matplotlib-devel Mailing List for matplotlib (Page 3)

Quoting John Hunter <jdh...@ni...>:
> Perhaps Fernando or some other knowledgeable wx person can comment on
> the appropriate workaround if there is one.

Currently my workaround is to use Tk :(

Cheers,

f

What I was trying to say is that insofar as the SVG backend is a work in progress, the PyChart backend could be used an example to aid development. Just trying to be helpful :)

Chris

>>>>> "cfuller" == cfuller  <cf...@th...> writes:

    cfuller> A nice source for some SVG inspiration might be PyChart,
    cfuller> another python plotting system, that has a fully
    cfuller> functional SVG backend:
    cfuller> http://www.hpl.hp.com/personal/Yasushi_Saito/pychart.  I
    cfuller> like the looks/features of this package, but it lacks
    cfuller> interactivity/GUI support and relies on ghostscript for
    cfuller> rendering images.

Are you aware that matplotlib has an SVG backend already?  Jared
Wahlstrand submitted one back in late May, which I haven't announced
yet.  Jared what is the status on backend_svg

 - font support?

 - mathtext support?

 - image support?

JDH

>>>>> "cfuller" == cfuller  <cf...@th...> writes:

    cfuller> wx won't start up on my linux machine, running Fedora
    cfuller> Core 2 and a compiled-from-source wxPython 2.5.1. There
    cfuller> was some incompatability with the new GTK libraries,
    cfuller> since the wxPython rpm was comppiled on RH9. I'm guessing
    cfuller> the wx backend is having similar troubles, although I
    cfuller> compiled it from source, naturally. That's a long
    cfuller> compile, BTW!  The wx backend works for me at work, under
    cfuller> Server 2003 and the same version of wx. It still leaks,
    cfuller> although I upgraded to the .60b installation binary from
    cfuller> John's website. Compiling distutil packages under windows
    cfuller> is something I've yet to master, whether with the Borland
    cfuller> compiler or the (not so new) Visual C++ Tooolkit
    cfuller> 2003. Visual Studio just isn't something I can justify,
    cfuller> even if the lab would pay for it. Especially with the VC6
    cfuller> vs VC7 funny business. At least Tk stopped leaking!

Could you send me the script that is leaking?

    cfuller> Since I have a wx that works, its not so bad. I'll still
    cfuller> work on the FC2 issues, but I can see about that toolbar
    cfuller> addition as well. I should get around to installing
    cfuller> windows again at home, I've been putting off an upgrade
    cfuller> to XP, and my old win2k pro install is useless, after a
    cfuller> mobo upgrade.

I believe this is the bug Fernando Perez was writing about which is
specific to recent releases of wx that use private GTK symbols that
are no longer present in the gtk libs in Fedora core 2.  Here is a
snip from an email he sent me earlier on the subject.  As far as I
know, there is nothing we can do about it on the matplotlib side,
except perhaps check the wx/wxpython list to see if it has been fixed
yet in CVS, and agitate for a fix if not.  

Perhaps Fernando or some other knowledgeable wx person can comment on
the appropriate workaround if there is one.

From: Fernando Perez <Fer...@co...>
Subject: Re: matplotlib, ipython and other comments
To: John Hunter <jdh...@ni...>
Date: Wed, 09 Jun 2004 14:29:54 -0600
Organization: Applied Mathematics, University of Colorado at Boulder

...snip ...

> I'll look into this later.  My experience with WX and WXAgg is that
> both work under linux, but WX is a bit slow and buggy.  Is the problem
> you are describing Fedora specific?  (Sorry I can't easily read these
> links now since URL cut-and-paste from my xterm on OSX laptop to my
> browser window doesn't work).

No, the bug is in current WX.  Here's a traceback:

In [6]: import wx
---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)

/home/fperez/<console>

/usr/lib/python2.3/site-packages/wx/__init__.py
      43 __revision__ = "$Revision: 1.1.2.4 $"[11:-2]
      44
---> 45 from wxPython import wx
      46
      47 _newnames = {}

/usr/lib/python2.3/site-packages/wxPython/__init__.py
      18 # Ensure the main extension module is loaded, in case the add-on modules
      19 # (such as utils,) are used standalone.
---> 20 import wxc
      21
      22 
#----------------------------------------------------------------------------

ImportError: /usr/lib/libwx_gtk2-2.4.so.0: undefined symbol: 
_gtk_accel_group_detach

Apparently the WX guys chose to use _gtk_* symbols which the GTK documentation 
_explicitly warned_ were private and could go away at any time.  Now, in the 
version of GTK shipped with Fedora they _did_ go away, so Wx broke.  This just 
needs to be fixed by the Wx team (maybe it already is in CVS, I'm using Wx as 
shipped with Fedora).  So don't worry about this, it will get fixed in time by 
those responsible.

wx won't start up on my linux machine, running Fedora Core 2 and a compiled-from-source wxPython 2.5.1. There was some incompatability with the new GTK libraries, since the wxPython rpm was comppiled on RH9. I'm guessing the wx backend is having similar troubles, although I compiled it from source, naturally. That's a long compile, BTW!

The wx backend works for me at work, under Server 2003 and the same version of wx. It still leaks, although I upgraded to the .60b installation binary from John's website. Compiling distutil packages under windows is something I've yet to master, whether with the Borland compiler or the (not so new) Visual C++ Tooolkit 2003. Visual Studio just isn't something I can justify, even if the lab would pay for it. Especially with the VC6 vs VC7 funny business. At least Tk stopped leaking!

Since I have a wx that works, its not so bad. I'll still work on the FC2 issues, but I can see about that toolbar addition as well. I should get around to installing windows again at home, I've been putting off an upgrade to XP, and my old win2k pro install is useless, after a mobo upgrade.

Chris

A nice source for some SVG inspiration might be PyChart, another python plotting system, that has a fully functional SVG backend: http://www.hpl.hp.com/personal/Yasushi_Saito/pychart.

I like the looks/features of this package, but it lacks interactivity/GUI support and relies on ghostscript for rendering images.

>>>>> "cfuller" == cfuller  <cf...@th...> writes:

    cfuller> Ack, I'll try not to let it go to my head! Hacking on
    cfuller> backends is more then I bargained for! I'm no good at
    cfuller> making bitmaps, so someone else should figure out what
    cfuller> the button should look like. I'll make my demo into a
    cfuller> proper dialog, hook it up to generic button, and I'll
    cfuller> even take a whack at the wx backend, but that will be
    cfuller> tricky, since it won't even run on my machine for some
    cfuller> reason. I suppose you just got a reluctant wx developer,
    cfuller> if I'm going to debug that :)

That would be great - there are some known gtk related problems with
the latest wx release that are independent of matplotlib.  What
platform are you having trouble with?


As for the pixmap, what we've done in the past is I pick some suitable
gtk stock icon and we just copy that into a pixmap.

    cfuller> If all goes well, I'll see about the ticklabels.

    cfuller> As for your question about axes_demo, I was thinking
    cfuller> along these lines while developing the code. It would be
    cfuller> simple enough to generalize to working with a list of
    cfuller> axes, rather than passing a figure and querying the
    cfuller> figure to get the list. A bounding box would have to be
    cfuller> passed along also, to stand in for the full figure
    cfuller> [0,1],[0,1] box. A default bounding box could be defined
    cfuller> by the axes. It should be fairly simple to implement. I
    cfuller> just thought it was getting a little too ambitious :) (or
    cfuller> at least more then my needs required, if you prefer!)

I don't think its necessary to support Axes, just to fail gracefully.
They are only used when people place them explicity, so I don't know
that it's necessary to interactively resize them.  Failing gracefully
with something along the lines of

if not isinstance(ax, axes.Subplot):
    error_msg('only works with subplots')

would satisfy me.

    cfuller> For that matter, it doesn't have to work on only
    cfuller> axes. Any rectangular region would do. The Text class
    cfuller> even has get_position and set_position methods! See the
    cfuller> kind of trouble I get into?

Whoa now, easy there.  

    cfuller> Chris Fuller


    cfuller> -------------------------------------------------------
    cfuller> This SF.Net email sponsored by Black Hat Briefings &
    cfuller> Training.  Attend Black Hat Briefings & Training, Las
    cfuller> Vegas July 24-29 - digital self defense, top technical
    cfuller> experts, no vendor pitches, unmatched networking
    cfuller> opportunities. Visit www.blackhat.com
    cfuller> _______________________________________________
    cfuller> Matplotlib-devel mailing list
    cfuller> Mat...@li...
    cfuller> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

Ack, I'll try not to let it go to my head! Hacking on backends is more then I bargained for! I'm no good at making bitmaps, so someone else should figure out what the button should look like. I'll make my demo into a proper dialog, hook it up to generic button, and I'll even take a whack at the wx backend, but that will be tricky, since it won't even run on my machine for some reason. I suppose you just got a reluctant wx developer, if I'm going to debug that :)

If all goes well, I'll see about the ticklabels.

As for your question about axes_demo, I was thinking along these lines while developing the code. It would be simple enough to generalize to working with a list of axes, rather than passing a figure and querying the figure to get the list. A bounding box would have to be passed along also, to stand in for the full figure [0,1],[0,1] box. A default bounding box could be defined by the axes. It should be fairly simple to implement. I just thought it was getting a little too ambitious :) (or at least more then my needs required, if you prefer!)

For that matter, it doesn't have to work on only axes. Any rectangular region would do. The Text class even has get_position and set_position methods! See the kind of trouble I get into?


Chris Fuller

>>>>> "cfuller" == cfuller  <cf...@th...> writes:

    cfuller> I've been playing with matplotlib a couple of weeks
    cfuller> now. That nasty memory leak was starting to turn me off,
    cfuller> and then got fixed right on cue! I've had a hell of a
    cfuller> time finding a good plotting system to replace BLT, so
    cfuller> that was a relief.

There have been some additional memory leak fixes since 0.54.2.  Glad
we could keep you on board!

    cfuller> Anyhow, I noticed my titles overwriting the x axes of
    cfuller> adjacent subplots, and so I wrote a little function that
    cfuller> shifts subplots around. That's not terribly interesting,
    cfuller> but I have a compulsive habit of making everything I do
    cfuller> as general as possible. The result is a system for
    cfuller> controlling figure layout based on six parameters, three
    cfuller> for each axis. As well as a rather elaborate demo. The
    cfuller> basic demo was done in an hour, but I couldn't leave well
    cfuller> enough alone!

Ahh, just the kind of developer we need....  

    cfuller> I tried to get the docstrings right, but haven't checked
    cfuller> them. I also included a mathematical derivation at the
    cfuller> end of the file, and I don't know what the documentation
    cfuller> scripts will make of that.

    cfuller> There's no error checking. None should be needed, as long
    cfuller> as all the parameters are numeric. Some curious results
    cfuller> are very possible, but nothing breaks.

    cfuller> If you find the documentation and API (and whatever else)
    cfuller> worthy, you can include it in the next release/current
    cfuller> CVS if you think it appropriate.

I really liked your demo.  So much that I think we should include it
into the distribution and incorporate versions of your widgets for
each backend.  I think we should have a button on the standard toolbar
that launches your dialog box.  I can do the one for GTK - perhaps you
or Todd would like to incorporate your changes into backend_tkagg so
that a toolbar button launches the dialog box you wrote.

It would also be nice to add checkboxes for turning off y tick labels
for all but the first column and turning off x tick labels for all but
the last row.

Any takers for doing the wx version?

Another question: what would happen if your code were run on non
subplot axes, eg with those examples/axes_demo.py?

A word of warning: in your demo code, you define a function set.
There is a matplotlib.matlab function with the same name...

    cfuller> Oh, and if the user wants to shift individual subplots
    cfuller> around, that function is still present.

I've been playing with matplotlib a couple of weeks now. That nasty memory leak was starting to turn me off, and then got fixed right on cue! I've had a hell of a time finding a good plotting system to replace BLT, so that was a relief.

Anyhow, I noticed my titles overwriting the x axes of adjacent subplots, and so I wrote a little function that shifts subplots around. That's not terribly interesting, but I have a compulsive habit of making everything I do as general as possible. The result is a system for controlling figure layout based on six parameters, three for each axis. As well as a rather elaborate demo. The basic demo was done in an hour, but I couldn't leave well enough alone!

I tried to get the docstrings right, but haven't checked them. I also included a mathematical derivation at the end of the file, and I don't know what the documentation scripts will make of that.

There's no error checking. None should be needed, as long as all the parameters are numeric. Some curious results are very possible, but nothing breaks.

If you find the documentation and API (and whatever else) worthy, you can include it in the next release/current CVS if you think it appropriate.

Oh, and if the user wants to shift individual subplots around, that function is still present.

Chris Fuller