matplotlib-users Mailing List for matplotlib

Hi Everyone,
There is a new formatter in ticker.py called NewScalarFormatter. If you have 
scientific notation in your plots, you may like the results. If you would 
like to try it out, you need to change ScalarFormatter->OldScalarFormatter, 
and NewScalarFormatter->ScalarFormatter. It will then be the default for 
linear scale axes. I would appreciate feedback, it will hopefully become the 
default at some point.
Darren

Jonathan Taylor wrote:
>Hi,
>	Please excuse the simple questions.
>
>i) Is it possible to draw step functions in matplotlib, i.e instead of
>joining points with lines, using steps instead. In R
>(http://cran.r-project.org) this would be done with something like,
>type='s'. I can think of a simple hack to do it, but thought it might
>exist in pylab, but couldn't find an example.
> 
>
try giving plot() an options linestyle='steps'. This was modeled on how 
gnuplot does it.
>ii) Masking in imshow: I want to plot a 2d image, but only show a
>particular region, setting the rest black (or some other color). Is
>there a way to get to the RGB data that is output from the colormap to
>explicitly set some regions black? Or maybe change the alpha level in
>these regions?
>
> 
>
i do something similar using the 'fill' command. Does the job.
-- 
Peter Groszkowski Gemini Observatory
Tel: +1 808 9742509 670 N. A'ohoku Place
Fax: +1 808 9359235 Hilo, Hawai'i 96720, USA 

Hi,
	Please excuse the simple questions.
i) Is it possible to draw step functions in matplotlib, i.e instead of
joining points with lines, using steps instead. In R
(http://cran.r-project.org) this would be done with something like,
type='s'. I can think of a simple hack to do it, but thought it might
exist in pylab, but couldn't find an example.
ii) Masking in imshow: I want to plot a 2d image, but only show a
particular region, setting the rest black (or some other color). Is
there a way to get to the RGB data that is output from the colormap to
explicitly set some regions black? Or maybe change the alpha level in
these regions?
Thanks,
Jonathan Taylor
-- 
Jonathan Taylor Tel: 650.723.9230
Dept. of Statistics Fax: 650.725.8977
Sequoia Hall, 137 www-stat.stanford.edu/~jtaylo
390 Serra Mall
Stanford, CA 94305

>>>>> "Rich" == Rich Drewes <dr...@in...> writes:
 Rich> Beautiful, thanks! I didn't even have to upgrade my 0.80
 Rich> matplotlib.
BTW, Rich, I read your brainlab article in the linux journal (I
subscribe) -- it was very nice. One of the main applications I had in
mind when I wrote the line collections code was efficiently making
spike raster plots, where you have bazillions of independent line
segments. Might be useful in brainlab...
By the way, some people in our group are using p-genesis and a
collaboration with Argonne National Labs (supercomputer) to do large
scale simulations of neocortex.
JDH

On Mon, 2 May 2005, Darren Dale wrote:
> What a coincidence, I just wrote an example two days ago that you might find 
> useful. You can find it in CVS; look for line_collection.py in the examples 
> directory.
Beautiful, thanks! I didn't even have to upgrade my 0.80 matplotlib.
Rich

Hi Rich,
On Monday 02 May 2005 12:57 pm, Rich Drewes wrote:
> Hello,
>
> I am trying to show many hundreds to a few thousand *independent* line
> segments efficiently. I can of course make one call to plot() per
> segment, but that is slooow and I am trying to keep up with a C
> application generating the plot data.
>
> As I read the docs for plot(), I can specify multiple points to be
> connected together using array arguments, but in order to specify
> multiple, independent, *disconnected* line segments in one call to plot()
> I would have to use multiple sets of arguments like this:
>
> plot(x1, y1, 'b-', x2, y2, 'r-', x3, y3, 'b-', . . .)
>
> That is as slow as multiple calls to plot and leads to huge argument
> lists.
>
> What would be great would be the ability to specify a marker *list* that
> maps to the segments, rather than just a marker string that is applied to
> all segments in the list. Something like this:
>
> plot([0, 1, 2, 3, 4], [2, 5, 7, 8, 9], ['b-', 'w-', 'b-', 'w-', 'b-'])
>
> Then by changing the color of certain segments to the background color I
> could effectively make many separate segments.
>
> Any thoughts on the wisdom of this? Or a suggestion for doing this with
> the current interface that didn't occur to me?
>
What a coincidence, I just wrote an example two days ago that you might find 
useful. You can find it in CVS; look for line_collection.py in the examples 
directory.
-- 
Darren S. Dale
Bard Hall
Department of Materials Science and Engineering
Cornell University
Ithaca, NY. 14850
dd...@co...

Hello,
I am trying to show many hundreds to a few thousand *independent* line
segments efficiently. I can of course make one call to plot() per
segment, but that is slooow and I am trying to keep up with a C
application generating the plot data.
As I read the docs for plot(), I can specify multiple points to be
connected together using array arguments, but in order to specify
multiple, independent, *disconnected* line segments in one call to plot()
I would have to use multiple sets of arguments like this:
plot(x1, y1, 'b-', x2, y2, 'r-', x3, y3, 'b-', . . .)
That is as slow as multiple calls to plot and leads to huge argument
lists.
What would be great would be the ability to specify a marker *list* that
maps to the segments, rather than just a marker string that is applied to
all segments in the list. Something like this:
plot([0, 1, 2, 3, 4], [2, 5, 7, 8, 9], ['b-', 'w-', 'b-', 'w-', 'b-'])
Then by changing the color of certain segments to the background color I 
could effectively make many separate segments.
Any thoughts on the wisdom of this? Or a suggestion for doing this with
the current interface that didn't occur to me?
Rich

On Sunday 01 May 2005 9:28 pm, John Hunter wrote:
> >>>>> "Fernando" == Fernando Perez <Fer...@co...> writes:
>
> Fernando> It's something I wrote long ago, and it also ships in
> Fernando> IPython.numutils. But John was kind enough to fold some
> Fernando> of my numutils pet favorites into pylab a while ago, if
> Fernando> nothing else so I'd shut up and actually start using
> Fernando> matplotlib :)
>
> It actually resides in matplotlib.mlab --- pylab just imports it.
> Since Darren is doing API work, he'll want it from mlab.
>
That's a nifty function, exactly what I am looking for. Thank you! And thank 
you John, you answered my next question.
-- 
Darren S. Dale
Bard Hall
Department of Materials Science and Engineering
Cornell University
Ithaca, NY. 14850
dd...@co...

>>>>> "Fernando" == Fernando Perez <Fer...@co...> writes:
 Fernando> It's something I wrote long ago, and it also ships in
 Fernando> IPython.numutils. But John was kind enough to fold some
 Fernando> of my numutils pet favorites into pylab a while ago, if
 Fernando> nothing else so I'd shut up and actually start using
 Fernando> matplotlib :)
It actually resides in matplotlib.mlab --- pylab just imports it.
Since Darren is doing API work, he'll want it from mlab.
JDH

Darren Dale wrote:
> On Sunday 01 May 2005 9:04 pm, John Hunter wrote:
> 
>>On 5/1/05, Darren Dale <dd...@co...> wrote:
>>
>>>When I do a=arange(-1,0.0001,.01), a[-1] is equal to something like
>>>2e-17. I
>>
>>Are you sure? I'm getting something different
> 
> 
> Yeah, I'm sure. I just tried reinstalling Numeric without linking to my 
> external linalg libraries (I had reason to suspect this due to a recent 
> discussion on gentoo-science), but the problem persists.
> 
> Fernando, is frange a homegrown function? Where can I find it?
pylab.frange
It's something I wrote long ago, and it also ships in IPython.numutils. But 
John was kind enough to fold some of my numutils pet favorites into pylab a 
while ago, if nothing else so I'd shut up and actually start using matplotlib :)
Best,
f

On Sunday 01 May 2005 9:04 pm, John Hunter wrote:
> On 5/1/05, Darren Dale <dd...@co...> wrote:
> > When I do a=arange(-1,0.0001,.01), a[-1] is equal to something like
> > 2e-17. I
>
> Are you sure? I'm getting something different
Yeah, I'm sure. I just tried reinstalling Numeric without linking to my 
external linalg libraries (I had reason to suspect this due to a recent 
discussion on gentoo-science), but the problem persists.
Fernando, is frange a homegrown function? Where can I find it?
-- 
Darren S. Dale
Bard Hall
Department of Materials Science and Engineering
Cornell University
Ithaca, NY. 14850
dd...@co...