matplotlib-users Mailing List for matplotlib

On Fri, Jul 20, 2012 at 10:12:43AM -0500, Benjamin Root wrote:
> On Fri, Jul 20, 2012 at 8:50 AM, Damon McDougall
> <dam...@gm...>wrote:
> 
> > Howdy all,
> >
> > Not sure if I'm being a giant noob, but is there any way to plot a
> > vector field (a la quiver) on the (x, y)-plane of an Axes3D object? The
> > behaviour I desire is exactly that of
> >
> > http://matplotlib.sourceforge.net/mpl_toolkits/mplot3d/tutorial.html#filled-contour-plots
> >
> > But instead of a contour plot on the (x, y)-plane, I want a quiver plot
> > there.
> >
> > Any ideas?
> > Thank you :)
> >
> >
> I just tried to see if it was possible with only a few tweaks, and it
> revealed some limitations in the mplot3d code with respect to handling
> collection objects subclassed from PolyCollections (and others).
> Unfortunately, I don't see any immediate work-around. Could you please
> file a feature request? I may or may not be able to address it this
> weekend.
>
Mate, you are a machine. I was thinking of tweaking myself, but it seems
like since you hit a brick wall I'll wait to see if you get a branch
going. When you do, I'll fork that bad boy.
Thanks for following up!
> 
> Ben Root
-- 
Damon McDougall
http://damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

On Fri, Jul 20, 2012 at 8:50 AM, Damon McDougall
<dam...@gm...>wrote:
> Howdy all,
>
> Not sure if I'm being a giant noob, but is there any way to plot a
> vector field (a la quiver) on the (x, y)-plane of an Axes3D object? The
> behaviour I desire is exactly that of
>
> http://matplotlib.sourceforge.net/mpl_toolkits/mplot3d/tutorial.html#filled-contour-plots
>
> But instead of a contour plot on the (x, y)-plane, I want a quiver plot
> there.
>
> Any ideas?
> Thank you :)
>
>
I just tried to see if it was possible with only a few tweaks, and it
revealed some limitations in the mplot3d code with respect to handling
collection objects subclassed from PolyCollections (and others).
Unfortunately, I don't see any immediate work-around. Could you please
file a feature request? I may or may not be able to address it this
weekend.
Ben Root

On Thu, Jul 19, 2012 at 3:17 PM, ObsessiveMathsFreak <
obs...@gm...> wrote:
> I used sagemath which uses matplotlib as its plotting interface. After
> extensive investigation I was extremely disappointed to find that
> matplotlib has no fundamental support for drawing arrows at the ends of
> axes.
>
> Is there no way that such basic functionality could be included in the
> next matplotlib release. Add-ons such as artistsaxis are not really
> appropriate as they require extensive refactoring of code using splines
> or axes.
>
> A simple "endarrow=True" option on either splines or axes would be a
> very useful addition to the axis interface. Many people would be very
> grateful if you could consider adding this feature in the next release.
>
>
>
Agreed, this would be a wonderful feature, and is often requested. The
AxesGrid toolkit does provide this functionality
http://matplotlib.sourceforge.net/examples/axes_grid/demo_axisline_style.html
but it would be nice to eventually move some of this into the standard
matplotlib axes object.
Two of the hurdles I foresee are:
 * API -- there are many kinds of arrows and I am sure once this ability
is added, users are going to want to tweak the appearance. We will need to
have an easy way to turn them on, select which axes to have them for, the
type of arrow and their size and other properties. Finally, I am sure this
will need to be at least partly accessible through rcparams.
 * Locating -- If we place the arrowhead at the very end of the axis
spine, we possibly run the risk of the arrowhead being clipped by the
bounding box rectangle. We would also need it to be taken into account for
tight_layout and for bbox_inches='tight'. On the other hand, one could
position the arrow head such that the tip of the arrow is at the end of the
spine, but one runs the risk of colliding with the tick and tick labels.
By no means should this discourage someone from attempting this. But I
wanted to outline some of the considerations one would need to take into
account while creating this feature.
Cheers!
Ben Root

Howdy all,
Not sure if I'm being a giant noob, but is there any way to plot a
vector field (a la quiver) on the (x, y)-plane of an Axes3D object? The
behaviour I desire is exactly that of
http://matplotlib.sourceforge.net/mpl_toolkits/mplot3d/tutorial.html#filled-contour-plots
But instead of a contour plot on the (x, y)-plane, I want a quiver plot
there.
Any ideas?
Thank you :)
-- 
Damon McDougall
http://damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

On Fri, Jul 20, 2012 at 5:39 AM, Damon McDougall
<dam...@gm...>wrote:
> On Fri, Jul 20, 2012 at 12:14:08PM +0200, Fabien Lafont wrote:
> > Hello everyone,
> >
> > Is it possible to have automaticaly more than 3 colors when Iplot a
> graph?
> > When I plot it put the first in blue the second in green the third in red
> > and the fourth in blue again. I want to use more colors to differenciate
> > the curves.
> >
>
> Strange, this is the default color cycle (colormaps are something
different).
['b','g','r','c','m','y','k']
You can modify this via the "axes.color_cycle" rcparam.
Cheers!
Ben Root

On Fri, Jul 20, 2012 at 12:14:08PM +0200, Fabien Lafont wrote:
> Hello everyone,
> 
> Is it possible to have automaticaly more than 3 colors when Iplot a graph?
> When I plot it put the first in blue the second in green the third in red
> and the fourth in blue again. I want to use more colors to differenciate
> the curves.
>
Sure. Here's an example I cooked up for you:
https://gist.github.com/3150091
Hope that helps.
> 
> Is it possible?
> 
> Fabien
-- 
Damon McDougall
http://damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom

Hello everyone,
Is it possible to have automaticaly more than 3 colors when Iplot a graph?
When I plot it put the first in blue the second in green the third in red
and the fourth in blue again. I want to use more colors to differenciate
the curves.
Is it possible?
Fabien

On Thu, Jul 19, 2012 at 11:34 PM, JonBL <jc....@bi...> wrote:
>
> I have a Python program which calls matplotlib's show() method to display a
> plot, but control does not return to my program until I close the displayed
> figure. I want control to immediately return to my program so that I can
> display additional figures as well.
>
> The doco (matplotlib 1.1.1) for the show() method mentions an experimental
> key word arg named 'block', that can be set to True or False. This looks
> promising, but plt.show(block = False) raises type error "got an unexpected
> keyword argument 'block'". A call to plt.show() works fine. The method
> appears to accept no arguments.
>
> Can anyone suggest how to bypass the blocking behaviour of the show()
> method?
>
> TIA,
> Jon
>
"The only thing worse than no comments are outdated comments"
IIRC, "block" was long removed from matplotlib. If you want non-blocking
behavior, just call "plt.ion()" to turn interactivity on. There are other
approaches to this problem that you would want to examine if you were truly
embedding into an application, but in the most simple cases, just turning
interactivity on should do the trick.
Cheers!
Ben Root

Solved - just discovered methods ion() and ioff() which do the job.
JonBL wrote:
> 
> I have a Python program which calls matplotlib's show() method to display
> a plot, but control does not return to my program until I close the
> displayed figure. I want control to immediately return to my program so
> that I can display additional figures as well.
> 
> The doco (matplotlib 1.1.1) for the show() method mentions an experimental
> key word arg named 'block', that can be set to True or False. This looks
> promising, but plt.show(block = False) raises type error "got an
> unexpected keyword argument 'block'". A call to plt.show() works fine. The
> method appears to accept no arguments.
> 
> Can anyone suggest how to bypass the blocking behaviour of the show()
> method?
> 
> TIA,
> Jon
> 
-- 
View this message in context: http://old.nabble.com/How-to-Turn-Off-Blocking-by-Method-show%28%29-tp34188043p34188078.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

I have a Python program which calls matplotlib's show() method to display a
plot, but control does not return to my program until I close the displayed
figure. I want control to immediately return to my program so that I can
display additional figures as well.
The doco (matplotlib 1.1.1) for the show() method mentions an experimental
key word arg named 'block', that can be set to True or False. This looks
promising, but plt.show(block = False) raises type error "got an unexpected
keyword argument 'block'". A call to plt.show() works fine. The method
appears to accept no arguments.
Can anyone suggest how to bypass the blocking behaviour of the show()
method?
TIA,
 Jon
-- 
View this message in context: http://old.nabble.com/How-to-Turn-Off-Blocking-by-Method-show%28%29-tp34188043p34188043.html
Sent from the matplotlib - users mailing list archive at Nabble.com.