matplotlib-users Mailing List for matplotlib

I use the following on Debian Sarge linux (both x86_64 and i686, 
although I hope/think it would work on any architecture and hopefully 
any linux distro):

"ffmpeg -hq -b 8000 -f mpeg2video -r 30 -i frame%03d.png movie.mpeg"

You may now skip the rest of this email, which is an uncalled-for 
venting-of-frustration.

Unfortunately, it seems bazillions of software companies think they can 
make bazillions of dollars by releasing yet another codec encumbered by 
more-or-less (but usually more) restrictive licensing conditions, 
leaving only lowest-common denominator codecs available for those of us 
who prefer to work without such restrictions. I sought long and hard to 
do something better than the above, but I can say this about the above 
command:

 * it produces movies that play in Windows (including PowerPoint, which 
isn't a given, even if it plays in Windows Media Player -- c'mon 
Microsoft, this is 2006, we should be able to play movies in our 
presentations), Mac OS X, and linux (Debian sarge, amd64 and i386, at least)
 * it seems to work with a standard Debian setup, and doesn't require 
using DLLs imported from some Windows system to to the encoding
 * it seems very fragile -- changing the frame rate or the codec usually 
breaks one of the above points
 * mpeg2 is probably also burdened by some licensing restrictions which 
I'm unaware of
 * on some Windows boxes/programs, white backgrounds get displayed as 
gray for some reason

Wishing that in 2006 we as a human race could come up with a better, 
open video format, but willing to accept even minor improvements to the 
above script,
Andrew

Alan G Isaac wrote:

>How are you turning your PNGs into an animation?
>
>Thanks,
>Alan Isaac
>
>PS I know about
>http://www.pymolwiki.org/index.php/Category%3ASoftware_Movies
>  
>

hu...@ya... wrote:
> Hi,
> 
> just a small question about histogram. I saw that the result of the hist 
> function from pylab and histogram from numpy+scipy can be slightly different 
> when the array is big and with real data (not integer). I'll probably told 
> something stupid but perhaps that will be good to have consistancies between 
> both function, won't it?

There are lots of different, equally valid ways to construct a histogram.
pylab.hist() and scipy.stats.histogram() probably use different algorithms. It's
probably not worth changing one just to match the other. Much better would be to
provide a broader interface to let the user twiddle the various knobs he would
like to twiddle. I believe David Huard posted an improved histogram class that
implements a number of useful features.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco

Here a sample:

the data are in the file data.dat join. 

In [1]: import pylab
In [2]: import scipy
In [3]: import scipy.stats

In [4]: data1,data2=pylab.load('data.dat',unpack=True)
In [5]: pylab.hist(data1,20)
(Out[5]:
array([  4,   6,  23,  52,  90, 128, 184, 244, 283, 293, 297, 330, 321,
       231, 188, 140,  94,  48,  29,  15]),
 array([ 0.00998046,  0.01054459,  0.01110872,  0.01167285,  0.01223698,
        0.01280111,  0.01336524,  0.01392937,  0.0144935 ,  0.01505763,
        0.01562176,  0.01618589,  0.01675002,  0.01731415,  0.01787828,
        0.01844241,  0.01900654,  0.01957067,  0.0201348 ,  0.02069894]),
 <a list of 20 Patch objects>)

In [6]: scipy.stats.histogram(data1,20)
Out[6]:
(array([  1,   7,  17,  43,  75, 126, 185, 248, 303, 302, 314, 353, 315,
       241, 178, 145,  70,  51,  20,   6]),
 0.0096835454084847374,
 0.00059382155039052636,
 0)

>     humufr> Hi, just a small question about histogram. I saw that the
>     humufr> result of the hist function from pylab and histogram from
>     humufr> numpy+scipy can be slightly different when the array is
>     humufr> big and with real data (not integer). I'll probably told
>     humufr> something stupid but perhaps that will be good to have
>     humufr> consistancies between both function, won't it?
>
> Complete example, please...
>
> JDH
>
>
> -------------------------------------------------------
> All the advantages of Linux Managed Hosting--Without the Cost and Risk!
> Fully trained technicians. The highest number of Red Hat certifications in
> the hosting industry. Fanatical Support. Click to learn more
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=107521&bid=248729&dat=121642
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Allan Noriel Estrella wrote:
> can you give the url of wxmpl and Mplot, i had a hard time googling for 
> them

That's rare, but I had trouble too, some of the google links are dead, 
but I found this for wxmpl:

http://agni.phys.iit.edu/~kmcivor/wxmpl/

and this (just the files) for Mplot:

http://cars9.uchicago.edu/~newville/Python/MPlot/

If you search this list, you'll see discussion of the differences 
between them, but I'll try this short synopsis (Kevin and Matt, please 
correct as needed)

wxmpl is designed to make it a bit easier to embed matplotlib in a a 
wxPython app. Once you've got a PlotPanel (or PlotFrame), you do your 
plotting with the matplotlib OO API (and some of pylab)

Mplot is designed to provide a plotting window that the end-user of the 
application can interact with, doing at least some manipulation of the 
plot. It provides and alternative API for plotting that is easy to use, 
but not feature-complete.

Note that I haven't Mplot, so I may have that a bit wrong.

-Chris



-- 
Christopher Barker, Ph.D.
Oceanographer
                                     		
NOAA/OR&R/HAZMAT         (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chr...@no...

>>>>> "humufr" == humufr  <hu...@ya...> writes:

    humufr> Hi, just a small question about histogram. I saw that the
    humufr> result of the hist function from pylab and histogram from
    humufr> numpy+scipy can be slightly different when the array is
    humufr> big and with real data (not integer). I'll probably told
    humufr> something stupid but perhaps that will be good to have
    humufr> consistancies between both function, won't it?

Complete example, please...

JDH

Hi,

just a small question about histogram. I saw that the result of the hist 
function from pylab and histogram from numpy+scipy can be slightly different 
when the array is big and with real data (not integer). I'll probably told 
something stupid but perhaps that will be good to have consistancies between 
both function, won't it?

N.

How are you turning your PNGs into an animation?

Thanks,
Alan Isaac

PS I know about
http://www.pymolwiki.org/index.php/Category%3ASoftware_Movies

Bryan Cole wrote:

> To accomplish similar goals, I embed a matplotlib panel (including
> toolbar) in a wxPython application 

If you do this, be sure to take a look at wxmpl and/or Mplot , they make 
it a bit easier to embed MPL in wxPython.

-Chris


-- 
Christopher Barker, Ph.D.
Oceanographer
                                     		
NOAA/OR&R/HAZMAT         (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chr...@no...

On Fri, 26 May 2006, Giandomenico Sica apparently wrote: 
> Insert a number: 4 
> Couples: [[4, 3], [4, 2], [4, 1], [3, 2], [3, 1], [2, 1]] 
> What I'd like to do is drawing with Matplotlib a graph having as nodes 
> 1, 2, 3, 4 and as edges among the nodes the above couples. 

n = input("Insert a number" )
a = reversed(range(1,n+1))
coppie = [[i,j] for i in a for j in range(1,i)]
print coppie
G = NX.Graph()
for edge in coppie:
	G.add_edge(*edge)
print G.edges()
NX.draw(G)
pylab.savefig(r'c:\temp\temp.eps')

hth,
Alan Isaac

Hello,

I'm new in Matplotlib and am writing in order to ask a help.
I've developed a little software with Python that, by inserting a
number, calculate a list of couples.

Example

Insert a number: 4
Couples: [[4, 3], [4, 2], [4, 1], [3, 2], [3, 1], [2, 1]]

What I'd like to do is drawing with Matplotlib a graph having as nodes
1, 2, 3, 4 and as edges among the nodes the above couples. (sample
drawing: https://networkx.lanl.gov/)

Can you help me to make this operation please?
My code at the present is the following one.

n = input("Insert a number" )
a = range(n)
p = 0
q = 1
coppie = []
coppia = []
while len(a) - p != 1:
	coppia.append(len(a) - p)
	coppia.append(len(a) - p - q)
	coppie.append(coppia)
	coppia = []
	if len(a) - p - q != 0:
		q = q + 1
	if len(a) - p - q == 0:
		p = p + 1
		q = 1
print coppie

I suppose it is quite easy but in spite of this have some difficulty.
Really many thanks for the help!

Best,
Nico

I am having some problems with the axes frame being clipped.


Consider the following:

from pylab import *
fill([0,1,1,0],[0,0,1,1],facecolor=(1,1,1),edgecolor=(1,1,1))
axis([-1,1,-1,1])
gca().get_frame().set_linewidth(20)
show()

I have made the frame thick to emphasize the problem: When I plot
things, they overwrite part of the frame which looks bad.  What is the
simplest/best way to avoid this?

I would like the axes frame to be drawn only outside of the actual
plot area, but I cannot figure out how to achieve this.  (It seems
like this might also be desirable default behaviour...).  Otherwise,
I would settle for the frame being drawn over top of the plot so that
it is smooth.

A crude hack would be to plot a blank rectangle.  It should also be
possible to use clipping somehow, but I do not have much experience
with that.

Is there a simple solution?

Thanks for any help,
Michael