matplotlib-users Mailing List for matplotlib

Hello all, I've been trying for days but I can't seem to get the
result I'm looking for. I have a 2d array of type "numpy.ndarray"
which I'd like to plot as a simple color map.  I'd like to plot it in
the upper-lefthand corner of the client area in a wxPython frame. The
plotting needs to be a very simple 1:1 ratio, for example if the numpy
array has 400 rows and 500 columns, I would like to plot it so that it
assumes 400x500 pixels in the wxPython frame. I do not need axis ticks
and labels, just the colormap plot itself. I can get my figure to plot
(with tick marks and labels since I haven't figured out how to turn
those off) but I cannot size it properly. I've copied a tutorial
example I found and modify it and through tedious trial and error have
gotten half-way to where I need:


# First attempt to render data to a window:

import matplotlib
matplotlib.use('WXAgg')

from matplotlib import rcParams
import numpy

import matplotlib.cm as cm

from matplotlib.backends.backend_wxagg import FigureCanvasWxAgg


#from matplotlib.figure import Figure

from wx import *

import DataFileTypes as DFT


class DataFrame(Frame):

    def __init__(self):

        Frame.__init__(self, None, -1, "Data filename here",
                       size=DisplaySize())

    def displayData(self):

        data = None

        # Load data into "data" object using my custom IntData(...) class:
        try:
            data = DFT.INTData("C:\SAR Test files\Tibet2008.int")

        except DFT.DataFileError:
            print("Error opening data file")

        except DFT.ResourceFileError:
            print("Error opening resource file")


        if data:

            # Assume a screen dpi of 96...seems very flakey to me:
            ScreenDPI = 96.0


            # compute the width and height of figure using this dpi
            # and the rows and columns of the data for a 1:1 display ratio:
            FigureWidthInInches  = (data.numcolumns / ScreenDPI)
            FigureHeightInInches = (data.numrows / ScreenDPI)
            print(FigureWidthInInches, FigureHeightInInches)

            # Instantiate Figure based on these parameters:
            self.fig =
matplotlib.figure.Figur((FigureWidthInInches,FigureHeightInInches),
dpi = ScreenDPI)
            self.canvas = FigureCanvasWxAgg(self, -1, self.fig)

            # Put everything into a sizer:
            sizer = BoxSizer(VERTICAL)
            #sizer.Add(self.canvas, 1, LEFT | TOP | GROW)
            sizer.Add(self.canvas, 0, LEFT | TOP)
            self.SetSizer(sizer)
            # self.Fit()

            a = self.fig.add_axes([0.075, 0.1, 0.75, 0.85])
            self.im = a.imshow(data.getNumpyArray(),
interpolation=None, cmap = data.getCustomColorMap())




if __name__ == '__main__':

    app = PySimpleApp()
    frame = DataFrame()
    frame.displayData()

    frame.Show()
    app.MainLoop()


It displays but the plot is inside the figure i.e. the colormap of the
data is within the figure that I've sized. matplotlib does this by
design, of course, but I cannot figure out how to defeat it. For one
thing, I don't think I'm sizing the figure correctly by setting
(guessing at) the dpi and computing the inches...just seems wrong, but
I can't find any tutorials or examples that show anything that sizes
figures using pixels or screen coords.

I always know the dimensions of my data a priori, so let's assume the
following very simple situation:

- I have a numpy.ndarray of data with 350 rows and 500 columns. How do
I display it in the upper-left hand corner of the frame client with no
tick marks/labels, etc...just the colormap at screen
coords(0,0)->(349,499)  (rows,columns)? Could someone post a few lines
to do this? Thanks so much in advance!

-L

If you have any intention of continuing to install updated versions of 
matplotlib as they come along, then please start now to eliminate the 
use of the matplotlib.numerix subpackage, if you have not already done 
so.  It has not been used internally in matplotlib, or made available 
via pylab, for well over a year.

In the next release of matplotlib based on the svn trunk, importing 
numerix will trigger a warning.  In some later release, numerix will be 
removed entirely.

Note that the sole present function of numerix is essentially to make 
numpy behave like the old Numeric.

Eric

Hi,

When I combine mathtext with the stixsans fonts while mathtext.default is
set to 'regular', horizontal text is not vertically aligned properly, like
the baseline is moving.  I've attached an image of what I see when I run the
following code.  It should be noted that I don't see anything like this on
the vertical axis, and it all goes away if I stop using mathtext.

import matplotlib.pyplot as plt
from matplotlib import rcParams
rcParams['font.size'] = 12.0
rcParams['mathtext.default'] = 'regular'
rcParams['mathtext.fontset'] = 'stixsans'

fig = plt.figure()

plt.xlabel('ItLooksToMeLikeTheBaselineMoves $(g m^{-3})$')
plt.ylabel('ItLooksToMeLikeTheBaselineMoves $(db km^{-1})$')
plt.show()


Thoughts? Am I doing something bad?  Can anyone else reproduce this?

Ryan

-- 
Ryan May
Graduate Research Assistant
School of Meteorology
University of Oklahoma
Sent from: Norman Oklahoma United States.

Here is some additional  information regarding this problem

OS -- Linux 2.6.22-16-generic #1 SMP Mon Jan 26 00:07:52 GMT 2009 i686
GNU/Linux

Moved matplotlibrc to  "os.environ['HOME'] =  '/tmp/matlibconfig'" because
of write permission problems with apache child processes. 

No customizations to matplotlibrc

Ubuntu Gutsy

Python-matplotlib     	0.90.1-2ubuntu1
Python-matplotlib-data 	0.90.1-2ubuntu1
Python-matplotlib-doc 	0.90.1-2ubuntu1


Code 
# This code is called by another python script
# and I am not even using the parameters that 
# are passed
# Code works from python idle

import sys, math
import matplotlib_config_path # this sets configpath as described above

import matplotlib 
matplotlib.use('Agg') # use non gui backend

from matplotlib import pylab as plt
import numpy as np
import os,sys 

def VT(batt_v, batt_temp0, batt_temp1, batt_ts, plotfile, dur):  

    t =[]  
    
    for x in batt_ts:    # time axis to 0 based
        t.append((x - batt_ts[0])/divsor[dur]) 

    fig = plt.figure()
    ax1 = fig.add_subplot(111)   # subplot 1 col x 1 row           

    ax1.plot([1,2,3])
    fig.savefig('test.png')

Error returned is the same


rlp_GMC wrote:
> 
> When I try to save a fig to a file  I get the error 
> 
> TclError Couldn't Connect to display ":0.0"  
> 
> I looked in the Maplotlib users manual  (latest version pg 144-146)  and
> tried the following with no luck. BTW why is the same module named
> differently Linux/Win ?
> 
> import matplotlib
> matplotlib.use('Agg')
> import matplotlib.pylab as plt (linux )   or,  import matplotlib.pyplot as
> plt   (Windows)
> fig = plt.figure()
> ax = fig.add_subplot(111)
> ax.plot([1,2,3])
> fig.savefig('test.png')
> 
> Is Matplotlib trying to display a plot to the monitor? 
> 
> rlp
> 


-- 
View this message in context: http://www.nabble.com/TclError-Couldn%27t-Connect-to-display-%22%3A0.0%22-tp22176192p22207062.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

I think the new version of matplotlib corrected the bug.

I have the problem with version 0.91.2, and there is no problem with the
version 0.98.3.


david.froger wrote:
> 
> Hi, I have the same problem, I have joined examples.
> 
> 
> 
>  http://www.nabble.com/file/p22180990/example.py example.py 
> http://www.nabble.com/file/p22180990/example.pdf example.pdf 
> http://www.nabble.com/file/p22180990/example.pdf example.pdf 
> 
> Zack 24 wrote:
>> 
>> Hi Richard, what exactly wrong here?
>> I didn't found anything...
>> 
>> On 4 May 2007, Richard Vernhes wrote:
>>> Here is attached the eps file with misaligned labels and grid.
>>> Richard
>> 
>> 
>> 
>> -- 
>> Zack
>> 
>> -------------------------------------------------------------------------
>> This SF.net email is sponsored by DB2 Express
>> Download DB2 Express C - the FREE version of DB2 express and take
>> control of your XML. No limits. Just data. Click to get it now.
>> http://sourceforge.net/powerbar/db2/
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>> 
>> 
> 
> 

-- 
View this message in context: http://www.nabble.com/Problem-with-the-position-of-tick-labels-in-postscript-file-tp10324639p22197268.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

As I mentioned in a previous email last night, I have built an
installer for matplotlib (updated this afternoon) off the SVN trunk.
You are welcome to use it, however be aware that maplotlib does not
support Python 2.6 (and won't until after Numpy officially supports
Python 2.6) and there are some known issues - one of which is that I
can't get Tkinter to work properly when using plot.show().  As a work
around I strongly suggest installing WxPython and using that as the
default backend.  If you do chose to use this installer please know
that it is unsupported and you are on your own if something breaks.

http://code.patricktmarsh.com/builds

-Patrick



On Tue, Feb 24, 2009 at 7:22 PM, Philip Bloom <pb...@cr...> wrote:
> What is the recommended way to install matplotlib for python 2.6?
>
> I have numpy installed currently on my python 2.6 distribution.  Is there an
> existing binary installer for matplotlib?  Is the best thing to do to try
> and compile from the current .98.5 source?
>
> Thanks for any help or advice, it is much appreciated.
>
> ------------------------------------------------------------------------------
> Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
> -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
> -Strategies to boost innovation and cut costs with open source participation
> -Receive a $600 discount off the registration fee with the source code: SFAD
> http://p.sf.net/sfu/XcvMzF8H
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>



-- 
Patrick Marsh
Graduate Research Assistant
School of Meteorology
University of Oklahoma
http://www.patricktmarsh.com

Can't you simply use "text"?

text(0.5, 0.5, "s", va="center", ha="center")

Unfortunately, no TextCollection yet.

-JJ


On Tue, Feb 24, 2009 at 4:49 PM, Fred Mailhot
<fma...@co...> wrote:
> Hi,
>
> I've looked through the documentation, and see that it's possible to
> plot a variety of symbols, including some customization, but what I'd
> really like to do is use a text character as the point marker. The
> only way I can think of doing this right now is to annotate the point
> (with zero offset) with the text in question, and somehow hide the
> actual point itself (is that even possible?).
>
> Is there something easier/more obvious that I'm missing?
>
> Thanks!
> Fred.
>
> ------------------------------------------------------------------------------
> Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
> -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
> -Strategies to boost innovation and cut costs with open source participation
> -Receive a $600 discount off the registration fee with the source code: SFAD
> http://p.sf.net/sfu/XcvMzF8H
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

What is the recommended way to install matplotlib for python 2.6?

I have numpy installed currently on my python 2.6 distribution.  Is
there an existing binary installer for matplotlib?  Is the best thing to
do to try and compile from the current .98.5 source?

Thanks for any help or advice, it is much appreciated.