DOC: Correct subplot() doc #7760
Conversation
| Creating a new subplot with a position which is entirely inside a | ||
| pre-existing axes will trigger the larger axes to be deleted:: | ||
| Creating a new subplot with axes that overlap pre-existing axes will | ||
| trigger the pre-existing axes to be deleted:: |
There was a problem hiding this comment.
Something like "Creating a subplot will delete any pre-existing subplot that overlaps with it." seems clearer (possibly add "(if any)" at the end).
|
Good suggestion @anntzer |
Current coverage is 62.11% (diff: 100%)@@ master #7760 diff @@
==========================================
Files 174 174
Lines 56028 56028
Methods 0 0
Messages 0 0
Branches 0 0
==========================================
- Hits 34805 34803 -2
- Misses 21223 21225 +2
Partials 0 0
|
|
lgtm |
| @@ -976,8 +976,8 @@ def subplot(*args, **kwargs): | |||
|
|
|||
| .. note:: | |||
|
|
|||
| Creating a new subplot with a position which is entirely inside a | |||
| pre-existing axes will trigger the larger axes to be deleted:: | |||
| Creating a subplot will delete any pre-existing subplot that overlaps | |||
|
Thanks @petehuang |
|
I challenge anyone reading the proposed modification of the docstring, without poring over the code, to explain what "fully overlaps" means. Before reading the code, I was imagining that it might mean "encloses", as the original form of the docstring implied. Based on the Bbox method code, however, it means that the interiors intersect--that is, the intersection of the two rectangles encloses a finite area. Hence the subplot code is deleting any pre-existing subplots that have a finite intersection with the new subplot. Coinciding edges are not enough, because then the intersection encloses no area. |
|
Ok let me paraphrase to make sure I understand correctly: you're saying that it's not determined by axes lying over one another, but it's that the rectangles that the axes create intersect in some way? I just want to make sure we get this explanation and the docstring in plain english |
|
The practical problem is coming from the adverb "fully". The "overlaps" part is fine. The subtle distinction is that a shared boundary does not count as an "overlap". You could change "that fully overlaps with it" to "that has a finite area of overlap with it". Or "that overlaps with it beyond sharing a boundary". |
|
I knew what the function did (and I find this behavior extremely strange…). |
|
Thank you! |
|
Backported to |
Fixes #7393