matplotlib-users Mailing List for matplotlib (Page 8)

>>>>> "Dominique" == Dominique Orban <Dom...@po...> writes:

    Dominique> When using imshow(), why does there always seem to be a
    Dominique> blank zone along the southern and eastern edges of the
    Dominique> figure? For instance:

    Dominique> X = rand(10,10) imshow(X)

    Dominique> plots a luminance image of X, which seems fine, except
    Dominique> for the lower and rightmost edges, which are blank. I
    Dominique> may be misunderstanding the purpose of imshow, but
    Dominique> skimming through the code didn't give me an answer. I
    Dominique> am using matplotlib 0.52 on WinXP with either GTKAgg or
    Dominique> TkAgg.

Hi Dominique,

Your example did point me to a small bug in the image module, but it
is mostly unrelated to what you are observing.  In the axes.py
function imshow, replace

            self.set_image_extent(0, numcols-1, 0, numrows-1) 
with 
            self.set_image_extent(0, numcols, 0, numrows) 


This only affects the tick labeling (not the actual image display)
but it was wrong before and should be changed.

Now run this script

    from matplotlib.matlab import *
    X = rand(10,10)

    subplot(211)
    im = imshow(X)
    im.set_interpolation('nearest')

    subplot(212)
    im = imshow(X)
    show()

The key thing is that the white border you are seeing arises from
interpolation.  The points on the bottom and right have no neighbors
in those directions, and so they interpolate to the background color,
which is white.

You can set the axis limits so that these regions don't appear, or use
nearest neighbor interpolation.

Let me know if these suggestions don't work for you.

JDH

Hi john ,

I was doing a pure TeX plot (a bunch of equations inside a box)  and I 
noticed that the \sqrt{}command does not work even though it listed in 
the help page for mathtext.

\frac and \dfrac would be a nice addition too...

feel free to used this little script as an example of another use of 
mathtext...

Flavio

Hi everyone:

I was wondering whether it is possible to tell matplotlib how/when to 
connect data points. Consider this simple script:

from matplotlib.matlab import *
 
figure(1)
t = [0,1,2,3,4,5,105,106,107]
s = [1,4,5,3,9,11,-5,-8,3]
plot(t, s, antialiased=False)
 
grid(True)
show()


There are no data points between t=5 and t=105. By default the points 
(5,11) and (105,-5) are connected, but I would like to tell matplotlib 
NOT to do so. In my case I would like to pass the plot function a 
variable telling it what to do. So for example would have:

plot(t, s, max_delta=40)

This would mean that the points are only to be connected if the 
difference between the adjacent t values is less than 40.  

In my case this is relevant because sometimes there are "holes" in my 
data, and connecting the points makes the plots look very messy.

Would anyone find something like this useful? Would it be difficult to 
implement?

Thanks.
Peter

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

    Humufr>        Hello, I have a problem to obtain a good ps file
    Humufr> (with mathematical symbol) with matplotlib. That's work
    Humufr> very good with a png file but not with a postscript.

    Humufr> I tried to define the TTFPATH and AFMPATH but that change
    Humufr> nothing. I think that I miss one configuration somewhere
    Humufr> but I can't arrive to find what...

    Humufr> Thank you for this soft, I just begin to use but it's
    Humufr> seems very interesting

I haven't yet added mathtext to the postscript backend, though I will
as soon as I get the time.  See the mathtext documentation at
http://matplotlib.sf.net/matplotlib.mathtext.html for the latest
information about mathtext, which backends it works on, etc...

mathtext is designed around the BaKoMa truetype fonts.  There are also
postscript versions of these fonts.  Paul Barrett has been working on
unifying the interface between AFM (postscript) and truetype fonts,
but we're not there yet.  When this is done, it will be easier to
rewrite mathtext to work with either kind of font.  Another
possibility is to modify the postscript backend to use postscript
level 3, which supports truetype fonts.  This would be very nice
because it would improve compatibility between the postscript, image
and GUI backends.

Another thing that needs to be done is to add image support to the
postscript backend (draw_image).  This shouldn't be too hard.  The
matplotlib freetype module ft2font has a method to get the font raster
as a bitmap, which could be placed on the PS canvas with draw_image.
This wouldn't be an ideal solution because it's not scalable, but it
would be a fairly easy temporary solution.

Volunteers welcome!
JDH

       Hello,

I have a problem to obtain a good ps file (with mathematical symbol) 
with matplotlib. That's work very good with a png file but not with a 
postscript.

I tried to define the TTFPATH and AFMPATH but that change nothing. I 
think that I miss one configuration somewhere but I can't arrive to find 
what...

Thank you for this soft, I just begin to use but it's seems very 
interesting

    Nicolas

When using imshow(), why does there always seem to be a blank zone along 
the southern and eastern edges of the figure? For instance:

X = rand(10,10)
imshow(X)

plots a luminance image of X, which seems fine, except for the lower and 
rightmost edges, which are blank. I may be misunderstanding the purpose 
of imshow, but skimming through the code didn't give me an answer. I am 
using matplotlib 0.52 on WinXP with either GTKAgg or TkAgg.

Thanks !
Dominique

One thing I'm not sure of is just how polar plots are supposed to work. 
If you set the axis background color, what gets colored?  To get that
right (if "in the circle" is right), I guess you'd need a new Axes
class.

In gnuplot, everything is drawn in the rectangle.  "set polar; set grid;
plot sin(3*t)" gives you a rectangular grid.  You have to do "set grid
polar" to circles.  Of course gnuplot isn't what we're going for (How do
you change background color?  The only way I know is using xresources.)

I guess the answer should be what matlab does (unless we can think of
something better).  Now I've never used matlab, so this will take a
little research.  I borrowed a matlab graphics book and will take a
peek.

Keep up the good work.

Thanks,
Greg

PS  Say hi to hyde park for me.  I graduated in 98.

On Wed, 2004-03-31 at 07:54, John Hunter wrote:
> >>>>> "Greg" == Greg Whittier <gr...@th...> writes:
> 
>     Greg> As a temporary solution you might try just transforming your
>     Greg> r,theta data to x, y and then drawing a grid over it.
> 
> A nice workaround.  You may want to add 
> 
>   axis('off')
> 
> To get rid of the background axes.
>  
>     Greg> I looked at the classes and with the loglog already done as
>     Greg> an example, it shouldn't be too hard to add polar plotting.
>     Greg> So far though I haven't got any further than printing out
>     Greg> some of the code.
> 
> I was planning on taking a different tack, and derive a PolarAxes from
> Axes and RadialAxis from Axis, etc, which uses circles rather then
> lines for the radial gridlines, and so on. But you get so close with
> so little code that your approach may be better.  If you want to keep
> forging ahead, I'm happy to leave the ball in your court.
> 
> On an unrelated note,
> 
>   from Numeric import *
>   from matplotlib.matlab import *
> 
> is redundant because matplotlib.matlab imports all of numeric/numarray
> as well as MLab, fft, and some stuff from LinearAlgebra and Matrix.
> This is to provide a matlab like environment where most of the things
> you need are there.
> 
> JDH
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
-- 
Greg Whittier <gr...@th...>

It works great now. Thanks for the quick response and
the fix John!

Also to answer Todd Miller:

>I googled around on this and got a suggestion having
>to do with mainloop
> handling from an old Guido e-mail.  IDLE has a -n
>switch ...

In my initial message note the IDLE info:
>IDLE 1.0      ==== No Subprocess ====

The 'No Subprocess' part indicates I'm using the -n
flag.  In IDLE 1.0, from my understanding, the IDE
normally restarts the interpreter each time a program
is run.  The -n flag does not do this and runs
everything together so the Tkinter mainloop does not
get confused.  It's not a great solution but it works
and I have not had serious stability problems, at
least for my limited needs.  This trick is needed for
the turtle module, and probably anything that uses
Tkinter.

And now back to trying matplotlib!

-- David 



>>>>> "John" == John Hunter <jdhunter@ac...> writes:
 
     John> Ahh the golden piece of information.  Todd
Miller has
     John> reported that idle does not seem to respect
matplotlibrc.
     John> At this point we have no idea why.  We'll
try to get this
     John> figured out ASAP!


 >>>> "John" == John Hunter <jdhunter@ac...> writes:

> I found one problem.  IDLE sets the HOME variable,
>and there is a bug
> in the way matplotlib located the rc file. 
>Basically, if HOME is set,
> it expects the rc file to be there.  Here is a fix.
 
 >Edit
C:\Python23\Lib\site->packages\matplotlib\__init__.py
and replace
> the function matplotlib_fname with the following
 
 def matplotlib_fname():
     'Return the path to the rc file'
     if os.environ.has_key('MATPLOTLIBRC'):
         path =  os.environ['MATPLOTLIBRC']
         if os.path.exists(path):
             fname = os.path.join(path,
'.matplotlibrc')
             if os.path.exists(fname):
                 return fname
 
     if os.environ.has_key('HOME'):
         path =  os.environ['HOME']
         if os.path.exists(path):
             fname = os.path.join(path,
'.matplotlibrc')
             if os.path.exists(fname):
                 return fname
 
     path =  get_data_path() # guaranteed to exist or
raise
     fname = os.path.join(path, '.matplotlibrc')
     if not os.path.exists(fname):
         print >> sys.stderr, 'Could not find
.matplotlibrc; using defaults'
     return fname
 
> I tested this on windows with idle.  Idle now
>respects matplotlibrc
> and loads TkAgg.  However, at least on my system,
>there is still some
> idle bug because the Tk window launches and then
>freezes without
> displaying the figure I suspect Todd, the TkAgg
>author, will be
> looking at this soon.  [Todd, I had this problem
>with or without
> window_focus set in my test].
 
 >Sorry for the troubles,
 >JDH


__________________________________
Do you Yahoo!?
Yahoo! Small Business $15K Web Design Giveaway 
http://promotions.yahoo.com/design_giveaway/

On Thu, 2004-04-01 at 08:55, John Hunter wrote:
> >>>>> "John" == John Hunter <jdh...@ac...> writes:
> I tested this on windows with idle.  Idle now respects matplotlibrc
> and loads TkAgg.  However, at least on my system, there is still some
> idle bug because the Tk window launches and then freezes without
> displaying the figure I suspect Todd, the TkAgg author, will be
> looking at this soon.  [Todd, I had this problem with or without
> window_focus set in my test].

I googled around on this and got a suggestion having to do with mainloop
handling from an old Guido e-mail.  IDLE has a -n switch which you can
add to the end of the windows shortcut command line (Target:) under the
shortcut properties.  When I did this,  matplotlib is unfrozen.  The
e-mail is a little ominous about using the switch:  I think a failure in
matplotlib could also bring down IDLE and any open windows it might have
lying around (like, say, a plot script!).  Guido also suggests that
running the matplotlib Tkinter mainloop in its own thread to avoid using
-n is a bad idea.  This is looking dicey.  We might want to make a note
that using IDLE with -n can be done but is considered risky.

Here's Guido:
http://aspn.activestate.com/ASPN/Mail/Message/edu-sig/1818398

Regards,
Todd

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

>>>>> "Steven" == Steven Pang <pk...@si...> writes:

    Steven> Hi, I need a library for my app to produce .png charts and
    Steven> is currently trying matplotlib. I would need the whole app
    Steven> to be compiled to .exe using py2exe. I wonder if
    Steven> matplotlib is suitable?

Oops, I got confused by all the "Steve" posts this morning and
accidentally responded to Steve Chaplin trying to answer your
question.  Here is the same post again:

Agg is already built into the matplotlib windows installer.  No need
to build it yourself.  If for some reason, however, you want or need
to build it yourself, download
http://matplotlib.sourceforge.net/win32_static.tar.gz, untar it in
your matplotlib src dir, and follow the instructions in the README in
the win32_static dir and in setupext.py.

As for building matplotlib + agg into a py2exe installer, yes this is
possible.  See Matt Fischler's post on this list "How to freeze your
matplotlib apps using py2exe"

  http://sourceforge.net/mailarchive/message.php?msg_id=7467225

Hope this helps,
JDH

>>>>> "John" == John Hunter <jdh...@ac...> writes:

    John> Ahh the golden piece of information.  Todd Miller has
    John> reported that idle does not seem to respect matplotlibrc.
    John> At this point we have no idea why.  We'll try to get this
    John> figured out ASAP!

I found one problem.  IDLE sets the HOME variable, and there is a bug
in the way matplotlib located the rc file.  Basically, if HOME is set,
it expects the rc file to be there.  Here is a fix.

Edit C:\Python23\Lib\site-packages\matplotlib\__init__.py and replace
the function matplotlib_fname with the following

def matplotlib_fname():
    'Return the path to the rc file'
    if os.environ.has_key('MATPLOTLIBRC'):
        path =  os.environ['MATPLOTLIBRC']
        if os.path.exists(path):
            fname = os.path.join(path, '.matplotlibrc')
            if os.path.exists(fname):
                return fname

    if os.environ.has_key('HOME'):
        path =  os.environ['HOME']
        if os.path.exists(path):
            fname = os.path.join(path, '.matplotlibrc')
            if os.path.exists(fname):
                return fname

    path =  get_data_path() # guaranteed to exist or raise
    fname = os.path.join(path, '.matplotlibrc')
    if not os.path.exists(fname):
        print >> sys.stderr, 'Could not find .matplotlibrc; using defaults'
    return fname

I tested this on windows with idle.  Idle now respects matplotlibrc
and loads TkAgg.  However, at least on my system, there is still some
idle bug because the Tk window launches and then freezes without
displaying the figure I suspect Todd, the TkAgg author, will be
looking at this soon.  [Todd, I had this problem with or without
window_focus set in my test].

Sorry for the troubles,
JDH

>>>>> "Steve" == Steve Chaplin <ste...@ya...> writes:

    Steve> To answer my own question. Yes, AGG is tested on Windows,
    Steve> Linux, Sun, SGI, and Apple (MacOS X, X11)

Hi Steve, 

Agg is already built into the matplotlib windows installer.  No need
to build it yourself.  If for some reason, however, you want or need
to build it yourself, download
http://matplotlib.sourceforge.net/win32_static.tar.gz, untar it in
your matplotlib src dir, and follow the instructions in the README in
the win32_static dir and in setupext.py.

As for linux, OS X, and unix, the matplotlib src distribution comes
with agg.  No need to download it separately.  Ditto for CVS checkouts
(this is a recent change).

As for building matplotlib + agg into a py2exe installer, yes this is
possible.  See Matt Fischler's post on this list "How to freeze your
matplotlib apps using py2exe"

  http://sourceforge.net/mailarchive/message.php?msg_id=7467225

Hope this helps,
JDH

To answer my own question. Yes, AGG is tested on Windows, Linux, Sun,
SGI, and Apple (MacOS X, X11)

The website is a little out of date. Here's some info the author sent
me:

=============================================================
AGG has been considerably modified last time and its development is
still in
process, but major design issues are stabilized now.
The good news is that there many new things were implemented, such as:

- Gradients and Gouraud Shading
- Image affine transformations
- Strokes with different types of line joins and line caps
- Dashed line generator
- Markers, such as arrowheads/arrowtails
- Fast vectorial polygon clipping to a rectangle
- Low-level clipping to multiple rectangular regions
- Alpha-Masking
- New fast anti-aliased line algorithm
- Using arbitrary images as line patterns
- Rendering in separate color channels
- Perspective and bilinear transformations of vector and image data
- General polygon clipping (and, or, xor, sub) based on Alan Murta's GPC

The library is platform independent and it's considerably redesigned.
You can find many examples, including image transformation ones in
agg2/examples. The examples are platform independent too and you
can find the building environments in the respective directories.
If you use Win32 platform with VC++ 6 or later just open
the projects, the library doesn't need any preliminary building. If you
use
Unix/Linux with X11 and gcc, first execute make in the agg2 directory
and then make in examples/X11/.

There is also a simple SVG viewer, it's in agg2/svg. You will
need Expat XML parser.

The latest snapshot of AGG is always here:
http://www.antigrain.com/agg2.zip (MS-DOS end-of-line style)
http://www.antigrain.com/agg2.tar.gz (Unix end-of-line style)
or in the CVS repository at SourceForge:
http://sourceforge.net/cvs/?group_id=42020

Also there's a lite, minimalistic version:
http://www.antigrain.com/lite/agg2_lite.zip

There is a mailing list:
http://lists.sourceforge.net/lists/listinfo/vector-agg-general
where I publish news from time to time and where some very nice people
are
discussing AGG issues. They are giving me a lot of recommendations and
new
ideas.

I'd recommend you to subscribe and take part in it. Look at the mailing
archives for details:
http://sourceforge.net/mailarchive/forum.php?forum=vector-agg-general
====================================================

> Hi Steve, thanks for letting me know.  There is an off-by-one error
> and it looks like and easy fix.  For future reference, you may want to
> consider using GTKAgg as your default GUI.  This has the GTK widget
> but uses agg for rendering.  Agg supports subpixel rendering and so
> isn't susceptible to one pixel errors that crop up in GTK in a number
> of contexts.  At low resolutions, these become particularly
> noticeable.  Other benefits over the GTK backend are alpha blending,
> anti-aliased drawing, and faster image support -
> http://matplotlib.sourceforge.net/backends.html#GTKAgg

I had a look at http://www.antigrain.com/ and downloaded
agg-2002-05-21.zip.
There's no install document, configure.ac file or Makefile that would
allow me to compile the software. It does not look like this is portable
software that will compile on Linux.

Regards,
Steve

Hi,

I need a library for my app to produce .png charts and is
currently trying matplotlib. I would need the whole app
to be compiled to .exe using py2exe. I wonder if
matplotlib is suitable?

I come across some instructions from John on how
to do this using Agg without a GUI:
Download the src distribution, edit setup.py and
set BUILD_AGG = True, do a normal setup.py install.

However, I got this error on my Win98:
running build_ext
error: Python was built with version 6 of Visual Studio, and extensions 
need to
be built with the same version of the compiler, but it isn't installed.

Thanks for any help,
Steven.

> Now on to your problem.
> 
> In backend_gtk draw_rectangle, change
> 
>         x, y = int(x), self.height-int(math.ceil(y+height))
> 
> to
>         x, y = int(x), self.height-int(y+height)
> 
> and the GraphicsContext.set_clip_rectangle method to
> 
>     def set_clip_rectangle(self, rectangle):
>         GraphicsContextBase.set_clip_rectangle(self, rectangle)
>         l,b,w,h = rectangle
>         rectangle = (int(l), self.renderer.height-int(b+h)+1,
>                      int(w), int(h))
>         self.gdkGC.set_clip_rectangle(rectangle)        

I tried this but it made things worse. With simple_plot.py the plot now
extends above the top axis line.

Steve

>>>>> "David" == David Brown <db...@ya...> writes:

    David> I installed the matplotlib-1.52.win32-py2.3 binary on
    David> python23 (Enthought).  I don't want to install a lot of
    David> extra packages so I found the existing .matplotlibrc in
    David> C:\Python23\share\matplotlib and modified the backend to
    David> "backend : TkAgg".  But when I try to import matplotlib it
    David> appears to be still looking for the default Gtk backend.

    David> I noticed that sys.path did not contain
    David> C:\Python23\share\matplotlib so I tried adding it.  Still
    David> no luck.

It has nothing to do with sys.path, so you can safely ignore this.

    David> Here's the error message:

    David> IDLE 1.0 ==== No Subprocess ====
            ^^^

Ahh the golden piece of information.  Todd Miller has reported that
idle does not seem to respect matplotlibrc.  At this point we have no
idea why.  We'll try to get this figured out ASAP!

As a test case, try to run a matplotlib script from the DOS prompt

C:> python somefile.py

and see if TkAgg launches.  If so, we can be fairly sure it's idle.
Probably just double clicking on one of the examples from the zip
distribution will be enough, but you lose the traceback this way if
there is an error.

Let me know,
JDH

I installed the matplotlib-1.52.win32-py2.3 binary on
python23 (Enthought).  I don't want to install a lot
of extra packages so I found the existing
.matplotlibrc in C:\Python23\share\matplotlib and
modified the backend to "backend      : TkAgg".  But
when I try to import matplotlib it appears to be still
looking for the default Gtk backend.  

I noticed that sys.path did not contain
C:\Python23\share\matplotlib so I tried adding it. 
Still no luck.

Here's the error message:

IDLE 1.0      ==== No Subprocess ====
>>> from matplotlib.matlab import *

No module named pygtk
matplotlib requires pygtk-1.99.16 or greater -- trying
anyway.  Please hold on
Traceback (most recent call last):
  File "<pyshell#0>", line 1, in ?
    from matplotlib.matlab import *
  File
"C:\Python23\Lib\site-packages\matplotlib\matlab.py",
line 127, in ?
    from backends import new_figure_manager,
error_msg, \
  File
"C:\Python23\Lib\site-packages\matplotlib\backends\__init__.py",
line 16, in ?
    from backend_gtk import  \
  File
"C:\Python23\Lib\site-packages\matplotlib\backends\backend_gtk.py",
line 13, in ?
    import gobject
ImportError: No module named gobject
>>> 


Any ideas on why this does not work?  Should my python
path contain the path to share\matplotlib ?
-- David

__________________________________
Do you Yahoo!?
Yahoo! Finance Tax Center - File online. File on time.
http://taxes.yahoo.com/filing.html