matplotlib-devel Mailing List for matplotlib

On Wed, Jul 30, 2008 at 4:31 PM, Eric Bruning <eri...@gm...> wrote:
> I tried with the Tk backend, and I'm seeing the same problem as Eric F
> has with GTK. The problem seems to be with idle_event not getting
> triggered on some of the backends.
>
> The checkins having to do with idle_event from the end of June
> indicate that idle_event is still somewhat experimental.
>
> What's the current status of idle_event? Can I help in some way?

I was working on idle event, to make a backend neutral api for idle
processing, animation, etc...  There was a particularly nasty problem
with tk since I had to do the threading calls myself and I am no
threading guru, so I reverted some of the changes, mainly because of
the problem of cross thread signal handling blocked the CTRL-C
interrupts.  These changes were added in r5652

  r5652 | jdh2358 | 2008-06-23 16:39:11 -0500 (Mon, 23 Jun 2008) | 1 line

  draft idle/timeout api

and subsequent refactoring

  r5653 | jdh2358 | 2008-06-23 23:17:21 -0500 (Mon, 23 Jun 2008) | 1 line

  replaced idle handler with idle event

but then reverted when Nils reported the CTRL-C bug in

  r5655 | jdh2358 | 2008-06-24 08:56:55 -0500 (Tue, 24 Jun 2008) | 1 line

  removed idle support from tkagg until I figure out interrupts

So the implementation is incomplete and tkagg needs to most work,
since for other GUIs it is fairly easy to pass through the native
enent handling.  I am not sure what the problem is with GTK right now,
and haven't had a chance to look at it.

JDH

I tried with the Tk backend, and I'm seeing the same problem as Eric F
has with GTK. The problem seems to be with idle_event not getting
triggered on some of the backends.

The checkins having to do with idle_event from the end of June
indicate that idle_event is still somewhat experimental.

What's the current status of idle_event? Can I help in some way?

If idle_event can't be made robust, I'll need to look for another
method to address the following problem. When I close off the polygon,
I determined that I needed to fall back to the idle state to be able
to see the closed polygon. My code, itself triggered off a right mouse
click, was apparently blocking drawing to the screen.  It is desirable
to see the closed polygon on screen before the polygon callback in the
event that the callback takes a long time.

Thanks,
Eric

On Wed, Jul 30, 2008 at 10:58 AM, Nils Wagner
<nw...@ia...> wrote:

> Sorry for the delay and thank you for your patience.
> Here is the output of the "never-ending story"

Could you try again with r5932

On Tue, 29 Jul 2008 16:43:42 -0500
  "John Hunter" <jd...@gm...> wrote:
> On Tue, Jul 29, 2008 at 4:38 PM, Paul Kienzle 
><pki...@ni...> wrote:
> 
>> Okay, how about
>>    if not hasattr(self,'IsVisible'):
>>        self.IsVisible = lambda self: True
>> in __init__
>>
>> That way it will always think it is visible.
> 
> OK, Nils, I committed a minor variant of that in r5926. 
> Test again...
> 
> JDH

Hi John,

Sorry for the delay and thank you for your patience.
Here is the output of the "never-ending story"

Traceback (most recent call last):
   File 
"/usr/lib/python2.4/site-packages/matplotlib/backends/backend_wx.py", 
line 1099, in _onPaint
     self.draw(drawDC=drawDC)
   File 
"/usr/lib/python2.4/site-packages/matplotlib/backends/backend_wxagg.py", 
line 64, in draw
     self.gui_repaint(drawDC=drawDC)
   File 
"/usr/lib/python2.4/site-packages/matplotlib/backends/backend_wx.py", 
line 1000, in gui_repaint
     if self.IsShownOnScreen():
TypeError: <lambda>() takes exactly 1 argument (0 given)
Traceback (most recent call last):
   File 
"/usr/lib/python2.4/site-packages/matplotlib/backends/backend_wx.py", 
line 1101, in _onPaint
     self.gui_repaint(drawDC=drawDC)
   File 
"/usr/lib/python2.4/site-packages/matplotlib/backends/backend_wx.py", 
line 1000, in gui_repaint
     if self.IsShownOnScreen():
TypeError: <lambda>() takes exactly 1 argument (0 given)
  
BTW, the entries are identical for both revisions.
Is that intended ?

r5923 | jdh2358 | 2008-07-29 20:01:16 +0200 (Di, 29 Jul 
2008) | 1 line

special case contains test for degenerate rectangle
------------------------------------------------------------------------
r5922 | jdh2358 | 2008-07-29 19:47:50 +0200 (Di, 29 Jul 
2008) | 1 line

special case contains test for degenerate rectangle

On Tue, Jul 29, 2008 at 11:12 PM, Eric Firing <ef...@ha...> wrote:
> Eric Bruning wrote:
>>
>> On Tue, Jul 29, 2008 at 10:53 PM, Eric Firing <ef...@ha...> wrote:
>>>
>>> Eric Bruning wrote:
>
>>> I gave it a quick try with gtk backend.  The polygon is drawn based on
>>> clicks, but nothing gets printed.  Is this expected?
>>
>> That is unexpected. The callback should change the color of the
>> lassoed dots (from green to blue, for my default colormap) and print
>> the verts, the coords of the lassoed points, and the 'charge' array.
>>
>> Try adding a print in do_callback. One possibility is that the trigger
>> on idle_event isn't working.
>
> do_callback is not getting run.
>
> Eric

On Wed, Jul 30, 2008 at 5:40 AM, Mark Bakker <ma...@gm...> wrote:
> Hello Eric -
>
> I am interested. I wouldn't mind seeing it in mpl, but before we get there
> is there a chance you may want to share the code with me? I presume it is
> not long.

Hi Mark,

The code was attached to the original post. If you can't dig it out of
the archive, I'll send it personally. Let me know.

> My other question is: why do you want to click a polygon? Wouldn't drawing a
> rectangular box be less stress on your carpal tunnel?

My use case requires selecting from a set of points that are
distributed in a way that is unsuitable for selection with a
rectangle.

(cc'ing the mpl-dev list)
-Eric B

Eric Bruning wrote:
> I wanted something more precise (and easier on the carpal tunnel) than
> the included Lasso widget. Inspired by the existing Lasso, I forked
> out an alternate approach that creates a polygon with mouse clicks and
> returns a list of vertices to a callback function.
> 
> See the attached for a demo. It's been tested with WxAgg running off
> recent svn. I'm using idle_event, which isn't mentioned in the docs
> for the released version.
> 
> If there's interest, I'll put together a patch to add this to widgets.py.

Yes, there is interest--we should have this functionality if it can work 
on all interactive backends.

I gave it a quick try with gtk backend.  The polygon is drawn based on 
clicks, but nothing gets printed.  Is this expected?

Eric

> 
> Thanks,
> Eric B
> 
> 
> ------------------------------------------------------------------------
> 
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

I wanted something more precise (and easier on the carpal tunnel) than
the included Lasso widget. Inspired by the existing Lasso, I forked
out an alternate approach that creates a polygon with mouse clicks and
returns a list of vertices to a callback function.

See the attached for a demo. It's been tested with WxAgg running off
recent svn. I'm using idle_event, which isn't mentioned in the docs
for the released version.

If there's interest, I'll put together a patch to add this to widgets.py.

Thanks,
Eric B