matplotlib

On Tuesday, August 30, 2011, Katie Boyle <katielboyle@...> wrote:
> Hi All,
>
> I have a 2D grid comprised of 100 nodes in X, 300 nodes in Z, and a single
node in Y. I have a scalar value defined at each node point.
>
> I want to display this planar grid in 3D by first rendering a contour plot
(no problems there) and then rotating it such that it plots as a vertical
sheet when viewed in 3D. Right now it is automatically forced to reside in
the X-Y plane, but I want it to reside in the X-Z plane. Here is an excerpt
from my code:
>
> fig = plt.figure(figsize=(25,20))
> ax = Axes3D(fig)
> ax.contourf(xx, yy, zz, 100, cmp = plt.cm.jet,zdir='y',extend3D=False)
> plt.hold(True)
> ax.bar3d(5000, 625, 1500, 1, 1, 1, color='r')
> ax.set_xlabel("X [m]")
> ax.set_ylabel("Y [m]")
> ax.set_zlabel("Z [m]")
> plt.savefig("Feature.png",dpi=200)
>
> I think the 'zdir' flag is supposed to do this exact task, but setting it
to 'y' seems to have no effect on my plot. Also, the plot is being extended
into 3D even though I set the extend3D flag to False. Can anyone tell me
what I'm doing wrong?
>
> Cheers!
>

What version of mpl are you using?  I don't think those kwargs were added
until the upcoming release, v1.1.0.  Does the graph look right if you do
just contour?

Plus, from the top of my head, I don't think extend3d works for contourf
(but I could be wrong).

Ben Root

Hi Ben,

I'm using mpl 0.99 on Ubuntu 10.10, and whether I use contour or contourf I
get a plot that is laying in the X-Y plane with contour loops extending into
the +z direction. The contour plot itself looks right (if I could get it to
be 2D instead of 3D), it's just its orientation and non-planarity that pose
problems.

Cheers,
Katie


On Tue, Aug 30, 2011 at 5:22 PM, Benjamin Root <ben.root@...> wrote:

>
>
> On Tuesday, August 30, 2011, Katie Boyle <katielboyle@...> wrote:
> > Hi All,
> >
> > I have a 2D grid comprised of 100 nodes in X, 300 nodes in Z, and a
> single node in Y. I have a scalar value defined at each node point.
> >
> > I want to display this planar grid in 3D by first rendering a contour
> plot (no problems there) and then rotating it such that it plots as a
> vertical sheet when viewed in 3D. Right now it is automatically forced to
> reside in the X-Y plane, but I want it to reside in the X-Z plane. Here is
> an excerpt from my code:
> >
> > fig = plt.figure(figsize=(25,20))
> > ax = Axes3D(fig)
> > ax.contourf(xx, yy, zz, 100, cmp = plt.cm.jet,zdir='y',extend3D=False)
> > plt.hold(True)
> > ax.bar3d(5000, 625, 1500, 1, 1, 1, color='r')
> > ax.set_xlabel("X [m]")
> > ax.set_ylabel("Y [m]")
> > ax.set_zlabel("Z [m]")
> > plt.savefig("Feature.png",dpi=200)
> >
> > I think the 'zdir' flag is supposed to do this exact task, but setting it
> to 'y' seems to have no effect on my plot. Also, the plot is being extended
> into 3D even though I set the extend3D flag to False. Can anyone tell me
> what I'm doing wrong?
> >
> > Cheers!
> >
>
> What version of mpl are you using?  I don't think those kwargs were added
> until the upcoming release, v1.1.0.  Does the graph look right if you do
> just contour?
>
> Plus, from the top of my head, I don't think extend3d works for contourf
> (but I could be wrong).
>
> Ben Root

On Tuesday, August 30, 2011, Katie Boyle <katielboyle@...> wrote:
> Hi Ben,
>
> I'm using mpl 0.99 on Ubuntu 10.10, and whether I use contour or contourf
I get a plot that is laying in the X-Y plane with contour loops extending
into the +z direction. The contour plot itself looks right (if I could get
it to be 2D instead of 3D), it's just its orientation and non-planarity that
pose problems.
>
> Cheers,
> Katie
>

Posting an image would be very helpful.  In addition, I don't think v0.99
supports what you are trying to do.  But perhaps the image would make things
clearer.

Ben Root

On 08/30/2011 12:14 PM, Craig Finch wrote:
> I successfully built Python 2.7.2, NumPy 1.61, and Scipy 0.9.0 in my
> home directory on a Centos5 system. I am trying to build matplotlib
> 1.0.1 in my home directory, but I am having a problem with the GTK
> backends. I also built pycairo 1.2.2, pygobject 2.14.2, and pygtk 2.10.6
> in my home directory (using the libraries and headers installed in the
> default system locations). I downloaded a few pygtk examples from the
> tutorial section of their web site and ran them to verify that pygtk
> works correctly.
>
> Here is the error I get when I run a script that contains only the line
> "import matplotlib.pyplot":
>
> python2.7 test.py --verbose-helpful
> $HOME=/home/cfinch
> CONFIGDIR=/home/cfinch/.matplotlib
> matplotlib data path
> /home/cfinch/.local/lib/python2.7/site-packages/matplotlib/mpl-data
> loaded rc file
> /home/cfinch/.local/lib/python2.7/site-packages/matplotlib/mpl-data/matplotlibrc
> matplotlib version 1.0.1
> verbose.level helpful
> interactive is False
> units is False
> platform is linux2
> Using fontManager instance from /home/cfinch/.matplotlib/fontList.cache
> Traceback (most recent call last):
> File "test.py", line 1, in <module>
> import matplotlib.pyplot
> File
> "/home/cfinch/.local/lib/python2.7/site-packages/matplotlib/pyplot.py",
> line 95, in <module>
> new_figure_manager, draw_if_interactive, show = pylab_setup()
> File
> "/home/cfinch/.local/lib/python2.7/site-packages/matplotlib/backends/__init__.py",
> line 25, in pylab_setup
> globals(),locals(),[backend_name])
> File
> "/home/cfinch/.local/lib/python2.7/site-packages/matplotlib/backends/backend_gtkagg.py",
> line 10, in <module>
> from matplotlib.backends.backend_gtk import gtk, FigureManagerGTK,
> FigureCanvasGTK,\
> File
> "/home/cfinch/.local/lib/python2.7/site-packages/matplotlib/backends/backend_gtk.py",
> line 28, in <module>
> from matplotlib.backends.backend_gdk import RendererGDK, FigureCanvasGDK
> File
> "/home/cfinch/.local/lib/python2.7/site-packages/matplotlib/backends/backend_gdk.py",
> line 29, in <module>
> from matplotlib.backends._backend_gdk import pixbuf_get_pixels_array
> ImportError: No module named _backend_gdk
>
> Here is the beginning of my matplotlib build log, which shows that the
> build process is actually finding the GTK libraries:
>
> basedirlist is: ['/home/cfinch']
> ============================================================================
> BUILDING MATPLOTLIB
> matplotlib: 1.0.1
> python: 2.7.2 (default, Aug 30 2011, 12:57:00) [GCC 4.1.2
> 20080704 (Red Hat 4.1.2-50)]
> platform: linux2
>
> REQUIRED DEPENDENCIES
> numpy: 1.6.1
> freetype2: 9.10.3
> * WARNING: Could not find 'freetype2' headers in any
> * of '/home/cfinch/include', '.',
> * '/usr/include/freetype2'.

Freetype is a requirement, so my guess is that this is the origin of the 
problem, even though it is is showing up as a puzzling import error for 
_backend_gdk.

Eric


>
> OPTIONAL BACKEND DEPENDENCIES
> libpng: 1.2.10
> Tkinter: no
> * TKAgg requires Tkinter
> wxPython: no
> * wxPython not found
> Gtk+: gtk+: 2.10.4, glib: 2.12.3, pygtk: 2.10.6,
> pygobject: 2.14.2
> Mac OS X native: no
> Qt: no
> Qt4: no
> Cairo: 1.2.2
>
> OPTIONAL DATE/TIMEZONE DEPENDENCIES
> datetime: present, version unknown
> dateutil: 1.5
> pytz: 2010o
>
> OPTIONAL USETEX DEPENDENCIES
> dvipng: 1.5
> ghostscript: 8.70
> latex: 3.141592
> pdftops: 3.00
> ------------ SNIP -----------
>
> There are no errors in the build or install logs, but I can post the
> rest of it if necessary. Also, the library _backend_gdk.so is present in
> ~/lib/python2.7/site-packages/matplotlib/backends/.
>
> Any suggestions would be appreciated!
>

On Mon, Aug 29, 2011 at 10:26 PM, Trevor J Christensen <
trevor@...> wrote:

> I have a number of questions, but I'll start with one.  Consider the
> following code:
>
> import matplotlib as mpl
> import matplotlib.pyplot as plt
> import numpy as np
> x = np.arange(1, 6)
> plt.plot(x, x*1.5)
> plt.show()
> s = raw_input('Enter anything to continue:')
> plt.show()
> print 'The end'
>
> Question:  The first plt.show() launches an interactive window.  Control
> returns to the script upon its close.  The second plt.show() does nothing
> that I can see.  Why?
>
> Ubuntu 11.04, Python 2.7.1, Matplotlib 0.99.3
>


That is because closing the figure destroys that object.  If you want full
interactive mode, I suggest updating to v1.0.1 (or the upcoming v1.1.0
release) and use plt.ion().  The show() will trigger the figures to appear,
but control immediately returns to the script.

Interactivity in earlier versions of mpl is a bit more hit-or-miss, so you
can try it out first before updating.  Also note that multiple calls to
plt.show() in earlier versions are not recommended, but are allowed in v1.0
and up.

I hope this helps,
Ben Root

How about Numpy's loadtxt()?  Genfromtext is more of a specialty function.

Ben Root

On Monday, August 29, 2011, Alexa Villaume <alexa7890@...> wrote:
> Hi there,
> I have a 2d array that I'd like to contour. The plot isn't turning out
right and I think it's because the array isn't be read in correctly.
> It's a 25x41 array that is in a text file. I have tried using genfromtxt
and the asciitable module to read the data in and neither gives the correct
output.
> Does anyone know of any command either specific for reading in 2d arrays
or is able to handle 2d arrays?
> Thanks,
> Alexa

On Sun, Aug 28, 2011 at 3:03 PM, Bruce Ford <bruce@...:

> Getting a strange result trying to divide two 3d arrays.  I am getting
> a matrix of NaNs regardless of how I divide and I can't determine why.
>
> #opened a NetCDF file using python-netcdf4
>
> var1 = nc_file.variables['var1']  ###shape = [31,181,360] with a
> values ranging from 0 - 243 and NO NaNs in the array, dtype float32
> var2 = nc_file.variables['var2']  ###shape = [31,181,360] with a
> values ranging from 0 - 4 (mostly zeros) and NO NaNs in the array,
> dtype float32
>
> np.seterr(all='ignore')  #in case problem has something do to with
> dividing by zero
>
> var1/var2  ###gives array of NaNs with shape of 31,181,360
>
> #doing the division one slice at a time doesn't help...
> for x in range(1,var1.shape[0]):
>            var[x,:,:] = var1[x,:,:]/var2[x,:,:]  ###gives array of
> NaNs with shape of 31,181,360
>
> var = np.divide(var1,var2)  ###gives array of NaNs with shape of 31,181,360
>
>
> print "<p>where max: " + np.where(var1 == np.max(var1))  #prints
> (array([28]), array([79]), array([182]))
> print var1[28,79,182]  #print 545
> print var2[28,79,182]   ##prints 6
>
> #so there are values in this location that should not result in an
> NaN.  Instead I get an entire array of NaNs
>
> What am I missing?
>
> Bruce
>
>
Just to be sure, what might appear to be all NaNs is not really.  To find
out for sure, do this:

print np.any(np.isfinite(var1/var2))

If it is True, then the array is not full of NaNs and Infs, if False, then
that would be odd.  When you print an array to the screen, it will often
only print part of the array.

Furthermore, the author of the post that you link to in your other email
misses some important facts about representation of numbers in computers.
The reason 0 / 0 (integers) did not return a NaN or 2 / 0 return an Inf is
that integers can not represent those values.  Only floating point numbers
can.  Furthermore, there is an important difference between Inf and NaN.  It
is merely out of convenience that nan_to_num() would convert the Infs into
the largest (or smallest for -Inf) floating point number.  Other functions
designed for NaN (such as nansum(), nanmax(), etc) do not treat Infs in this
manner.

Ben Root

On Mon, Aug 29, 2011 at 2:56 PM, Mark Bakker <markbak@...> wrote:
> Hello List,
>
> Does anybody know of any plans to include support for the new datetime64
> data type in numpy? If this is the new numpy standard for doing dates and
> times, it would be great if it would work with plot_date, for example.
>
> Just wondering (but boy, would I do a little dance when all this datetime
> stuff is fully operational and integrated),


Should be fairly straightforward to write a converter.  Just follow
the example of DateConverter in matplotlib.dates
:https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/dates.py:

class DateConverter(units.ConversionInterface):
    """
Converter for datetime.date and datetime.datetime data,
or for date/time data represented as it would be converted
by :func:`date2num`.

The 'unit' tag for such data is None or a tzinfo instance.
"""

    @staticmethod
    def axisinfo(unit, axis):
        """
Return the :class:`~matplotlib.units.AxisInfo` for *unit*.

*unit* is a tzinfo instance or None.
The *axis* argument is required but not used.
"""
        tz = unit

        majloc = AutoDateLocator(tz=tz)
        majfmt = AutoDateFormatter(majloc, tz=tz)
        datemin = datetime.date(2000, 1, 1)
        datemax = datetime.date(2010, 1, 1)

        return units.AxisInfo( majloc=majloc, majfmt=majfmt, label='',
                               default_limits=(datemin, datemax))

    @staticmethod
    def convert(value, unit, axis):
        """
If *value* is not already a number or sequence of numbers,
convert it with :func:`date2num`.

The *unit* and *axis* arguments are not used.
"""
        if units.ConversionInterface.is_numlike(value):
            return value
        return date2num(value)

    @staticmethod
    def default_units(x, axis):
        'Return the tzinfo instance of *x* or of its first element, or None'
        try:
            x = x[0]
        except (TypeError, IndexError):
            pass

        try:
            return x.tzinfo
        except AttributeError:
            pass
        return None


units.registry[datetime.date] = DateConverter()
units.registry[datetime.datetime] = DateConverter()

Okay, so using cm.set_over() and set_under works, but is there any way of
adding the set colors to the edges of the colorbar?

On 29 August 2011 15:47, James McMurray <jamesmcm03@...> wrote:

> Hi,
>
> Is it possible to color values outside the colorbar limits in a different
> color than the colorbar itself. So for example, an otherwise greyscale plot
> could have the values outside the upperlimit in blue and those below the
> lower limit in red, so that these points are not confused with the
> legitimate data points (as they can be when they are just more black and
> white points as default).
>
> Thanks,
> James McMurray
>

On Mon, Aug 29, 2011 at 9:59 AM, James McMurray <jamesmcm03@...:

> Okay, so using cm.set_over() and set_under works, but is there any way of
> adding the set colors to the edges of the colorbar?
>
>
>
This example should do what you want:

http://matplotlib.sourceforge.net/examples/api/colorbar_only.html

I hope that helps!
Ben Root

On 08/29/2011 05:44 AM, Benjamin Root wrote:
> On Mon, Aug 29, 2011 at 9:59 AM, James McMurray <jamesmcm03@...
> <mailto:jamesmcm03@...>> wrote:
>
>     Okay, so using cm.set_over() and set_under works, but is there any
>     way of adding the set colors to the edges of the colorbar?
>
>
>
> This example should do what you want:
>
> http://matplotlib.sourceforge.net/examples/api/colorbar_only.html

Also 
http://matplotlib.sourceforge.net/examples/pylab_examples/contourf_demo.html.

Note that "extend=both" is the relevant kwarg.  With contourf, the kwarg 
goes in the contourf call, and colorbar automatically gets it from the 
ContourSet. With other types of color mappable plot, the kwarg needs to 
be given directly to the colorbar call.

Eric

>
> I hope that helps!
> Ben Root
>
>
>
> ------------------------------------------------------------------------------
> EMC VNX: the world's simplest storage, starting under $10K
> The only unified storage solution that offers unified management
> Up to 160% more powerful than alternatives and 25% more efficient.
> Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
>
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users@...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On Mon, Aug 29, 2011 at 12:30 PM, Eric Firing <efiring@...> wrote:

> On 08/29/2011 05:44 AM, Benjamin Root wrote:
> > On Mon, Aug 29, 2011 at 9:59 AM, James McMurray <jamesmcm03@...
> > <mailto:jamesmcm03@...>> wrote:
> >
> >     Okay, so using cm.set_over() and set_under works, but is there any
> >     way of adding the set colors to the edges of the colorbar?
> >
> >
> >
> > This example should do what you want:
> >
> > http://matplotlib.sourceforge.net/examples/api/colorbar_only.html
>
> Also
>
> http://matplotlib.sourceforge.net/examples/pylab_examples/contourf_demo.html
> .
>
> Note that "extend=both" is the relevant kwarg.  With contourf, the kwarg
> goes in the contourf call, and colorbar automatically gets it from the
> ContourSet. With other types of color mappable plot, the kwarg needs to
> be given directly to the colorbar call.
>
> Eric
>
>
Crap! Is that why my colorbars from my pcolors (or was it pcolormesh?) never
worked right?  I have been making colorbars directly for a while now because
I couldn't figure out why those colorbars didn't work right.

I think I will file a feature request to fix this inconsistency.  Probably
will be a bit involved to fix it, though.

Ben Root

On 08/29/2011 07:40 AM, Benjamin Root wrote:
>
>
> On Mon, Aug 29, 2011 at 12:30 PM, Eric Firing <efiring@...
> <mailto:efiring@...>> wrote:
>
>     On 08/29/2011 05:44 AM, Benjamin Root wrote:
>      > On Mon, Aug 29, 2011 at 9:59 AM, James McMurray
>     <jamesmcm03@... <mailto:jamesmcm03@...>
>      > <mailto:jamesmcm03@... <mailto:jamesmcm03@...>>> wrote:
>      >
>      >     Okay, so using cm.set_over() and set_under works, but is
>     there any
>      >     way of adding the set colors to the edges of the colorbar?
>      >
>      >
>      >
>      > This example should do what you want:
>      >
>      > http://matplotlib.sourceforge.net/examples/api/colorbar_only.html
>
>     Also
>     http://matplotlib.sourceforge.net/examples/pylab_examples/contourf_demo.html.
>
>     Note that "extend=both" is the relevant kwarg.  With contourf, the kwarg
>     goes in the contourf call, and colorbar automatically gets it from the
>     ContourSet. With other types of color mappable plot, the kwarg needs to
>     be given directly to the colorbar call.
>
>     Eric
>
>
> Crap! Is that why my colorbars from my pcolors (or was it pcolormesh?)
> never worked right?  I have been making colorbars directly for a while
> now because I couldn't figure out why those colorbars didn't work right.
>
> I think I will file a feature request to fix this inconsistency.
> Probably will be a bit involved to fix it, though.

It's not entirely an inconsistency.  The logic is this: extend='both' 
(or whatever) directly affects the levels used in contourf; there is 
nothing similar in imshow or pcolor*.  Adding the extend kwarg to them 
would simply be putting it in place for later use by colorbar.

Eric

>
> Ben Root
>
>
>
> ------------------------------------------------------------------------------
> EMC VNX: the world's simplest storage, starting under $10K
> The only unified storage solution that offers unified management
> Up to 160% more powerful than alternatives and 25% more efficient.
> Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
>
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users@...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users