matplotlib-users Mailing List for matplotlib

On Wed, Mar 2, 2011 at 8:46 AM, Yuri D'Elia <wa...@th...> wrote:
>
>
>
> On Wed, 2 Mar 2011 22:01:02 +0900
> Jae-Joon Lee <lee...@gm...> wrote:
>
> > >> > Is this a bug?
> > >>
> > >> Unfortunately, bbox_inches option is never meant to be complete in
> > >> figuring out the exact size of the figure area.
> > >
> > > Why not? What's the purpose of bbox_inches='tight' otherwise?
> >
> > Figuring out enclosing bbox when arbitrary spline paths are involved
> > is difficult (I think there is no exact solution). So I only intended
> > to support common cases.
>
> Ok, I can understand that, but shouldn't all artists used to construct the
> picture, as suptitle, be considered?
>
> > >> However, you can use "bbox_extra_artists" keyword argument to specify
> > >> additional artists that should be considered when dertermining the
> > >> plot size.
> > >>
> > >> mytitle = fig.suptitle("Horray!", fontsize=20)
> > >>
> > >> ...
> > >>
> > >> fig.savefig("out.png", bbox_inches='tight',
> bbox_extra_artists=[mytitle])
> > >
> > > That doesn't work for me either.
> >
> > Can you be more specific? Does it throw an exception? Or the code runs
> > without any error but the output is still wrong?
>
> No error/exception are produced. The output is simply identical to the one
> without bbox_extra_artists.
>
> This also works in my previous example:
>
> import matplotlib as mpl
> import matplotlib.figure
> import matplotlib.backends.backend_agg
>
> fig = mpl.figure.Figure()
> cvs = mpl.backends.backend_agg.FigureCanvasAgg(fig)
> fig.set_size_inches((20,20))
> plot = fig.add_subplot(111)
> plot.set_title("Subtitle")
> plot.plot([1,2,3], [3,2,1])
> st = fig.suptitle("Horray!", fontsize=20)
> fig.savefig("out.png", bbox_inches='tight', bbox_extra_artists=[st])
>
>
Which version of matplotlib are you using? This example works for me using
the latest matplotlib from source. Also, why the awkward usage and
imports? If you want to force the Agg backend to be used, you could just
do:
import matplotlib
matplotlib.use("Agg")
before any other matplotlib imports.
Ben Root

On Thu, Mar 3, 2011 at 9:54 PM, George Washington <gw...@ho...>wrote:
> I am new to Matplotlib and am having some problems plotting the following
> set of coordinates (in python 2.6 and Win 7 32)
> This is just a small sample of the data:
>
> Seq.No. x-scale y-scale z-scale
> 01.000000 1579446.055280 5361974.495490 1342.967407
> 02.000000 1579446.646620 5361972.813700 1342.967407
> 03.000000 1579448.047050 5361968.830880 1341.237305
> 04.000000 1579450.992084 5361963.830880 1337.739502
> 05.000000 1579453.937117 5361958.830880 1336.262817
> ...
> ...
>
> with the following outcome:
> (plot3d.png)
>
> *Problem*: while the x-scale is ok and the z-scale looks ok, the y-scale
> is definitely not ok. The numbers in the image are *not* the ones in the
> y-array (I double checked.)
>
George,
This is a known issue where very large axis values were being represented
using an "offset" (much like in 2-d plots with very large axis values). The
problem was that the offset was not displayed for 3d plots. This is
definitely fixed in the latest development branch, but I can't remember if I
fixed it in the 1.0.1 release (probably not).
>
> I also have a number of questons:
> *Question1*: how does one create a label for the z scale? (zlabel is not
> valid)
>
This should be fixed by the next release. Most functions like set_xlim(),
set_ylabel() and such merely call that function for the appropriate axis
object. If a particular function is missing, you can call it yourself doing
something like the following:
ax.zaxis.set_label()
Note that if you have an earlier version of matplotlib, you might have to
do:
ax.w_zaxis.set_label()
> *Queston2*: Is it possible to fill below the line (so it looks like a
> mountain) and how
>
Never tried considered anything like that. Might involve creating a 3D
patch object with some sort of path completion. File a feature request on
the matplotlib tracker here:
http://sourceforge.net/tracker2/?group_id=80706
> *Question3*: Is it possible to traverse sequentially all the points
> plotted in the image so as to make computations (such as distance to the
> next point, etc..).
> The points themselves come from a text file but are not
> in sequence. They are sequenced by being plotted in their right position in
> 3D space.
>
>
I am not exactly sure what you mean, but there is a nice data structure that
I use to do efficient data operations on spatial data called KDTrees, which
can be found in scipy:
http://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.KDTree.html
I hope that is helpful.
Ben Root

Thank you for the reply.
I discovered this myself yesterday. Now I have an official answer if people
want a colored + or the x symbol in the scatter plot.
Scott Hansen
efiring wrote:
> 
> On 03/02/2011 06:42 AM, Mr. Python wrote:
>>
>> I am unable to set the color of the scatter plot points using the code
>> below:
>>
>> import matplotlib
>> matplotlib.use('Qt4Agg')
>>
>> import matplotlib.pyplot as plt
>> import numpy as np
>>
>> plt.scatter(newSpdVals, newEngLoadVals,c='g',marker='+')
> 
> This is a bug in the handling of symstyle 2, which includes '+' and 'x'.
> 
>>
>> #Add the graph title and x/y axis labels
>> plt.title("Gear Span")
>> plt.xlabel("Vehicle Speed (mph)")
>> plt.ylabel("Engine Load %")
>> plt.show()
>>
>> No matter what I change the c parameter to ('b','r', etc.), the color of
>> the
>> scatter plot points are always black. The lists newSpdVals and
>> newEngLoadVals have 100 dimmensions and are of type numpy.ndarray.
>>
>> I am using the following software and hardware:
>>
>> Windows XP
>> Python 2.7
>> PyQt4 (I launch the plot from a PyQt GUI. I don't embed the plot in the
>> GUI).
>> matplotlib 1.01
>>
>> In the beginning of the same function where my scatter plot code is, I
>> added
>> the following code to verify if the problem is due to the matplotlib/PyQt
>> interface:
>>
>> a = np.linspace(-10,100,33)
>> b = np.linspace(-32,400,33)
>> plt.scatter(a, b,color='g',marker='+')
> 
> Note the difference: here you are using a 'color' kwarg that is getting 
> passed directly to the AsteriskPolygonCollection, whereas before you 
> were using the 'c' kwarg that is being handled incorrectly in this case 
> inside of the scatter code.
> 
> Eric
> 
>> plt.title("Gear Span")
>> plt.xlabel("Vehicle Speed (mph)")
>> plt.ylabel("Engine Load %")
>> plt.show()
>>
>> The points for this plot are colored green. I'm able to change the point
>> color to blue, red, etc.
>>
>> Any ideas why I can't change the point color of my first graph? Thank
>> you
>> for your help.
>>
>>
> 
> 
> ------------------------------------------------------------------------------
> What You Don't Know About Data Connectivity CAN Hurt You
> This paper provides an overview of data connectivity, details
> its effect on application quality, and explores various alternative
> solutions. http://p.sf.net/sfu/progress-d2d
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
> 
-- 
View this message in context: http://old.nabble.com/Can%27t-set-scatter-plot-point-color-tp31051071p31070224.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

On 03/02/2011 06:42 AM, Mr. Python wrote:
>
> I am unable to set the color of the scatter plot points using the code below:
>
> import matplotlib
> matplotlib.use('Qt4Agg')
>
> import matplotlib.pyplot as plt
> import numpy as np
>
> plt.scatter(newSpdVals, newEngLoadVals,c='g',marker='+')
This is a bug in the handling of symstyle 2, which includes '+' and 'x'.
>
> #Add the graph title and x/y axis labels
> plt.title("Gear Span")
> plt.xlabel("Vehicle Speed (mph)")
> plt.ylabel("Engine Load %")
> plt.show()
>
> No matter what I change the c parameter to ('b','r', etc.), the color of the
> scatter plot points are always black. The lists newSpdVals and
> newEngLoadVals have 100 dimmensions and are of type numpy.ndarray.
>
> I am using the following software and hardware:
>
> Windows XP
> Python 2.7
> PyQt4 (I launch the plot from a PyQt GUI. I don't embed the plot in the
> GUI).
> matplotlib 1.01
>
> In the beginning of the same function where my scatter plot code is, I added
> the following code to verify if the problem is due to the matplotlib/PyQt
> interface:
>
> a = np.linspace(-10,100,33)
> b = np.linspace(-32,400,33)
> plt.scatter(a, b,color='g',marker='+')
Note the difference: here you are using a 'color' kwarg that is getting 
passed directly to the AsteriskPolygonCollection, whereas before you 
were using the 'c' kwarg that is being handled incorrectly in this case 
inside of the scatter code.
Eric
> plt.title("Gear Span")
> plt.xlabel("Vehicle Speed (mph)")
> plt.ylabel("Engine Load %")
> plt.show()
>
> The points for this plot are colored green. I'm able to change the point
> color to blue, red, etc.
>
> Any ideas why I can't change the point color of my first graph? Thank you
> for your help.
>
>

On 03/01/2011 10:22 PM, Jason Grout wrote:
> I tried building the standalone html docs using:
>
> cd doc
> python make.py html
>
> I notice that there are around 30 .pyc files left in the
> build/html/pyplots/ directory. Are these needed in the html
> documentation build directory?
> 
No -- but they are also a little hard to avoid. I have made a change to 
clean these out after the fact.
> Also, it seems that the files in _images are redundant, as they are
> referenced in their original directory, not _images.
>
> from the build/html directory:
>
> % find . -name multiline.pdf
> ./_images/multiline.pdf
> ./plot_directive/mpl_examples/pylab_examples/multiline.pdf
> % grep -ri "multiline.pdf" *
> examples/pylab_examples/multiline.html:<p>[<a class="reference external"
> href="../../plot_directive/mpl_examples/pylab_examples/multiline.py">source
> code</a>,<a class="reference external"
> href="../../plot_directive/mpl_examples/pylab_examples/multiline.hires.png">hires.png</a>,
> <a class="reference external"
> href="../../plot_directive/mpl_examples/pylab_examples/multiline.pdf">pdf</a>]</p>
> 
This one is harder. Sphinx annoyingly always copies every image that is 
displayed on a page to the _images directory. However, it is impossible 
to link (i.e. through an <a href="...">) to an image in the _images 
directory. Since our plot directive results need to both display the 
image and link to different resolutions of them, we have to at least 
have two copies of the "normal resolution" pngs, unless we can solve the 
problem in Sphinx itself or do some sort of postprocessing on the HTML 
output.
However, the .pdf files ending up in _images is the result of the 
peculiar way in which HTML and LaTeX are generated from the same 
source. Sphinx seems to think it is displaying both png and pdf images 
in the HTML and copies them both over. I have added a post-processing 
step to our build that subsequently removes the pdf files under _images.
You can see the changes here:
https://github.com/matplotlib/matplotlib/pull/28
Mike
> Any comments about trimming down the size of the build/html directory?
>
> Thanks,
>
> Jason
>
> ------------------------------------------------------------------------------
> Free Software Download: Index, Search& Analyze Logs and other IT data in
> Real-Time with Splunk. Collect, index and harness all the fast moving IT data
> generated by your applications, servers and devices whether physical, virtual
> or in the cloud. Deliver compliance at lower cost and gain new business
> insights. http://p.sf.net/sfu/splunk-dev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
-- 
Michael Droettboom
Science Software Branch
Space Telescope Science Institute
Baltimore, Maryland, USA