matplotlib-users Mailing List for matplotlib

On 09/30/2011 09:12 AM, John Hunter wrote:
> On Thu, Sep 29, 2011 at 11:56 PM, Uri Laserson<las...@mi...> wrote:
>> Hi all,
>> Does anyone have a good solution for taking a line segment and coloring it
>> according to some gradient (where the start and end colors are the only two
>> things specified for the line coloring)?
>
> This example shows how to apply a colormap to a line segment
>
> http://matplotlib.sourceforge.net/examples/pylab_examples/line_collection2.html
>
> JDH
This example might be closer to the request:
http://matplotlib.sourceforge.net/examples/pylab_examples/multicolored_line.html
Eric

On Thu, Sep 29, 2011 at 11:56 PM, Uri Laserson <las...@mi...> wrote:
> Hi all,
> Does anyone have a good solution for taking a line segment and coloring it
> according to some gradient (where the start and end colors are the only two
> things specified for the line coloring)?
This example shows how to apply a colormap to a line segment
http://matplotlib.sourceforge.net/examples/pylab_examples/line_collection2.html
JDH

Hi Jeff,
I computed the convex hull for the points in paths returned by Basemap.drawcountries() and I got the attached map (convex hull algorithm: http://www.scipy.org/Cookbook/Finding_Convex_Hull). You can see the magenta convex hull polygon with U.S. border lines drawn in black as a reference. 
Unfortunately, adding the coastlines to the computation of the convex hull just hungs the program.
I tried running the computation of the convex hull using the polygons retrieved from US Shapefile in http://www.geophysique.be/en/2011/01/27/matplotlib-basemap-tutorial-07-shapefiles-unleached/, but the program hungs due to the huge amount of points too.
I am running out of time to complete this project, so I cannot implement a more efficient version of the convex hull computation to include the coastlines in the process right now. 
QUESTIONS:
- if I plot an open path in the area next to U.S./ Mexico border and I'd like to have the filling of such path to go all the way to the border, I figured I could add extra points to close the path over Mexico in order to fill the area I need. What I do not understand is how can I mask the filled portion over Mexico without masking any portion of the filling over U.S.. i.e. the portion that is outside the polygon shown in the map attached. 
- if I have similar situation in Florida, for which I would like to mask all points that are not over land using is_land() method, how can I do this ?
Thanks

Hi,
I'd like to show a background image after plotting dates. I use imshow,
which works, but makes the plotted line erased.
If somebody can help, it would be valuable.
thanks in advance
jm
Here is the code. If the line "ax.imshow(img,alpha=0.5)" is commented out,
the graph is correctly displayed.
# -*- coding: utf8 -*-
import datetime, csv
import matplotlib.pyplot as plt
from pylab import figure, show
from matplotlib import image
from matplotlib.dates import MONDAY, SATURDAY
from matplotlib.dates import DayLocator, MonthLocator, WeekdayLocator,
DateFormatter
''' data reading. Here is the data :
Nombre de signalements par jour du 01/08/2011 au 02/10/2011
20110801,8
20111002,10
'''
dates, data = [], []
dataReader = csv.reader(open('data.csv', 'rb'))
for i,row in enumerate(dataReader):
 if i == 0:
 title = row[0]
 else:
 w = row[0]
 w = datetime.date(int(w[0:4]),int(w[4:6]),int(w[6:8])).toordinal()
 dates += [w]
 data += [row[1]]
#
days = DayLocator()
months = MonthLocator()
mondays = WeekdayLocator(MONDAY)
dateFmt = DateFormatter("%d/%m/%Y")
fig = figure()
img = image.imread('carte quartiers Creil.png')
fig.set_facecolor('w')
ax = fig.add_subplot(111)
ax.get_frame().set_alpha(0)
ax.plot_date(dates, data,'-',color='r',alpha=0.75,linewidth=2)
loc = months
if dates[-1] - dates[0] <= 10:
 loc = days
ax.xaxis.set_major_locator(loc)
ax.xaxis.set_major_formatter(dateFmt)
ax.xaxis.set_minor_locator(mondays)
ax.grid(True)
plt.ylabel('nombre de signalements')
plt.title(title)
fig.autofmt_xdate()
ax.imshow(img,alpha=0.5)
show()

> That is why you have to work with the 3D objects from
> mpl_toolkits.mplot3d.art3d directly. There isn't a lot of documentation, so
> your best bet is to look at the source code of the module.
Indeed, I was able to extract the necessary pieces from the source
code. In fact I split the function `plot_surface` into two functions
`polyMeshFromMeshGrid` and `plot_surfaceMesh`. The first functions
takes the meshgrid arguments and produces a list of polygons that
should be drawn, the second functions finally displays them (the code
in plot_surface is already structured like this). With this approach I
can plot functions which are made up of several parts which each have
meshgrid structure. Also I can add raw polygons to the mesh.
Maybe this splitting makes sense in the online code as well?
-Holger
def polyMeshFromMeshGrid(self, X, Y, Z, *args, **kwargs):
 rows, cols = Z.shape
 tX, tY, tZ = np.transpose(X), np.transpose(Y), np.transpose(Z)
 rstride = kwargs.pop('rstride', 10)
 cstride = kwargs.pop('cstride', 10)
 if 'facecolors' in kwargs:
 fcolors = kwargs.pop('facecolors')
 else:
 color = np.array(colorConverter.to_rgba(kwargs.pop('color', 'b')))
 fcolors = None
 cmap = kwargs.get('cmap', None)
 norm = kwargs.pop('norm', None)
 vmin = kwargs.pop('vmin', None)
 vmax = kwargs.pop('vmax', None)
 shade = kwargs.pop('shade', cmap is None)
 lightsource = kwargs.pop('lightsource', None)
 # Shade the data
 if shade and cmap is not None and fcolors is not None:
 fcolors = self._shade_colors_lightsource(Z, cmap, lightsource)
 polys = kwargs.pop('polys', [])
 normals = kwargs.pop('normals', [])
 #colset contains the data for coloring: either average z or the facecolor
 colset = kwargs.pop('colset', [])
 for rs in np.arange(0, rows-1, rstride):
 for cs in np.arange(0, cols-1, cstride):
 ps = []
 corners = []
 for a, ta in [(X, tX), (Y, tY), (Z, tZ)]:
 ztop = a[rs][cs:min(cols, cs+cstride+1)]
 zleft = ta[min(cols-1, cs+cstride)][rs:min(rows, rs+rstride+1)]
 zbase = a[min(rows-1, rs+rstride)][cs:min(cols, cs+cstride+1):]
 zbase = zbase[::-1]
 zright = ta[cs][rs:min(rows, rs+rstride+1):]
 zright = zright[::-1]
 corners.append([ztop[0], ztop[-1], zbase[0], zbase[-1]])
 z = np.concatenate((ztop, zleft, zbase, zright))
 ps.append(z)
 # The construction leaves the array with duplicate points, which
 # are removed here.
 ps = zip(*ps)
 lastp = np.array([])
 ps2 = []
 avgzsum = 0.0
 for p in ps:
 if p != lastp:
 ps2.append(p)
 lastp = p
 avgzsum += p[2]
 polys.append(ps2)
 if fcolors is not None:
 colset.append(fcolors[rs][cs])
 else:
 colset.append(avgzsum / len(ps2))
 # Only need vectors to shade if no cmap
 if cmap is None and shade:
 v1 = np.array(ps2[0]) - np.array(ps2[1])
 v2 = np.array(ps2[2]) - np.array(ps2[0])
 normals.append(np.cross(v1, v2))
 return polys, normals, colset
def plot_surfaceMesh(self, polys, normals, colset, *args, **kwargs):
 had_data = self.has_data()
 rows, cols = Z.shape
 rstride = kwargs.pop('rstride', 10)
 cstride = kwargs.pop('cstride', 10)
 if 'facecolors' in kwargs:
 fcolors = kwargs.pop('facecolors')
 else:
 color = np.array(colorConverter.to_rgba(kwargs.pop('color', 'b')))
 fcolors = None
 cmap = kwargs.get('cmap', None)
 norm = kwargs.pop('norm', None)
 vmin = kwargs.pop('vmin', None)
 vmax = kwargs.pop('vmax', None)
 linewidth = kwargs.get('linewidth', None)
 shade = kwargs.pop('shade', cmap is None)
 lightsource = kwargs.pop('lightsource', None)
 polyc = art3d.Poly3DCollection(polys, *args, **kwargs)
 if fcolors is not None:
 if shade:
 colset = self._shade_colors(colset, normals)
 polyc.set_facecolors(colset)
 polyc.set_edgecolors(colset)
 elif cmap:
 colset = np.array(colset)
 polyc.set_array(colset)
 if vmin is not None or vmax is not None:
 polyc.set_clim(vmin, vmax)
 if norm is not None:
 polyc.set_norm(norm)
 else:
 if shade:
 colset = self._shade_colors(color, normals)
 else:
 colset = color
 polyc.set_facecolors(colset)
 self.add_collection(polyc)
 x, y, z = [], [], [];
 for pol in polys:
 for p in pol:
 x.append(p[0])
 y.append(p[1])
 z.append(p[2])
 self.auto_scale_xyz(x, y, z, had_data)
 return polyc

Dear all,
I am trying to draw a bar graph in which the bars are filled with a striped pattern rather than a solid color. For example an alternating pattern of red and yellow stripes with an equal width, at a 45 degree angle with respect to the bar. Is there some easy way to do this in matplotlib?
Thank you in advance,
--Michiel.

On Thu, Sep 29, 2011 at 4:31 PM, Holger Brandsmeier <
hol...@sa...> wrote:
> Hi Ben,
>
> in your example
> > http://matplotlib.sourceforge.net/examples/mplot3d/polys3d_demo.html
> the essential line is:
> ax.add_collection3d(poly, zs=zs, zdir='y')
> according to everything I see this can only draw something that is
> made out of several axes parallel parts. I don't think how with such a
> command you could end up with some picture like:
>
> http://matplotlib.sourceforge.net/plot_directive/mpl_examples/mplot3d/surface3d_demo.hires.png
That is correct. It is only for simple conversions.
>
> What I want to plot should look like that picture (with is part of
> your gallery), just that I don't have a meshgrid data underlying.
>
>
That is why you have to work with the 3D objects from
mpl_toolkits.mplot3d.art3d directly. There isn't a lot of documentation, so
your best bet is to look at the source code of the module.
> Your second example seems to have the same limitation
> http://matplotlib.sourceforge.net/examples/mplot3d/pathpatch3d_demo.html
> just it works with the command
> art3d.pathpatch_2d_to_3d(p, z=0, zdir="x")
>
> The only way that I currently see is that I call
> ax.plot_surface()
> for each quadrilateral of my mesh.
>
>
That would be overkill. However, try checking out the source code for
plot_surface() to see how it creates its Poly3DCollection object.
Ben Root

On Fri, Sep 30, 2011 at 4:27 AM, Holger Brandsmeier <
hol...@sa...> wrote:
> Dear List,
>
> is it possible to not only assign once color per polygon that is
> plotted, but one color for each vertex, so that the result looks like
> a properly smooth function. Even if I sample the points closely
> enough, in the current approach in almost all of your examples you
> always see the underlying grid (plus it gets very slow in 3D). I don't
> really care much if you interpolate the colors on the 2D projection of
> the polygon or on the real 3D polygon, just any color interpolation is
> better than none.
>
What values for rstride and cstride are you using? By default,
plot_surface() will sample every 10th point of the data array (for
performance reasons). Also, color interpoltion can be turned on by setting
shade to True.
>
> What do you actually understand behind the "antialiasing" parameter of
> 34D plots. It just puzzles me to have such a concept without have a
> concept of pixel shading (and thus an easy way for the above
> interpolation).
>
>
Again, plot_surface() has a shade kwarg. Use it and let me know if it meets
your expectations.
> I noticed that even 2D plots from matplot lib are very pixel dominated:
> http://matplotlib.sourceforge.net/examples/pylab_examples/pcolor_demo.html
>
> But there is the function imshow which doesn't have this artifacts:
> http://matplotlib.sourceforge.net/examples/pylab_examples/pcolor_demo2.html
>
>
for pcolor(), you can set the antialiaseds kwarg to True. The only reason
why imshow looks good in the second example is because the interpolation was
set to bilinear, and I think that automatically sets antialiasing to True.
> So is it possible to do something like imshow for 3D as well?
>
>
imshow() in 2d was not designed in such a way to yet be used in 3d, but I
think there is already a feature request filed for that.
Try shade=True for plot_surface() and see how that looks for you.
Cheers!
Ben Root

David Hoese wrote:
> Neal,
> 
> I do something similar to this where data that I'm plotting in a 2D line
> plot comes from a UDP socket and some memory mapped files. To
> accomplish the live updating I have a mix of your #2 and #3. I have a
> main GUI thread that displays the plots, then I have a second thread
> that gets data from a python generator and signals the GUI thread when
> that new data has arrived. Inside the generator is where I have socket
> communications and select calls. Hope this helps.
> 
Haven't tried it, but I did notice QSocketNotifier, which sounds like it may be 
what I need if I want to hook into qt event loop to implement as single-thread, 
single-process.
> -Dave
> 
> On 9/30/11 9:10 AM,
> mat...@li...
> wrote:
>> Message: 4
>> Date: 2011年9月30日 07:37:49 -0400
>> From: Neal Becker<ndb...@gm...>
>> Subject: [Matplotlib-users] how would you do this (animated bargraph)
>> To:mat...@li...
>> Message-ID:<j649me$ug41ドル...@do...>
>> Content-Type: text/plain; charset="ISO-8859-1"
>>
>> I just put together an animated bargraph to show results from a realtime
>> process.
>>
>> I used this as an example:
>> http://matplotlib.sourceforge.net/examples/animation/animation_blit_qt4.html
>>
>> The tricky part for me was that in the original design, there was a realtime
>> process running. I have to write some information to my hardware (at a rate
>> of
>> about 1/360ms). To do this, I have a kernel driver that uses 'read' to tell
>> the user space when an interrupt has occurred, which is when the user space
>> should write new information to the hardware.
>>
>> So I don't know how or if this could be hooked into qt event processing. 
>> Just for a quick-and-dirty demo, I just removed the realtime processing from
>> the user-space and put it in the kernel driver, so now my bargraph display
>> can simply update on a periodic timer, and the userspace code has no realtime
>> requirement. But this is just a kludge.
>>
>> So I wonder how other's would solve this? I'm thinking it would be either:
>>
>> 1) multi-process
>> 2) multi-thread
>> 3) 1 process, but somehow hook my realtime events into qt's event loop.
>>
>> For #3, my device driver has a filedescriptor, that could use select to
>> detect the interrupt (rather than blocking read call).
>>
>> #1 and #2 seem complicated.
> 

2011年9月30日 Александров Петр <pet...@ma...>
> How to make a plot of a discontinuous 1d function without line at a
> function jump. An example is shown on a picture. This was plotted by:
> import numpy
> from matplotlib import pyplot
> x = numpy.linspace(-1.0, 1.0, 100)
> y = numpy.sign(x) * numpy.cos(x)
> axes = pyplot.gca()
> axes.set_ylim(-1.1, 1.1)
> pyplot.plot(x, y)
>
>
If you happen to know where the discontinuity is, then you could do two
separate plot commands. Automatically, it is harder. Maybe something like
this would help:
threshold = 1000.0
# Use gradient instead of diff because it returns an array of the same shape
dydx = numpy.gradient(y) / numpy.gradient(x)
idx = (numpy.abs(dydx) < threshold)
axes.plot(x[idx], y[idx])
Although, that gets rid of valid data points. Another trick would be to
rebuild the x and y arrays, but with NaNs placed between datapoints that
would form the discontinuity. matplotlib automatically treats NaNs as a
indicator to break a line.
Ben Root

Hey,
On Fri, 2011年09月30日 at 17:31 +0400, Александров Петр wrote:
> How to make a plot of a discontinuous 1d function without line at a
> function jump. An example is shown on a picture. This was plotted by:
> import numpy
> from matplotlib import pyplot
> x = numpy.linspace(-1.0, 1.0, 100)
> y = numpy.sign(x) * numpy.cos(x)
> axes = pyplot.gca()
> axes.set_ylim(-1.1, 1.1)
> pyplot.plot(x, y)
> 
You can plot two lines, or use nan (or masked arrays I guess) to break
the line. in this case, this would be something like:
x = np.linspace(-1.0, 1.0, 101) # 101 as then 0 is included
s = np.sign(x)
s[s==0] = np.nan
y = s * np.cos(x)
plt.plot(x, y)
Regards,
Sebastian
> 
> ------------------------------------------------------------------------------
> All of the data generated in your IT infrastructure is seriously valuable.
> Why? It contains a definitive record of application performance, security
> threats, fraudulent activity, and more. Splunk takes this data and makes
> sense of it. IT sense. And common sense.
> http://p.sf.net/sfu/splunk-d2dcopy2
> _______________________________________________ Matplotlib-users mailing list Mat...@li... https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Le vendredi 30 septembre 2011 à 17:31 +0400, Александров Петр a écrit :
> How to make a plot of a discontinuous 1d function without line at a
> function jump. An example is shown on a picture. This was plotted by:
> import numpy
> from matplotlib import pyplot
> x = numpy.linspace(-1.0, 1.0, 100)
> y = numpy.sign(x) * numpy.cos(x)
> axes = pyplot.gca()
> axes.set_ylim(-1.1, 1.1)
> pyplot.plot(x, y)
idx = (x>0)
pyplot(x[idx],y[idx])
idx = numpy.logical_not(idx)
pyplot(x[idx],y[idx])
-- 
Fabrice Silva

Neal,
I do something similar to this where data that I'm plotting in a 2D line 
plot comes from a UDP socket and some memory mapped files. To 
accomplish the live updating I have a mix of your #2 and #3. I have a 
main GUI thread that displays the plots, then I have a second thread 
that gets data from a python generator and signals the GUI thread when 
that new data has arrived. Inside the generator is where I have socket 
communications and select calls. Hope this helps.
-Dave
On 9/30/11 9:10 AM, mat...@li... wrote:
> Message: 4
> Date: 2011年9月30日 07:37:49 -0400
> From: Neal Becker<ndb...@gm...>
> Subject: [Matplotlib-users] how would you do this (animated bargraph)
> To:mat...@li...
> Message-ID:<j649me$ug41ドル...@do...>
> Content-Type: text/plain; charset="ISO-8859-1"
>
> I just put together an animated bargraph to show results from a realtime
> process.
>
> I used this as an example:
> http://matplotlib.sourceforge.net/examples/animation/animation_blit_qt4.html
>
> The tricky part for me was that in the original design, there was a realtime
> process running. I have to write some information to my hardware (at a rate of
> about 1/360ms). To do this, I have a kernel driver that uses 'read' to tell the
> user space when an interrupt has occurred, which is when the user space should
> write new information to the hardware.
>
> So I don't know how or if this could be hooked into qt event processing. Just
> for a quick-and-dirty demo, I just removed the realtime processing from the
> user-space and put it in the kernel driver, so now my bargraph display can
> simply update on a periodic timer, and the userspace code has no realtime
> requirement. But this is just a kludge.
>
> So I wonder how other's would solve this? I'm thinking it would be either:
>
> 1) multi-process
> 2) multi-thread
> 3) 1 process, but somehow hook my realtime events into qt's event loop.
>
> For #3, my device driver has a filedescriptor, that could use select to detect
> the interrupt (rather than blocking read call).
>
> #1 and #2 seem complicated.

I just put together an animated bargraph to show results from a realtime 
process.
I used this as an example:
http://matplotlib.sourceforge.net/examples/animation/animation_blit_qt4.html
The tricky part for me was that in the original design, there was a realtime 
process running. I have to write some information to my hardware (at a rate of 
about 1/360ms). To do this, I have a kernel driver that uses 'read' to tell the 
user space when an interrupt has occurred, which is when the user space should 
write new information to the hardware.
So I don't know how or if this could be hooked into qt event processing. Just 
for a quick-and-dirty demo, I just removed the realtime processing from the 
user-space and put it in the kernel driver, so now my bargraph display can 
simply update on a periodic timer, and the userspace code has no realtime 
requirement. But this is just a kludge.
So I wonder how other's would solve this? I'm thinking it would be either:
1) multi-process
2) multi-thread
3) 1 process, but somehow hook my realtime events into qt's event loop.
For #3, my device driver has a filedescriptor, that could use select to detect 
the interrupt (rather than blocking read call).
#1 and #2 seem complicated.

Dear List,
is it possible to not only assign once color per polygon that is
plotted, but one color for each vertex, so that the result looks like
a properly smooth function. Even if I sample the points closely
enough, in the current approach in almost all of your examples you
always see the underlying grid (plus it gets very slow in 3D). I don't
really care much if you interpolate the colors on the 2D projection of
the polygon or on the real 3D polygon, just any color interpolation is
better than none.
What do you actually understand behind the "antialiasing" parameter of
34D plots. It just puzzles me to have such a concept without have a
concept of pixel shading (and thus an easy way for the above
interpolation).
I noticed that even 2D plots from matplot lib are very pixel dominated:
http://matplotlib.sourceforge.net/examples/pylab_examples/pcolor_demo.html
But there is the function imshow which doesn't have this artifacts:
http://matplotlib.sourceforge.net/examples/pylab_examples/pcolor_demo2.html
So is it possible to do something like imshow for 3D as well?
-Holger

On 2011年9月29日 14:56:21 +0200
 "Nils Wagner" <nw...@ia...> wrote:
> Hi all,
> 
> How can I add a printer button to the Navigation toolbar 
> http://matplotlib.sourceforge.net/users/navigation_toolbar.html 
> ?
> 
> A small example for QT would be awesome.
> 
> Thanks in advance.
> 
> Nils
> 
How can I adapt the example
http://matplotlib.sourceforge.net/examples/user_interfaces/printing_in_wx.html
for QT ?
Nils

Hi all,
Does anyone have a good solution for taking a line segment and coloring it
according to some gradient (where the start and end colors are the only two
things specified for the line coloring)?
Thanks!
Uri
...................................................................................
Uri Laserson
Graduate Student, Biomedical Engineering
Harvard-MIT Division of Health Sciences and Technology
M +1 917 742 8019
las...@mi...