[Matplotlib-users] Fwd: Empty squares at end of data after interpolation with griddata?
Meant to copy this to the list as well for completeness. -- Forwarded message -- From: Rick Muller rpmul...@gmail.com Date: Sun, Jul 5, 2009 at 2:03 PM Subject: Re: [Matplotlib-users] Empty squares at end of data after interpolation with griddata? To: Jeff Whitaker jsw...@fastmail.fm The link should work now. What you say about extrapolation makes perfect sense. I wonder why those points aren't being computed. In any case, thanks for the help! Rick On Sun, Jul 5, 2009 at 7:27 AM, Jeff Whitaker jsw...@fastmail.fm wrote: http://files.getdropbox.com/u/533499/griddata-example-text.png Rick: That URL doesn't work. griddata won't do extrapolation, that is it won't interpolate outside the convex hull of the data. That's probably why you see those empty squares at the edges. -- Rick Muller rpmul...@gmail.com -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
Re: [Matplotlib-users] contourf: black lines connecting contour levels?
Beautiful! The SVN version worked well for this. I forgot to check whether the workaround for version 0.98.5.3 worked; you may be right, and I may have simply forgotten to redraw the figure, but I think I did so (I normally kill my windows between plotting). Thanks very much for your help with this! On Fri, Jul 3, 2009 at 1:52 PM, Jae-Joon Lee lee.j.j...@gmail.com wrote: The example in the cookbool works fine with svn version of mpl. So, it seems that this bug has been fixed. Anyhow, which command (contour or contourf) draws the vertical lines? I bet it is contourf. And set_edgecolor(none) for return value of contourf should have some effect. Did you redraw the figure? Regards, -JJ On Thu, Jul 2, 2009 at 9:20 PM, Rick Mullerrpmul...@gmail.com wrote: Oh, and I'm using the Agg backend, I think, whatever is the default. On Thu, Jul 2, 2009 at 7:19 PM, Rick Muller rpmul...@gmail.com wrote: JJ Thanks for the tips. I had seen one of those posts whilst googling around for the bug, but discounted it because I'm not using an alpha value. Here are links to one of the cookbook examples, and one of the files that I want to plot: http://files.getdropbox.com/u/533499/griddata-test.png http://files.getdropbox.com/u/533499/silicon_donor_10_newplot.png I'm using Mac OS 10.5.7, Python 2.6.2, and MPL 0.98.5.3. I don't know which bug in the thread you were referring to. I tried the for c in CS.collections: c.set_edgecolor(none) fix, but it didn't have any effect. On Thu, Jul 2, 2009 at 4:49 PM, Jae-Joon Lee lee.j.j...@gmail.com wrote: The dropbox link is broken (you need a public url). What version of mpl and what backend are you using? There was a similar problem which has now been fixed. Try the work-around described in the thread below, and see if works. http://www.nabble.com/problems-with-contourf---alpha-td22553269.html Regards, -JJ On Thu, Jul 2, 2009 at 4:26 PM, Rick Mullerrpmul...@gmail.com wrote: When I do contourf plots in matplotlib, I get lines connecting the contour levels. This doesn't only appear to be an artifact of my plotting algorithms, it appears in this example from the matplotlib cookbook: http://www.scipy.org/Cookbook/Matplotlib/Gridding_irregularly_spaced_data at least on my mac. (I think this is the link to the output I get from that: https://dl-web.getdropbox.com/get/Photos/griddata-test.png?w=007c9af9 ) Is there a way to keep these lines from happening? If not, is there a way to turn off all of the black lines separating the contour levels? Thanks in advance, Rick -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users -- Rick Muller rpmul...@gmail.com -- Rick Muller rpmul...@gmail.com -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
[Matplotlib-users] Empty squares at end of data after interpolation with griddata?
Having one last problem with matplotlib. I have some data that I'm interpolating with griddata, and then plotting with contourf. For reasons that escape me, the upper right and the lower left squares are not being plotted. I'm printing out a 10x10 version of this to exaggerate the effect: http://files.getdropbox.com/u/533499/griddata-example-text.png In reality, I interpolate/plot this 200x200 squares, and none of it is all that noticeable. However, I'm worried that I'm doing something wrong here, and that the mistake is going to come back and bite me later on. Has anyone seen anything like this? Thanks for any help you can offer with this, and thank for all of the help the list members have already given me. Rick -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
Re: [Matplotlib-users] contourf: black lines connecting contour levels?
JJ Thanks for the tips. I had seen one of those posts whilst googling around for the bug, but discounted it because I'm not using an alpha value. Here are links to one of the cookbook examples, and one of the files that I want to plot: http://files.getdropbox.com/u/533499/griddata-test.png http://files.getdropbox.com/u/533499/silicon_donor_10_newplot.png I'm using Mac OS 10.5.7, Python 2.6.2, and MPL 0.98.5.3. I don't know which bug in the thread you were referring to. I tried the for c in CS.collections: c.set_edgecolor(none) fix, but it didn't have any effect. On Thu, Jul 2, 2009 at 4:49 PM, Jae-Joon Lee lee.j.j...@gmail.com wrote: The dropbox link is broken (you need a public url). What version of mpl and what backend are you using? There was a similar problem which has now been fixed. Try the work-around described in the thread below, and see if works. http://www.nabble.com/problems-with-contourf---alpha-td22553269.html Regards, -JJ On Thu, Jul 2, 2009 at 4:26 PM, Rick Mullerrpmul...@gmail.com wrote: When I do contourf plots in matplotlib, I get lines connecting the contour levels. This doesn't only appear to be an artifact of my plotting algorithms, it appears in this example from the matplotlib cookbook: http://www.scipy.org/Cookbook/Matplotlib/Gridding_irregularly_spaced_data at least on my mac. (I think this is the link to the output I get from that: https://dl-web.getdropbox.com/get/Photos/griddata-test.png?w=007c9af9 ) Is there a way to keep these lines from happening? If not, is there a way to turn off all of the black lines separating the contour levels? Thanks in advance, Rick -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
Re: [Matplotlib-users] contourf: black lines connecting contour levels?
Oh, and I'm using the Agg backend, I think, whatever is the default. On Thu, Jul 2, 2009 at 7:19 PM, Rick Muller rpmul...@gmail.com wrote: JJ Thanks for the tips. I had seen one of those posts whilst googling around for the bug, but discounted it because I'm not using an alpha value. Here are links to one of the cookbook examples, and one of the files that I want to plot: http://files.getdropbox.com/u/533499/griddata-test.png http://files.getdropbox.com/u/533499/silicon_donor_10_newplot.png I'm using Mac OS 10.5.7, Python 2.6.2, and MPL 0.98.5.3. I don't know which bug in the thread you were referring to. I tried the for c in CS.collections: c.set_edgecolor(none) fix, but it didn't have any effect. On Thu, Jul 2, 2009 at 4:49 PM, Jae-Joon Lee lee.j.j...@gmail.com wrote: The dropbox link is broken (you need a public url). What version of mpl and what backend are you using? There was a similar problem which has now been fixed. Try the work-around described in the thread below, and see if works. http://www.nabble.com/problems-with-contourf---alpha-td22553269.html Regards, -JJ On Thu, Jul 2, 2009 at 4:26 PM, Rick Mullerrpmul...@gmail.com wrote: When I do contourf plots in matplotlib, I get lines connecting the contour levels. This doesn't only appear to be an artifact of my plotting algorithms, it appears in this example from the matplotlib cookbook: http://www.scipy.org/Cookbook/Matplotlib/Gridding_irregularly_spaced_data at least on my mac. (I think this is the link to the output I get from that: https://dl-web.getdropbox.com/get/Photos/griddata-test.png?w=007c9af9 ) Is there a way to keep these lines from happening? If not, is there a way to turn off all of the black lines separating the contour levels? Thanks in advance, Rick -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users -- Rick Muller rpmul...@gmail.com -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
[Matplotlib-users] Help building
I've been a long time, happy user of MPL, but I'm having some problems building the latest version (0.98.5.3) on the Mac (python 2.6.2). The bottom line is that when I build, I get the error message: ld: in /Developer/SDKs/MacOSX10.4u.sdk/usr/local/lib/libPng.dylib, file is not o f required architecture for architecture ppc collect2: ld returned 1 exit status lipo: can't open input file: /var/folders/tu/tuLdYMqlEm00c5EVXrMQ+U+++TQ/-Tmp-// cci4zDq4.out (No such file or directory) error: command 'g++' failed with exit status 1 However, I have a freshly-built version of libpng installed from source in /usr/local/lib. I don't think MPL is finding it, though, since during config I see: libpng: found, but unknown version (no pkg-config) (However, when I install libpng, I do install the pc file, viz: cd /usr/local/lib/pkgconfig; rm -f libpng.pc cd /usr/local/lib/pkgconfig; ln -s libpng12.pc libpng.pc) I'm probably doing something dumb. Is there a workaround to this? Thanks in advance, Rick -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
Re: [Matplotlib-users] Help building
On Tue, Jun 30, 2009 at 9:14 AM, Michael Droettboom md...@stsci.edu wrote: What does pkg-config libpng --modversion --libs --cflags give? I guess I don't have pkg-config installed. Is this a requirement? I haven't had it installed before? The pkg-config error may actually be a red herring, since matplotlib will look in /usr/local even if pkg-config fails, and it obviously found the headers or it wouldn't have gotten to the linking stage. As the author of the pkg-config support I would like to rule that out as the problem though. I think it's more likely that the libpng isn't being built correctly to match the kind of binary matplotlib is building. I'm afraid that's where I hit the end of my knowledge as a non-Mac user, but I'm sure another Mac developer will chime in. I think it's just finding the wrong version of libpng. For whatever reason, it has decided that an old version in /Developer/SDKs/MacOSX10.4u.sdk/usr/local/lib/libPng.dylib is the proper version, despite there being a newer version in /usr/local/lib. Is there any way I can rearrange the search path so that it finds the /usr/local/lib version? Rick -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
Re: [Matplotlib-users] Help building
On Tue, Jun 30, 2009 at 9:14 AM, Michael Droettboom md...@stsci.edu wrote: What does pkg-config libpng --modversion --libs --cflags give? The pkg-config error may actually be a red herring, since matplotlib will look in /usr/local even if pkg-config fails, and it obviously found the headers or it wouldn't have gotten to the linking stage. As the author of the pkg-config support I would like to rule that out as the problem though. I think it's more likely that the libpng isn't being built correctly to match the kind of binary matplotlib is building. I'm afraid that's where I hit the end of my knowledge as a non-Mac user, but I'm sure another Mac developer will chime in. Okay, I just installed pkg-config, and it now finds the proper libpng... libpng: 1.2.37 However, I still get exactly the same compile problem: g++ -arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk -g -bundle -u ndefined dynamic_lookup build/temp.macosx-10.3-fat-2.6/src/ft2font.o build/temp. macosx-10.3-fat-2.6/src/mplutils.o build/temp.macosx-10.3-fat-2.6/CXX/cxx_extens ions.o build/temp.macosx-10.3-fat-2.6/CXX/cxxsupport.o build/temp.macosx-10.3-fa t-2.6/CXX/IndirectPythonInterface.o build/temp.macosx-10.3-fat-2.6/CXX/cxxextens ions.o -L/usr/local/lib -L/usr/lib -L/usr/X11R6/lib -lfreetype -lz -lstdc++ -lm -o build/lib.macosx-10.3-fat-2.6/matplotlib/ft2font.so ld: in /Developer/SDKs/MacOSX10.4u.sdk/usr/local/lib/libPng.dylib, file is not o f required architecture for architecture ppc collect2: ld returned 1 exit status lipo: can't open input file: /var/folders/tu/tuLdYMqlEm00c5EVXrMQ+U+++TQ/-Tmp-// ccU5pNn1.out (No such file or directory) error: command 'g++' failed with exit status 1 -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
Re: [Matplotlib-users] Help building (problems with pkg-config and libpng)
I have a working system now. Here's what I did to fix it. 1. Installed libpng from source 2. Installed freetype2 from source 3. Installed pkg-config from source 4. Installed a new version of the Xcode developer tools for the mac. During the build I still saw a few 'incompatible architecture' errors flash by, but I guess they weren't serious enough for the build to stop. On Tue, Jun 30, 2009 at 9:51 AM, Rick Muller rpmul...@gmail.com wrote: On Tue, Jun 30, 2009 at 9:14 AM, Michael Droettboom md...@stsci.eduwrote: What does pkg-config libpng --modversion --libs --cflags give? The pkg-config error may actually be a red herring, since matplotlib will look in /usr/local even if pkg-config fails, and it obviously found the headers or it wouldn't have gotten to the linking stage. As the author of the pkg-config support I would like to rule that out as the problem though. I think it's more likely that the libpng isn't being built correctly to match the kind of binary matplotlib is building. I'm afraid that's where I hit the end of my knowledge as a non-Mac user, but I'm sure another Mac developer will chime in. Okay, I just installed pkg-config, and it now finds the proper libpng... libpng: 1.2.37 However, I still get exactly the same compile problem: g++ -arch ppc -arch i386 -isysroot /Developer/SDKs/MacOSX10.4u.sdk -g -bundle -u ndefined dynamic_lookup build/temp.macosx-10.3-fat-2.6/src/ft2font.o build/temp. macosx-10.3-fat-2.6/src/mplutils.o build/temp.macosx-10.3-fat-2.6/CXX/cxx_extens ions.o build/temp.macosx-10.3-fat-2.6/CXX/cxxsupport.o build/temp.macosx-10.3-fa t-2.6/CXX/IndirectPythonInterface.o build/temp.macosx-10.3-fat-2.6/CXX/cxxextens ions.o -L/usr/local/lib -L/usr/lib -L/usr/X11R6/lib -lfreetype -lz -lstdc++ -lm -o build/lib.macosx-10.3-fat-2.6/matplotlib/ft2font.so ld: in /Developer/SDKs/MacOSX10.4u.sdk/usr/local/lib/libPng.dylib, file is not o f required architecture for architecture ppc collect2: ld returned 1 exit status lipo: can't open input file: /var/folders/tu/tuLdYMqlEm00c5EVXrMQ+U+++TQ/-Tmp-// ccU5pNn1.out (No such file or directory) error: command 'g++' failed with exit status 1 -- Rick Muller rpmul...@gmail.com -- Rick Muller rpmul...@gmail.com -- ___ Matplotlib-users mailing list Matplotlib-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/matplotlib-users
