ENH have ax.get_tightbbox have a bbox around all artists attached to axes. by jklymak · Pull Request #10682 · matplotlib/matplotlib · GitHub

PR Summary

This PR takes the code that was in backend_bases.py in savefig for determining a figure bounding box, and puts it in ax.get_tightbbox and fig.get_tightbbox, via a refactoring into Artists.get_tightbbox.

Supercedes #10678 and extends already merged #9164

In previous PRs I was piecemeal adding new types of artists, but noted that fig.savefig(bbox_inches='tight') was actually doing the correct thing and putting the bbox around all the artists.

To get the old behaviour back, a new kwarg has been added:

ax.get_tightbbox(bbox_extra_artists=[])

So now the following code

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(5,2))
ax.set_xlim([0, 1])
ax.text(1, 0.5, 'This is a long overflow')
print('before', ax.get_tightbbox(fig.canvas.get_renderer(),bbox_extra_artists=[]))
print('now   ', ax.get_tightbbox(fig.canvas.get_renderer()))

returns:

before Bbox(x0=61.2, y0=-3.4, x1=922.125, y1=363.0)
now    Bbox(x0=61.2, y0=-3.4, x1=1210.25, y1=363.0)

This is a breaking change, in that anyone who was expecting the bbox to not include artists, but just the axes axis elements and labels, will now get a different behaviour. They can get the old behaviour as above (specifying the empty bbox_extra_artists kwarg. OTOH, willing be told this is too big an API change. (it of course still needs an API change note, and likely a test or two somewhere).

UPDATE 28 April:

This adds the artist property artist.set/get_in_layout (as discussed w/ @efiring et al on the weekly call) which specifies if an artist should be included in the tight_bbox calculation. This allows toggling artists in the tight layout or constrained layout calculations.

So, now we can also do:

fig, ax = plt.subplots(figsize=(5,2))
ax.set_xlim([0, 1])
t = ax.text(1, 0.5, 'This is a long overflow')
t.inbbox = True  # default
print('inbbox true', ax.get_tightbbox(fig.canvas.get_renderer()))
t.set_in_layout(False)
print('inbbox false ', ax.get_tightbbox(fig.canvas.get_renderer()))
t.set_in_layout(True)
print('inbbox true ', ax.get_tightbbox(fig.canvas.get_renderer()))

and get

in_layout  true  Bbox(x0=61.18055555555556, y0=-3.4444444444444535, x1=1210.25, y1=363.0)
in_layout false  Bbox(x0=61.18055555555556, y0=-3.4444444444444535, x1=922.125, y1=363.0)
in_layout  true  Bbox(x0=61.18055555555556, y0=-3.4444444444444535, x1=1210.25, y1=363.0)

Note that this also works for legends, which often get put outside axes and hence may not want to be part of the tight_layout machinery.

Discussion: Should artists that did not get included in the old tight_layout be set to False by default? I'm somewhat against this because it means users have to guess or query what state the attribute is in...

PR Checklist

Has Pytest style unit tests
Code is PEP 8 compliant
Documentation is sphinx and numpydoc compliant
Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way

Is this meant to be used by tight_layout, such that one could call fig.tight_layout(bbox_extra_artists=[])?

I think thats possible, though this PR doesn't do that. Not quite sure what happens if you specify an artist that isn't in the axes (tight layout would have to accept all the artists for each axes).

See update above: rebased and added artist.inbbox to determine if an artist is in a bbox calculation or not.

I am happy with this in principle, but we probably need to add the get_* and set_* methods to keep the API guessable.

We need a better name than inbbox as bbox is a bit arcane and inbbox reads to be as a question not an instruction.

please do not add this extra kwarg.

Sorry, I screwed up the PR during a rebase... The reason for this is so that the artist logic that was in backend_bases.print_figure could be refactored out and reused. So this needs to be here to make that
function work. Could make a private kwarg (is there such a thing?)

added setters and getters, and changed name to artist.set/get_in_layout and made artist._in_layout private.

@efiring Thanks for the review. Outstanding questions for others:

should this be a artist.set/get_in_layout or should it just be a property (artist.in_layout)?
Am I referring to the co-oridnate system of the bounding boxes correctly? get_tightbbox returns bounding boxes in pixels relative to the 0, 0 as the bottom left of the figure viewport. I'm referring to this as:

bbox : `.BboxBase`
            containing the bounding box (in figure pixel co-ordinates).

I don't think this was ever properly specified before this PR. Probably because get_tightbbox and get_window_extents verge on private methods.

re 1. From the pythonic point of view this would just be an attribute artist.in_layout. A property does not make sense here. Properties are basically there to attach code to reading or writing an attribute without changing the signature. You can just use the attribute and turn it into a property later if needed.

The only argument for set/get would be that it's widely used within matplotlib. However there are also a few counter-examples artist.axes, artist.stale.

It's up to one of the more senior devs to decide which way to go. Personally, I'd have a slight preference for the attribute solution.

From the pythonic point of view this would just be an attribute artist.in_layout.

Sorry, meant "attribute".

OTOH I think the issue is documentation - how does the user discover this attribute in the docs?

An attribute can have a docstring, and sphinx can include the documentation either by listing the attribute along with other explicit members, or via the autoattribute directive:

http://www.sphinx-doc.org/en/stable/ext/autodoc.html

I think @tacaswell really doesn't want to start doing this (using bare attributes) until the massive traitification is done, however. But maybe he will change his mind for cases like this.

I’d be all for either form, with slight preference for the attribute. @tacaswell?

I accept @tacaswell's argument that keeping the machinery provided by the setters and getters is worthwhile for this, so ignore my suggestion to reduce it to a bare attribute.

conditional on CI

self merging as reviewed and passed CI

Grrrrr. This has a funny interaction w/ zooming. If I zoom, then the clip path of an artist is not set properly until after a draw. But all the automatic layout depends on being able to get a sensible tighbbox during draw... Darn race conditions....

I just spent several hours tracking down why my annotations were getting placed incorrectly after updating from Matplotlib v2.2.3 to Matplotlib v3.3.1. The problem turned out to be ax.get_tightbbox() was not returning the same result as it did under Matplotlib v2.2.3. I am, however, pleased that I can get the old behavior with ax.get_tightbbox(bbox_extra_artists=[]). Thanks for keeping some degree of backwards compatibility.

You can also twiddle manually: artist.set_in_layout(False)...

…udes tick marks and tick mark labels, but not the axes lables, axes title, or other annotations. As of matplotlib 3.0, however, mpl_ax.get_tightbbox() returns the bounding box that includes the axes labels, axes title, and other annotations (see matplotlib/matplotlib#10682). This new behavior was messing up my axis label placement. Fortunately, one can still get the original behavior via mpl_ax.get_tightbbox(bbox_extra_artists = []), so the axes.tight_bbox property was updated to use this new kwarg.

jklymak added topic: geometry manager LayoutEngine, Constrained layout, Tight layout API: consistency Consistency of the matplotlib API, including naming, behavior, defaults, … API: changes Changes to the public API, typically requiring deprecation. labels Mar 5, 2018

jklymak mentioned this pull request Mar 5, 2018

FIX/ENH: include text in ax.get_tightbbox #10678

Closed

6 tasks

jklymak force-pushed the enh-ax-get-tightbbox branch 2 times, most recently from ec57567 to c53836c Compare March 5, 2018 15:42

jklymak added this to the v2.2.1 milestone Mar 5, 2018

jklymak force-pushed the enh-ax-get-tightbbox branch 2 times, most recently from f93b291 to 650f378 Compare March 5, 2018 20:02

tacaswell modified the milestones: v2.2.1, v3.0 Mar 5, 2018

jklymak force-pushed the enh-ax-get-tightbbox branch 4 times, most recently from 2660b18 to 30edb1f Compare March 6, 2018 05:06

jklymak mentioned this pull request Mar 8, 2018

Pie chart labels out of Figure, even with tight_layout #4251

Closed

jklymak force-pushed the enh-ax-get-tightbbox branch from 30edb1f to 485dd68 Compare March 24, 2018 21:07

jklymak mentioned this pull request Apr 12, 2018

legend(bbox_transform) don't work with tight layout in 2.2.x #11041

Closed

jklymak force-pushed the enh-ax-get-tightbbox branch 2 times, most recently from b90b15f to d948583 Compare April 29, 2018 02:02

jklymak mentioned this pull request Apr 29, 2018

Legend positioning behavior change from matplotlib {2.1.0 -> 2.2.2} #11126

Closed

jklymak force-pushed the enh-ax-get-tightbbox branch 2 times, most recently from 41727bd to fd918ca Compare April 29, 2018 15:38

tacaswell reviewed Apr 30, 2018

View reviewed changes

jklymak force-pushed the enh-ax-get-tightbbox branch from 7e5a28e to b352dbf Compare April 30, 2018 21:31

jklymak force-pushed the enh-ax-get-tightbbox branch from 5bada26 to 6e60ecf Compare July 5, 2018 21:30

jklymak force-pushed the enh-ax-get-tightbbox branch from 6e60ecf to 0ef8d7a Compare July 5, 2018 22:10

jklymak force-pushed the enh-ax-get-tightbbox branch 5 times, most recently from 25c45b2 to b31f829 Compare July 7, 2018 23:00

ENH have ax.get_tightbbox have a bbox around all artists

b18d16c

jklymak force-pushed the enh-ax-get-tightbbox branch from b31f829 to b18d16c Compare July 9, 2018 19:20

tacaswell approved these changes Jul 9, 2018

View reviewed changes

jklymak merged commit 632305e into matplotlib:master Jul 9, 2018

jklymak deleted the enh-ax-get-tightbbox branch July 9, 2018 20:09

jklymak mentioned this pull request Jul 11, 2018

FIX: CL avoid fully collapsed axes #11627

Merged

6 tasks

This was referenced Jul 23, 2018

fix tight_layout bug #11737 #11739

Closed

FIX: Stop tight layout collapsing under zoom #11742

Closed

jklymak mentioned this pull request Oct 1, 2018

Error for bbox_inches='tight' in savefig with make_axes_locatable #12355

Closed

jklymak mentioned this pull request Oct 27, 2018

Regression when calling annotate with nan values for the position #12648

Closed

jklymak mentioned this pull request Jan 30, 2019

Move the call to Formatter.set_locs into Formatter.format_ticks. #13323

Merged

6 tasks

jklymak mentioned this pull request Jul 21, 2020

Major speed regression introduced in "plt.bar" definition clipping between 3.0.3 and 3.3.0. #17974

Closed

-                  def get_tightbbox(self, renderer, call_axes_locator=True):
+                  def get_tightbbox(self, renderer, call_axes_locator=True,
+                          bbox_extra_artists=None):

ENH have ax.get_tightbbox have a bbox around all artists attached to axes. by jklymak · Pull Request #10682 · matplotlib/matplotlib · GitHub

Repository navigation

ENH have ax.get_tightbbox have a bbox around all artists attached to axes. - #10682

ENH have ax.get_tightbbox have a bbox around all artists attached to axes.#10682jklymak merged 1 commit intomatplotlib:mastermatplotlib/matplotlib:masterfrom jklymak:enh-ax-get-tightbboxjklymak/matplotlib:enh-ax-get-tightbboxCopy head branch name to clipboard

Conversation

jklymak commented Mar 5, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Summary

UPDATE 28 April:

PR Checklist

Uh oh!

ImportanceOfBeingErnest commented Apr 12, 2018

Uh oh!

jklymak commented Apr 12, 2018

Uh oh!

jklymak commented Apr 29, 2018

Uh oh!

tacaswell commented Apr 30, 2018

Uh oh!

tacaswell Apr 30, 2018

Choose a reason for hiding this comment

Uh oh!

jklymak Apr 30, 2018

Choose a reason for hiding this comment

Uh oh!

jklymak commented Apr 30, 2018

Uh oh!

jklymak commented Jul 5, 2018

Uh oh!

timhoffm commented Jul 5, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jklymak commented Jul 5, 2018

Uh oh!

efiring commented Jul 6, 2018

Uh oh!

jklymak commented Jul 8, 2018

Uh oh!

efiring commented Jul 9, 2018

Uh oh!

tacaswell left a comment

Choose a reason for hiding this comment

Uh oh!

jklymak commented Jul 9, 2018

Uh oh!

jklymak commented Jul 10, 2018

Uh oh!

breedlun commented Oct 20, 2020

Uh oh!

jklymak commented Oct 20, 2020

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

ENH have ax.get_tightbbox have a bbox around all artists attached to axes.#10682
jklymak merged 1 commit into
matplotlib:masterfrom
jklymak:enh-ax-get-tightbbox

jklymak commented Mar 5, 2018 •
edited
Loading

timhoffm commented Jul 5, 2018 •
edited
Loading