matplotlib-users Mailing List for matplotlib

Ah... That was not clear. I just retried my first example (with
show->savefig) for all of the backends that I have available: Qt4Agg,
TkAgg, PS, PDF, pgf, Cairo, GTK3Cairo, GTK3Agg. All of the *Agg backends
show the same problem: the other backends work as expected. I will file a
bug report now.

Ryan

On Tue, Dec 9, 2014 at 9:50 AM, Benjamin Root <ben...@ou...> wrote:

> Interesting. Just to double-check, when you say that it only happens for
> the "agg" backend, are you saying that backends like "tkagg" are
> unaffected? I think at this point there is enough information here to file
> a bug report.
>
> Ben Root
>
> On Mon, Dec 8, 2014 at 6:14 PM, Ryan Nelson <rne...@gm...> wrote:
>
>> Final update.
>>
>> I've done some more searching, and found a couple more things. It seems
>> that this problem occurs with the backend set to "Agg"
>> (`matplotlib.use("agg")), so it isn't related to the interactive backends.
>> In addition, the problem does not occur with a random Polygon object added
>> to an axes; however, I do see the problem when the same polygon is added to
>> the axes as a PolyCollection. See code below.
>>
>> Ryan
>>
>> #####
>> import numpy as np
>> import matplotlib
>> matplotlib.use("Agg")
>> import matplotlib.pyplot as plt
>> from matplotlib.collections import PolyCollection
>>
>> x = np.linspace(0, np.pi*2, 1000)
>> y = np.sin(x)
>>
>> ax = plt.axes()
>> data =np.array([(0,0), (1,0), (1,1), (0,1)])
>>
>> # These three lines work fine.
>> poly = plt.Polygon(data)
>> poly.set_linewidth(0)
>> ax.add_patch(poly)
>>
>> # Comment out the three lines above
>> # Uncomment next three lines, does not work.
>> #col = PolyCollection([data])
>> #col.set_linewidth(0.0)
>> #ax.add_collection(col)
>>
>> plt.axis([-2, 2, -2, 2])
>> plt.savefig('junk')
>> #####
>>
>> On Mon, Dec 8, 2014 at 5:02 PM, Ryan Nelson <rne...@gm...>
>> wrote:
>>
>>> Update 2.
>>>
>>> I made a new Anaconda Python 2.7 environment and cycled through some
>>> different MPL versions. Everything works as I would expect in 1.4.0;
>>> however, moving to 1.4.1 is when the problem occurs. I see this same
>>> problem if I do the OO commands instead of pyplot.
>>>
>>> import numpy as np
>>> import matplotlib.pyplot as plt
>>>
>>> x = np.linspace(0, np.pi*2, 1000)
>>> y = np.sin(x)
>>>
>>> ax = plt.axes()
>>> fill = ax.fill_between(x, y-0.1, y+0.1)
>>> fill.set_linewidth(0)
>>>
>>> plt.show()
>>>
>>> On Mon, Dec 8, 2014 at 3:38 PM, Ryan Nelson <rne...@gm...>
>>> wrote:
>>>
>>>> Update.
>>>>
>>>> This is a problem also in Anaconda Py3.4 with MPL 1.4.2, but it works
>>>> without a problem on MPL 1.4.0.
>>>>
>>>> Ryan
>>>>
>>>> On Mon, Dec 8, 2014 at 12:15 PM, Ryan Nelson <rne...@gm...>
>>>> wrote:
>>>>
>>>>> Hello all,
>>>>>
>>>>> I'm having an issue with fill_between. It seems that setting the
>>>>> keyword `linewidth=0` removes the entire patch, rather than the just the
>>>>> bounding lines. Example:
>>>>>
>>>>> ####
>>>>> import numpy as np
>>>>> import matplotlib.pyplot as plt
>>>>>
>>>>> x = np.linspace(0, 2*np.pi, 1000)
>>>>> y = np.sin(x)
>>>>>
>>>>> plt.fill_between(x, y-0.1, y+0.1, linewidth=0) # Setting this !=0
>>>>> works fine
>>>>> plt.plot(x, y, 'k')
>>>>> plt.show()
>>>>> ####
>>>>>
>>>>> I'm using MPL version 1.4.2 on Python 2.7.8 (Gentoo Linux). This used
>>>>> to work fine before, but maybe there is a new way to do what...
>>>>>
>>>>> Thanks
>>>>>
>>>>> Ryan
>>>>>
>>>>>
>>>>>
>>>>
>>>
>>
>>
>> ------------------------------------------------------------------------------
>> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
>> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
>> with Interactivity, Sharing, Native Excel Exports, App Integration & more
>> Get technology previously reserved for billion-dollar corporations, FREE
>>
>> http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>>
>

Interesting. Just to double-check, when you say that it only happens for
the "agg" backend, are you saying that backends like "tkagg" are
unaffected? I think at this point there is enough information here to file
a bug report.

Ben Root

On Mon, Dec 8, 2014 at 6:14 PM, Ryan Nelson <rne...@gm...> wrote:

> Final update.
>
> I've done some more searching, and found a couple more things. It seems
> that this problem occurs with the backend set to "Agg"
> (`matplotlib.use("agg")), so it isn't related to the interactive backends.
> In addition, the problem does not occur with a random Polygon object added
> to an axes; however, I do see the problem when the same polygon is added to
> the axes as a PolyCollection. See code below.
>
> Ryan
>
> #####
> import numpy as np
> import matplotlib
> matplotlib.use("Agg")
> import matplotlib.pyplot as plt
> from matplotlib.collections import PolyCollection
>
> x = np.linspace(0, np.pi*2, 1000)
> y = np.sin(x)
>
> ax = plt.axes()
> data =np.array([(0,0), (1,0), (1,1), (0,1)])
>
> # These three lines work fine.
> poly = plt.Polygon(data)
> poly.set_linewidth(0)
> ax.add_patch(poly)
>
> # Comment out the three lines above
> # Uncomment next three lines, does not work.
> #col = PolyCollection([data])
> #col.set_linewidth(0.0)
> #ax.add_collection(col)
>
> plt.axis([-2, 2, -2, 2])
> plt.savefig('junk')
> #####
>
> On Mon, Dec 8, 2014 at 5:02 PM, Ryan Nelson <rne...@gm...> wrote:
>
>> Update 2.
>>
>> I made a new Anaconda Python 2.7 environment and cycled through some
>> different MPL versions. Everything works as I would expect in 1.4.0;
>> however, moving to 1.4.1 is when the problem occurs. I see this same
>> problem if I do the OO commands instead of pyplot.
>>
>> import numpy as np
>> import matplotlib.pyplot as plt
>>
>> x = np.linspace(0, np.pi*2, 1000)
>> y = np.sin(x)
>>
>> ax = plt.axes()
>> fill = ax.fill_between(x, y-0.1, y+0.1)
>> fill.set_linewidth(0)
>>
>> plt.show()
>>
>> On Mon, Dec 8, 2014 at 3:38 PM, Ryan Nelson <rne...@gm...>
>> wrote:
>>
>>> Update.
>>>
>>> This is a problem also in Anaconda Py3.4 with MPL 1.4.2, but it works
>>> without a problem on MPL 1.4.0.
>>>
>>> Ryan
>>>
>>> On Mon, Dec 8, 2014 at 12:15 PM, Ryan Nelson <rne...@gm...>
>>> wrote:
>>>
>>>> Hello all,
>>>>
>>>> I'm having an issue with fill_between. It seems that setting the
>>>> keyword `linewidth=0` removes the entire patch, rather than the just the
>>>> bounding lines. Example:
>>>>
>>>> ####
>>>> import numpy as np
>>>> import matplotlib.pyplot as plt
>>>>
>>>> x = np.linspace(0, 2*np.pi, 1000)
>>>> y = np.sin(x)
>>>>
>>>> plt.fill_between(x, y-0.1, y+0.1, linewidth=0) # Setting this !=0 works
>>>> fine
>>>> plt.plot(x, y, 'k')
>>>> plt.show()
>>>> ####
>>>>
>>>> I'm using MPL version 1.4.2 on Python 2.7.8 (Gentoo Linux). This used
>>>> to work fine before, but maybe there is a new way to do what...
>>>>
>>>> Thanks
>>>>
>>>> Ryan
>>>>
>>>>
>>>>
>>>
>>
>
>
> ------------------------------------------------------------------------------
> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
> with Interactivity, Sharing, Native Excel Exports, App Integration & more
> Get technology previously reserved for billion-dollar corporations, FREE
>
> http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

Hello,

I am experiencing strange behavior using mpl_toolkits.basemap.Basemap.
My understanding is that pcolormesh is faster than pcolor, and thus
preferable.  Here is a minimal example (below) where I get a different
plot from pcolormesh than from pcolor.

On two systems (mac os x 10.9.5; Ubuntu 11.04 (GNU/Linux 2.6.32.28
x86_64); basemap 1.0.7 on the mac and 1.0.8 on the Ubuntu machine; both
using matplotlib 1.4.2) I get the expected grid of random colors from
pcolor, but a monochrome plot from pcolormesh.

Is this expected?  Are there circumstances in which I should be using
pcolor instead of pcolormesh?  Is this a bug in matplotlib or in
basemap?

Many thanks!

Tim

--

Timothy W. Hilton
Assistant Project Scientist
School of Engineering
University of California, Merced
th...@uc... 

#==================================================
# minimal example

import matplotlib.pyplot as plt
from mpl_toolkits.basemap import Basemap
import numpy as np

def setup_map(ax):

    m = Basemap(width=8.0e6,
                height=6.5e6,
                projection='aeqd',
                lat_0=54,
                lon_0=-105,
                resolution='l',
                area_thresh=1000, 
                rsphere=6371007.181000,
                fix_aspect=True,
                ax=ax)
    m.drawcoastlines()
    m.drawcountries()
    m.drawstates()
    return(m)

# create pseudo-data with longitudes and latitudes
lon, lat = np.meshgrid(np.arange(-180, 180), 
                       np.arange(90, -90, -1))
data = np.random.rand(*lon.shape) * 100

# two-panel figure
fig, ax = plt.subplots(nrows=1, ncols=2)

# plot pseoddata in left panel using pcolormesh
m0 = setup_map(ax[0])
cm = m0.pcolormesh(lon, lat, data, latlon=True, 
              vmin=0.0, vmax=100.0, cmap=plt.cm.get_cmap("Blues"))
plt.colorbar(cm, ax=ax[0])
ax[0].set_title('pcolormesh')

# plot pseoddata in left panel using pcolor
m1 = setup_map(ax[1])
cm = m1.pcolor(lon, lat, data, latlon=True, 
          vmin=0.0, vmax=100.0, cmap=plt.cm.get_cmap("Blues"))
plt.colorbar(cm, ax=ax[1])
ax[1].set_title('pcolor')

plt.show()