Repository navigation
allow kwargs transform to work for axvspan, axhspan, axvline, axhline - #91
kaufman wants to merge 323 commits into
Conversation
Sorry, something went wrong.
…n Granger svn path=/trunk/matplotlib/; revision=8664
…tion draw was putting data in the saved sequence, which needed to be removed. Also, handle the case where there is no saved sequence. Make sure to disconnect signals for first draw, otherwise saving the movie will start the animation. svn path=/trunk/matplotlib/; revision=8665
svn path=/trunk/matplotlib/; revision=8666
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8652 | efiring | 2010-08-21 13:49:53 -0400 (Sat, 21 Aug 2010) | 2 lines Axis.set_view_interval: when updating, respect original orientation ........ r8667 | leejjoon | 2010-08-28 11:54:32 -0400 (Sat, 28 Aug 2010) | 1 line fix gridspec bug that location is wrongly calculated for non-square grid ........ svn path=/trunk/matplotlib/; revision=8668
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8673 | efiring | 2010-09-01 09:44:36 -1000 (Wed, 01 Sep 2010) | 2 lines [3054467] fix bug in PatchCollection; restore attribute access to Patch.fill ........ svn path=/trunk/matplotlib/; revision=8674
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8675 | efiring | 2010-09-01 13:39:49 -1000 (Wed, 01 Sep 2010) | 2 lines PatchCollection: handle linestyles correctly; patch by Thomas Robitaille ........ svn path=/trunk/matplotlib/; revision=8676
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8677 | efiring | 2010-09-04 11:17:58 -1000 (Sat, 04 Sep 2010) | 2 lines doc patch by Christoph Gohlke ........ svn path=/trunk/matplotlib/; revision=8678
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8679 | efiring | 2010-09-04 14:38:45 -1000 (Sat, 04 Sep 2010) | 2 lines [3024007] Fix ancient bug in hist inherited from numpy, fixed in numpy 1.5 ........ svn path=/trunk/matplotlib/; revision=8680
The changeset includes refactoring to consolidate all Axes tracking in a single data structure. svn path=/trunk/matplotlib/; revision=8681
svn path=/trunk/matplotlib/; revision=8682
svn path=/trunk/matplotlib/; revision=8693
svn path=/trunk/matplotlib/; revision=8695
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8690 | mdboom | 2010-09-08 09:32:17 -0500 (Wed, 08 Sep 2010) | 2 lines [3058804] part of a line is not drawn ........ r8691 | mdboom | 2010-09-08 10:51:18 -0500 (Wed, 08 Sep 2010) | 2 lines Another pass at fixing simplification. ........ r8696 | ryanmay | 2010-09-11 14:02:38 -0500 (Sat, 11 Sep 2010) | 1 line Fix typo in griddata(). ........ r8697 | ryanmay | 2010-09-11 14:03:52 -0500 (Sat, 11 Sep 2010) | 1 line Handle multple calls to GTKTimer.start() by stopping any existing gtk timer id. ........ svn path=/trunk/matplotlib/; revision=8698
https://matplotlib.svn.sf.net/svnroot/matplotlib/branches/v1_0_maint ........ r8699 | mdboom | 2010-09-14 11:21:21 -0400 (Tue, 14 Sep 2010) | 2 lines Fix semi-transparent Gouraud triangle rendering in SVG backend. ........ r8700 | mdboom | 2010-09-14 11:53:59 -0400 (Tue, 14 Sep 2010) | 2 lines Fix mismatched opening/closing group. ........ svn path=/trunk/matplotlib/; revision=8701
svn path=/trunk/matplotlib/; revision=8702
svn path=/trunk/matplotlib/; revision=8703
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8704 | weathergod | 2010-09-16 11:45:43 -0500 (Thu, 16 Sep 2010) | 3 lines change pointer offset from type int to size_t, which caused problems with showing very high res images. This fixes bug ID 3054444. ........ svn path=/trunk/matplotlib/; revision=8705
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8706 | efiring | 2010-09-16 16:46:04 -1000 (Thu, 16 Sep 2010) | 5 lines include Python.h first in extension code; patch by Jason Grout. This cleans up the preprocessor directives and allows compilation on Solaris. A preprocessor directive is added to keep png.h from failing on Linux because Python.h has caused setjmp.h to be included. ........ svn path=/trunk/matplotlib/; revision=8707
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8708 | efiring | 2010-09-16 20:41:34 -1000 (Thu, 16 Sep 2010) | 2 lines including Python.h and png.h: a second try, special case for linux ........ svn path=/trunk/matplotlib/; revision=8709
svn path=/trunk/matplotlib/; revision=8710
https://matplotlib.svn.sf.net/svnroot/matplotlib/branches/v1_0_maint ........ r8711 | mdboom | 2010-09-20 12:58:14 -0400 (Mon, 20 Sep 2010) | 2 lines Fix clipped text in example. ........ r8712 | mdboom | 2010-09-21 16:13:25 -0400 (Tue, 21 Sep 2010) | 2 lines If a font file is looked up in the cache, but that font file no longer exists on disk, rebuild the cache. ........ svn path=/trunk/matplotlib/; revision=8713
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8714 | mdboom | 2010-09-21 15:18:48 -0500 (Tue, 21 Sep 2010) | 2 lines Bug # ........ r8715 | jdh2358 | 2010-09-22 07:46:38 -0500 (Wed, 22 Sep 2010) | 1 line added Bartosz Telenczuk's artist demo ........ svn path=/trunk/matplotlib/; revision=8716
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8717 | jdh2358 | 2010-09-24 04:54:34 +0900 (Fri, 24 Sep 2010) | 1 line added citation faq ........ r8718 | jdh2358 | 2010-09-24 05:07:12 +0900 (Fri, 24 Sep 2010) | 1 line wrap the mpl citation abstract ........ r8720 | leejjoon | 2010-09-27 10:30:19 +0900 (Mon, 27 Sep 2010) | 1 line fix bezier.get_parallerls to handle parallel input lines ........ svn path=/trunk/matplotlib/; revision=8721
https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/branches/v1_0_maint ........ r8722 | efiring | 2010-09-27 08:51:05 -1000 (Mon, 27 Sep 2010) | 2 lines Clarify docstrings for autoscale_view and relim. ........ r8723 | efiring | 2010-10-03 10:45:38 -1000 (Sun, 03 Oct 2010) | 2 lines docs: in event_handling, note that weak references to callbacks are used. ........ r8724 | efiring | 2010-10-03 11:04:10 -1000 (Sun, 03 Oct 2010) | 2 lines errorbar bugfix: plot everything regardless of the Axes hold state ........ r8725 | efiring | 2010-10-03 11:17:39 -1000 (Sun, 03 Oct 2010) | 2 lines bugfix: applied patch by Stan West to fix error in colormap reversal ........ svn path=/trunk/matplotlib/; revision=8726
…G files. svn path=/trunk/matplotlib/; revision=8727
svn path=/trunk/matplotlib/; revision=8728
svn path=/trunk/matplotlib/; revision=8729
…de to make contains() not cause a traceback if draw() has not been called. svn path=/trunk/matplotlib/; revision=8730
svn path=/trunk/matplotlib/; revision=8731
svn path=/trunk/matplotlib/; revision=8733
…eded by the bug tracker.
…ot_directive Conflicts: lib/matplotlib/sphinxext/plot_directive.py
…nor_doc_fixes Conflicts: lib/matplotlib/sphinxext/plot_directive.py
…into mdboom-minor_doc_fixes Conflicts: .gitignore CHANGELOG doc/faq/installing_faq.rst doc/sphinxext/inheritance_diagram.py doc/users/recipes.rst
modified: axes.py
There was a problem hiding this comment.
My preferred idiom would be:
kwargs.setdefault('transform', trans)
Then you don't need the comment above it. If this line immediately follows the definition of trans, then you don't need the comment above that, either.
Sorry, something went wrong.
There was a problem hiding this comment.
right, I guess that makes it obvious that I'm still a python newbie.
Sorry, something went wrong.
|
Apart from the stylistic suggestion in the line note, I am ambivalent about the change. The essence of these methods is in their blended transform; they exist to provide easy access to that advanced feature. This is stated right at the start of the docstring. Allowing a transform kwarg subverts that essence. Is the motivation that you want to be able to give the both sets of coordinates in axes units? In the spirit of these methods as convenience methods, hiding the transform complexity from the user, you could do that with a kwarg, e.g., On the positive side, I can see taking the point of view that the change is harmless because anyone who puts in a transform kwarg presumably understands how it is overriding the normal behavior; and anyone who does not put in the kwarg will never know the difference. Suitably condensed, the change slightly simplifies the code while expanding its capabilities. Now I see another problem, however: the scalex,scaley logic is based on the assumption that the default transform is being used. Yes, you could fix that; but it seems to me like just one more indication that it would be better to leave the methods alone, or possibly use the kwarg approach. In other words, there is a good reason these methods disallow passing an arbitrary transform. |
Sorry, something went wrong.
This patch removes the prohibition on using the 'transform' option for axhline and axvline and changes the code order so that a user-passed transform option will work with axhspan and axvspan.