matplotlib-users Mailing List for matplotlib

On 2005年10月14日, Steve Schmerler apparently wrote: 
> 1.) yes, the plots are fine 
> 2.) yes, the bounding box is large enough 
> I attached one of the .eps files I was talking about. In the printed 
> .pdf where this is included the xlabel "T/K" as well as the numbers on 
> the x-axis are missing. The "x"s in the legend box are also missing so 
> it isn't a bb problem. 
> I used dvipdf to convert .dvi -> .pdf (which I've learned only uses 
> dvips and gs with some pdfwrite options) and not dvipdfm. Is it 
> recommended to use dvipdfm over dvipdf? 
Works fine with dvipdfm.
(Ticks and labels are visible.)
Cheers,
Alan Isaac
(using Win2000, MiKTeX)

Hi,
here is the script:
#!/usr/bin/python
from scipy import *
from pylab import *
MAX_TIME =3D 7000
fn_social=3D"all-avg.dat"
def readStats(fn):
	q =3D [d for d in io.read_array(fn) if d[0]<=3DMAX_TIME]
	print q
	t=3D[s[0] for s in q]
	age=3D[s[1] for s in q]
	age_conf=3D[s[2] for s in q]
	wp=3D[s[3] for s in q]
	wp_conf=3D[s[4] for s in q]
	return t, age, age_conf, wp, wp_conf
hold(True)
t, age, age_conf, wp, wp_conf =3D readStats(fn_social)
xlabel('time [s]')
ylabel('average lifetime [s]')
errorbar(t,wp,yerr=3Dwp_conf, color=3D"blue",ecolor=3D"black", mfc=3D'red',=
=20
mec=3D'green', ms=3D200, mew=3D4)
show()
Am Freitag, 14. Oktober 2005 16:05 schrieb John Hunter:
> >>>>> "Willi" =3D=3D Willi Richert <ri...@c-...> writes:
>
> Willi> Thanks, that helped! Switching to QtAgg did the trick.
> Willi> However, I have still some issues: 1) the text placement of
> Willi> the titles does not really work:
> Willi> http://wr.objectis.net/Members/Papst/Matplotlib_Problems
>
> Willi> 2) The confidence intervals are only vertical lines. I
> Willi> would like to have some small horizontal "stoppers" at the
> Willi> upper and lower point of those error bars. Is that
> Willi> possible with matplotlib?
>
>
> Could you post the script -- hard to debug in a vacuum.
>
> JDH
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Power Architecture Resource Center: Free content, downloads, discussions,
> and more. http://solutions.newsforge.com/ibmarch.tmpl
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
=2D-=20
Dipl.-Inform. Willi Richert
 C-LAB - Cooperative Computing & Communication Laboratory
 der Universit=E4t Paderborn und Siemens
=46U.323
 F=FCrstenallee 11
 D-33102 Paderborn
 Tel: +49 52 51 60 - 61 20
 Fax: +49 52 51 60 - 60 65
 E-Mail: ri...@c-...
 Internet: http://www.c-lab.de

I am interested in creating graphs where I can interact with the data
plotted in the graph.
 
For example, suppose I plot the heat of formation of ~ 80 different alloy
structures (which each have a name) against the composition of each
structure. What I would like is to hold the mouse over a data point and see
a balloon pop up that displays the crystal structure name, or to left click
on a data point to launch a structure viewer for the structure at that
point, or right click to get a menu of different options, or to press "f" to
see what filename the data for that point is saved in, etc...
 
Another example would be a graph of many lines, where each line corresponds
to a structure. Then I would want to click on the lines to do various
things. 
 
Sometimes the data is very dense, so it would also be nice if you could move
the focus to different points by tabbing or using the arrow buttons.
 
It would be interesting to be able to select portions of data points to do
some kind of analysis on, or to modify their appearance someway, maybe
change symbol the size according to atomic volume, or color according to
some property, etc. 
 
Could these kinds of things be done in/with matplotlib? Or more importantly,
does a framework in matplotlib exist that this kind of thing could be
developed? I am interested in talking to anyone who has thoughts about this.
 
Thanks,
 
j 

Hi Jeff,
This looks great but unfortunately I get the following error:
astraw@aspiring:~/other-peoples-src/natgrid/example$ python example.py
Traceback (most recent call last):
 File "example.py", line 3, in ?
 import nat
 File "/usr/lib/python2.3/site-packages/nat.py", line 362, in ?
 import string, math, sys, Numeric, cdms, MA, natgridmodule
ImportError: No module named cdms
So I modified nat.py in the following way, and it now works. And there
was much rejoicing!
-import string, math, sys, Numeric, cdms, MA, natgridmodule
+import string, math, sys, Numeric, MA, natgridmodule
Jeff Whitaker wrote:
> Jesper Larsen wrote:
>
>> Hi matplotlib users,
>>
>> I have a long list of ungridded data that I would like to make a
>> contour plot of. The data is simply a list of (longitude, latitude,
>> datavalue) with the data value belonging the given longitude and
>> latitude. As far as I understand contour() only accepts gridded data
>> values.
>>
>> The solution is probably to interpolate the unstructured data to a
>> regular grid and then plot the data. Has anyone tried doing that or
>> know where to look for an interpolation/triangulation routine?
>>
>> Cheers,
>> Jesper
>>
>>
>>
>
> Jesper: Since this question has come up a couple of times, I decided to
> cook up an example. First you'll need to download and install the
> natgrid python module (included in CDAT, but I've separated it out from
> the huge tarball and put it at
> ftp://ftp.cdc.noaa.gov/Public/jsw/natgrid.tar.gz). Then try this:
>
>
> from RandomArray import uniform
> import pylab as p
> import nat
>
> def griddata(x,y,z,xi,yi):
> r = nat.Natgrid(y, x, yi, xi)
> return r.rgrd(z)
>
> npts = 500
> x = uniform(-2,2,npts); y = uniform(-2,2,npts)
> z = x*p.exp(-x**2-y**2)
>
> # x, y, and z are now vectors containing nonuniformly sampled data.
> # Define a regular grid and grid data to it.
> nx = 51; ny = 41
> x1 = p.linspace(-2,2,nx)
> y1 = p.linspace(-2,2,ny)
> xi, yi = p.meshgrid(x1, y1)
> zi = griddata(x,y,z,x1,y1)
>
> # Contour the gridded data, plotting dots at the nonuniform data points.
> CS = p.contour(xi,yi,zi,15,linewidths=0.5,colors=['k'])
> CS = p.contourf(xi,yi,zi,15,cmap=p.cm.jet)
> p.scatter(x,y,marker='o',c='b',s=5)
> p.xlim(-2,2)
> p.ylim(-2,2)
> p.show()
>
>
>
> It's interesting to see what happens when you vary npts (from 50 to 1000).
>
> HTH,
>
> -Jeff
>