
Thanks for any help,
Steffen

Hi,
As far as I know, the destination coordinate of trans* is a display
(canvas) coordinate, not the normalized figure coordinate. It has a
dimension of f.get_figwidth()*f.get_dpi(),
f.get_figheight()*f.get_dpi().
For example, transFigure transforms the normalized figure coordinate
to the display coordinate.
See if something like below works for you.
 ax = gca()
 f = gcf()
 x1, y1, x2, y2 = 0.2, 0.3, 0.4, 0.5
 trans = ax.transData + f.transFigure.inverted()
 ax_x1, ax_y1 = trans.transform_point([x1, y1])
 ax_x2, ax_y2 = trans.transform_point([x2, y2])
 ax_dx, ax_dy = ax_x2 - ax_x1, ax_y2 - ax_y1
 a = axes([ax_x1, ax_y1,ax_dx, ax_dy])
IHTH,
-JJ
On Mon, Sep 29, 2008 at 8:10 AM, Yves Revaz <yve...@ep...> wrote:
> Dear List,
>
> I would like to define a new second plot inside a first plot using the
> axes command.
> But I need the position and size do be defined not by the relative
> figure coordinates
> but by data coordinates.
>
> I have found the following trick :
>
> ax = gca()
> f = gcf()
> x,y = ax.transData.transform([x,y])/f.transFigure.transform([1,1])
> a = axes([x,y,dy,dy])
>
> However, if the position is now ok, the size (dx,dy) is still on
> relative figure coordiate.
> Ok, I could think a bit more and find how its possible to transform it,
> but I would
> like to ask the list if there is an easier way to do that.
>
> By the way, why do I have to divide by f.transFigure.transform([1,1]) ?
> I expected that ax.transData.transform did return already normalized
> values.
>
>
> Thanks in advance.
>
> yves
>
>
>
> --
> (o o)
> --------------------------------------------oOO--(_)--OOo-------
> Yves Revaz
> Laboratory of Astrophysics EPFL
> Observatoire de Sauverny Tel : ++ 41 22 379 24 28
> 51. Ch. des Maillettes Fax : ++ 41 22 379 22 05
> 1290 Sauverny e-mail : Yve...@ep...
> SWITZERLAND Web : http://www.lunix.ch/revaz/
> ----------------------------------------------------------------
>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hi people,
I'm new to unix, python and matplotlib. I've had trouble installing matplotlib. The machine I'm on is running solaris. I tried installing it on my friend's ubuntu box and it worked basically immediately. After some effort I managed to get it to compile and install on solaris.
In python as root when I type "from pylab import *" I get this error:
=========================
ImportError: ld.so.1: python: fatal: relocation error: file /opt/csw/lib/python/site-packages/matplotlib/_path.so: symbol __1c2N6FI_pv_: referenced symbol not found
=========================
As my normal account I get this:
=========================
Traceback (most recent call last):
 File "test.py", line 1, in <module>
 from pylab import *
 File "/opt/csw/lib/python/site-packages/pylab.py", line 1, in <module>
 from matplotlib.pylab import *
 File "/opt/csw/lib/python/site-packages/matplotlib/__init__.py", line 677, in <module>
 rcParams = rc_params()
 File "/opt/csw/lib/python/site-packages/matplotlib/__init__.py", line 598, in rc_params
 fname = matplotlib_fname()
 File "/opt/csw/lib/python/site-packages/matplotlib/__init__.py", line 548, in matplotlib_fname
 fname = os.path.join(get_configdir(), 'matplotlibrc')
 File "/opt/csw/lib/python/site-packages/matplotlib/__init__.py", line 242, in wrapper
 ret = func(*args, **kwargs)
 File "/opt/csw/lib/python/site-packages/matplotlib/__init__.py", line 435, in _get_configdir
 raise RuntimeError("'%s' is not a writable dir; you must set %s/.matplotlib to be a writable dir. You can also set environment variable MPLCONFIGDIR to any writable directory where you want matplotlib data stored "% (h, h))
RuntimeError: '/local/host/home/mctran' is not a writable dir; you must set /local/host/home/mctran/.matplotlib to be a writable dir. You can also set environment variable MPLCONFIGDIR to any writable directory where you want matplotlib data stored
=========================
when I tried to set MPLCONFIGDIR to a directory, I got the earlier error. I tried recompiling and reinstalling to no avail.
I searched on the mailing list for anything like this, but there were no hits. Any ideas how I can fix this?

Dear List,
I would like to define a new second plot inside a first plot using the 
axes command.
But I need the position and size do be defined not by the relative 
figure coordinates
but by data coordinates.
I have found the following trick :
ax = gca()
f = gcf() 
x,y = ax.transData.transform([x,y])/f.transFigure.transform([1,1])
a = axes([x,y,dy,dy])
However, if the position is now ok, the size (dx,dy) is still on 
relative figure coordiate.
Ok, I could think a bit more and find how its possible to transform it, 
but I would
like to ask the list if there is an easier way to do that.
By the way, why do I have to divide by f.transFigure.transform([1,1]) ?
I expected that ax.transData.transform did return already normalized
values.
Thanks in advance.
yves
-- 
 (o o)
--------------------------------------------oOO--(_)--OOo-------
 Yves Revaz
 Laboratory of Astrophysics EPFL
 Observatoire de Sauverny Tel : ++ 41 22 379 24 28
 51. Ch. des Maillettes Fax : ++ 41 22 379 22 05
 1290 Sauverny e-mail : Yve...@ep...
 SWITZERLAND Web : http://www.lunix.ch/revaz/
----------------------------------------------------------------

> Thanks, but this wasn't quite what I had in mind. the exportfig m - 
> file trims the size of the white bounding box on the figure, in 
> addition to saving the image. Is there an easy way to do that with 
> matplotlib?
If you produce pdf, pdfcrop is neat.
Best, Martin
-- 
Martin Lüthi an...@tn...