ENH add an inset_axes to the axes class #11026

jklymak · 2018-04-11T05:11:20Z

PR Summary

and we are back to `inset_axes`....

@efiring suggested just going back to inset_axes. Since that was what we started with, I didn't object. The original argument for the longer name was if the API was to include a loc='SW' API, instead of the bounds API I've used here. Thats still possible by type-checking the bounds argument.

EDIT:

Name bikeshedding:

Proposal after comments below:

Axes.inset_axes_from_bounds
Axes.indicate_inset_bounds
Axes.indicate_inset_zoom

rect was too ambiguous, lbwh too obscure, bounds is already used for bboxes (as is extents for lbrt)...

Comments on names still welcome...

Old PR descr

Adds ax.inset_axes_from_rect. Pretty straight forward, but note that the transData one moves under zoom and pan.

Note that I chose not to go with the bbox_to_anchor, and loc approach. I understand that stuff, but it seems needlessly complicated. Maybe as a refinement.

import matplotlib.pyplot as plt

fig, ax = plt.subplots()
ax.plot(range(10))
axin1 = ax.inset_axes_from_rect([0.8, 0.1, 0.15, 0.15])
axin2 = ax.inset_axes_from_rect([5, 7, 2.3, 2.3], transform=ax.transData)
plt.show()

Getting closer: This is a duplication of https://matplotlib.org/gallery/axes_grid1/inset_locator_demo2.html#sphx-glr-gallery-axes-grid1-inset-locator-demo2-py

import matplotlib.pyplot as plt

import numpy as np

def get_demo_image():
    from matplotlib.cbook import get_sample_data
    import numpy as np
    f = get_sample_data("axes_grid/bivariate_normal.npy", asfileobj=False)
    z = np.load(f)
    # z is a numpy array of 15x15
    return z, (-3, 4, -4, 3)

fig, ax = plt.subplots(figsize=[5, 4], constrained_layout=True)
Z, extent = get_demo_image()
Z2 = np.zeros([150, 150], dtype="d")
ny, nx = Z.shape
Z2[30:30 + ny, 30:30 + nx] = Z

# extent = [-3, 4, -4, 3]
ax.imshow(Z2, extent=extent, interpolation="nearest",
          origin="lower")

axins = ax.inset_axes_from_rect([0.5, 0.5, 0.47, 0.47])
axins.imshow(Z2, extent=extent, interpolation="nearest",
          origin="lower")
# sub region of the original image
x1, x2, y1, y2 = -1.5, -0.9, -2.5, -1.9
axins.set_xlim(x1, x2)
axins.set_ylim(y1, y2)
axins.set_xticklabels('')
axins.set_yticklabels('')

ax.zoom_inset_indicator(axins)
fig.canvas.draw()
plt.show()

EDIT: methods renamed inset_axes_rect to allow other APIs in the future.

PR Checklist

Has Pytest style unit tests
Code is PEP 8 compliant
New features are documented, with examples if plot related
Documentation is sphinx and numpydoc compliant
Added an entry to doc/users/next_whats_new/ if major new feature (follow instructions in README.rst there)
Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way

ImportanceOfBeingErnest · 2018-04-16T23:04:11Z

In how far is it bearable or confusing for users to have

an axes_grid1.inset_locator.inset_axes which creates a axes_grid1.parasite_axes.AxesHostAxes
an axes.Axes.inset_axes which creates a .axes.Axes ?

(I could imagine the problem arising because people are used to have the same function at different places, e.g. plt.plot vs. axes.Axes.plot)

jklymak · 2018-04-16T23:10:18Z

mplot3d call to plot objects return a different object than core plot. I don't personally think it should be a problem...

jklymak · 2018-04-19T18:32:59Z

I am done w/ this for now, and it can be reviewed if anyone has any suggestions. (note looks like a spare debug change got into the PR, but I'll remove after/during review).

anntzer · 2018-04-19T18:36:47Z

examples/mplot3d/lines3d.py

@@ -23,7 +23,7 @@
 x = r * np.sin(theta)
 y = r * np.cos(theta)

-ax.plot(x, y, z, label='parametric curve')
+l = ax.plot(x, y, z, label='parametric curve')


remove change

anntzer · 2018-04-19T18:39:12Z

lib/matplotlib/axes/_axes.py

+    """
+
+    def __init__(self, rect, trans, parent):
+        self._rect = rect


just create the transformedbbox here?

anntzer · 2018-04-19T18:39:54Z

lib/matplotlib/axes/_axes.py

+
+    def __call__(self, ax, renderer):
+        rect = self._rect
+        bbox = mtransforms.Bbox.from_bounds(rect[0], rect[1],


would typically write from_bounds(*rect), not that it really matters

anntzer · 2018-04-19T18:40:30Z

lib/matplotlib/axes/_axes.py

+        # This puts the rectangle into figure-relative coordinates.
+        bb = _inset_locator(rect, transform, self)(None, None)
+        # if not set, set to the same zorder as legend
+        zorder = kwargs.pop('zorder', 5)


kwargs.setdefault("zorder", Legend.zorder)

Well, I don't think I want it tied to the legend value. And legend's value is hard coded i.e. not an rcParam....

anntzer · 2018-04-19T18:44:37Z

lib/matplotlib/axes/_axes.py

+
+        zorder : number
+            Defaults to 5 (same as `.Axes.legend`).  Adjust higher or lower
+            to change whether it is above or below data plotted on the axes.


"... on the parent axes" (or however you call it)

anntzer · 2018-04-19T18:48:22Z

lib/matplotlib/axes/_axes.py

+            coordinates.
+
+        facecolor : Matplotlib color
+            Facecolor of the rectangle (default 'none')


full stops at end of lines (same below)

anntzer · 2018-04-19T18:48:30Z

lib/matplotlib/axes/_axes.py

+            is '0.5'
+
+        alpha : number
+            Transparency of the rectangle and connector lines.  Default 0.5


anntzer · 2018-04-19T18:48:55Z

lib/matplotlib/axes/_axes.py

+        -------
+
+        rectangle_patch, connector_lines :
+            `.Patches.Rectagle`, (four-tupple `.Patches.ConnectionPatch`) one


Rectangle tuple

anntzer · 2018-04-19T18:49:37Z

lib/matplotlib/axes/_axes.py

+            # decide which two of the lines to keep visible....
+            pos = inset_ax.get_position()
+            bboxins = pos.transformed(self.figure.transFigure)
+            rectbbox = mtransforms.Bbox.from_bounds(rect[0], rect[1],


from_bounds(*rect)

anntzer · 2018-04-19T18:49:58Z

lib/matplotlib/axes/_axes.py

+        -------
+
+        rectangle_patch, connector_lines :
+            `.Patches.Rectagle`, (four-tupple `.Patches.ConnectionPatch`) one


rectangle, tuple

anntzer · 2018-04-19T18:50:33Z

lib/matplotlib/axes/_base.py

@@ -1048,6 +1048,7 @@ def cla(self):
        self.tables = []
        self.artists = []
        self.images = []
+        self.child_axes = []


stash them into self.artists? I'd rather work towards unifying all these containers rather than adding more of them...

Seec omment below for why not implimented....

ImportanceOfBeingErnest · 2018-04-19T19:21:58Z

What I definitely like better about the axes_grid1.inset_locator.inset_axes is that you get some automatic padding, similar to a legend.
Here, any padding is either in percent of the parent axes width/height (i.e. the position would shift if the axes size changes) or you need to create an offset transform from scratch to supply to the transform argument.
So the equivalent of

from mpl_toolkits.axes_grid1.inset_locator import inset_axes

axins = inset_axes(ax, width="40%", height="30%")

would be

from matplotlib.transforms import ScaledTranslation

offset = ScaledTranslation(-5/72., -5/72., fig.dpi_scale_trans)
trans = ax.transAxes + offset
ax.inset_axes([.6,.7,.4,.3], transform=trans)

which is a bit cumbersome to write for such standard case.

anntzer · 2018-04-19T19:24:21Z

If that's a desirable feature (not certain yet, but possibly), certainly it should be pulled to core so that legend, text annotations and whatnot can also benefit from it?

ImportanceOfBeingErnest · 2018-04-19T19:37:14Z

@anntzer I think in its current form axes_grid1.inset_locator.inset_axes should not simply be pulled over, because it has some design flaws which become apparent once you leave the standard use case.
You may read #11060 and links therein to see what is confusing about it (and also review if you like ;)).

So for the core matplotlib, one should probably spend some more thoughts about how to best design such function to be powerful, but not as confusing.

jklymak · 2018-04-19T19:44:44Z

@anntzer and @ImportanceOfBeingErnest, thanks for the feedback. I think the logic for padded boxes is indeed in .offsets.AnchoredOffsetBox.

.offsets.AnchoredOffsetBox works from the idea that you want to put the box somewhere semi-automatically inside an axes (or other bbox). In that case a padding makes some sense. But I find the whole API hard to deal with if I just want to put a box (or axes) somewhere deterministic on the figure or parent_axes. So here, I was trying to be more deterministic with a more streamlined API.

What might make sense is to add a "pad" parameter in appropriate units (probably points, versus pixels), that would offset the new inset axes from the rectangle provided. Then if you want upper right you do:

w = 0.2, h=0.2
axins = ax.inset_axes([1-w, 1-h, w, h], pad=5)

I don't think thats hard to do, though I'll have to check how its done to be robust under figure size changes. Could have an hpad and wpad if desired. I'd tend to be against different pads left and right, up and down...

Not sure how all that would work with set aspect ratios for the child axes.

Would something like that satisfy the need?

Conversely, I'm not 100% against just moving the inset_axes into core. But I really don't like the API, so felt this would be a decent time to change it. I appreciate some power users have figured out how to make it work, and recent attempts to improve the documentation have been heroic, but...

EDIT: cross-post!

anntzer · 2018-04-19T19:54:54Z

To be clear, I think this should go in without padding support, because it's basically using the same positioning syntax as text, which is a good thing. Padding support should get its own discussion (and I'm not even certain it's something worth the complexity).

jklymak · 2018-04-19T20:05:48Z

I guess personally I don't have a problem with padding just being a fraction of the axes size (i.e. I can hard code it into rect), but I can see why it is nice to have in physical units for consistency between different sized-axes. But I'm more than happy to leave that for a future PR, either as a new method or by enhancing the API of these methods. I don't think there is anything architecturally that needs to change to make that work.

ImportanceOfBeingErnest · 2018-04-19T20:28:26Z

I'm thinking more in terms of ease of use. One can always argue that if people want to have full control they can reach their goal via the axes_grid1 inset or custom transforms etc.
So I'm comparing to legend here, because that is similar in the sense that you have a box with some extra information in. You may call legend() without any argument and get a nice legend. This has some padding to the outside and if you don't like the position you'd add loc=4 at most. Apart it's still deterministic and fully customizable, using all the remaining arguments.

In that sense, something like

def inset_axes(loc="upper right", pad=0.5, width=0.4, height=0.3, box=None, transform=None):

which can be called just as ax.inset_axes() could be more desireable.

anntzer · 2018-04-19T21:06:37Z

I think "loc=number" is a terrible API. At least that should be using strings (which are also allowed with legend, but I would want new APIs to only use strings).

jklymak · 2018-06-20T20:02:16Z

ping @efiring and @timhoffm if the new name suits. We could discuss on Monday's dev call, except I'll be unavailable. Happy to chat it out on gitter as well...

jklymak · 2018-07-04T19:12:19Z

Rebased and squashed. Be nice to get this in relatively soon because then I can move onto axes for secondary scales...

jklymak · 2018-07-08T14:03:24Z

recycled for CI

jklymak · 2018-07-10T17:46:24Z

rebased

jklymak · 2018-07-18T02:12:37Z

This just needs one more review to go in as a pretty useful 3.0 feature. OTOH, understand if it needs to get pushed 6 months...

efiring · 2018-07-18T19:08:20Z

lib/matplotlib/axes/_axes.py

+    _parent = parent
+
+    def inset_locator(ax, renderer):
+        bbox = _rect


Is this line actually doing anything other than renaming a variable? I think it is superfluous.

efiring

I would like to see this go in ASAP, preferably with one of the variations on naming that was discussed: drop the "from...", so it is "inset_axes()" and "indicate_inset()", and use "bounds" consistently for the corresponding variable name in private as well as public functions.

jklymak · 2018-07-20T04:50:33Z

Rebased and renamed to plain old inset_axes per @efiring. @timhoffm that may invalidate your approval? I'll leave it to the maintainers whether this has enough approval...

jklymak added the status: work in progress label Apr 11, 2018

jklymak added this to the v3.0 milestone Apr 11, 2018

jklymak force-pushed the enh-axes-add-axes branch 2 times, most recently from ea75792 to 0d3425a Compare April 16, 2018 21:16

jklymak mentioned this pull request Apr 16, 2018

Allow equal aspect box on shared axes #10961

Closed

2 tasks

jklymak force-pushed the enh-axes-add-axes branch from 4c63f1c to 082b8a6 Compare April 16, 2018 23:12

jklymak removed the status: work in progress label Apr 17, 2018

jklymak changed the title ~~WIP/ENH add an inset_axes to the axes class~~ ENH add an inset_axes to the axes class Apr 17, 2018

anntzer reviewed Apr 19, 2018

View reviewed changes

jklymak force-pushed the enh-axes-add-axes branch from 81f5aed to 435db99 Compare June 19, 2018 17:35

jklymak removed the status: work in progress label Jun 19, 2018

timhoffm approved these changes Jun 20, 2018

View reviewed changes

jklymak force-pushed the enh-axes-add-axes branch from 435db99 to b712c7b Compare July 4, 2018 19:11

jklymak mentioned this pull request Jul 6, 2018

WIP ENH secondary axes: #11589

Closed

5 tasks

jklymak closed this Jul 8, 2018

jklymak reopened this Jul 8, 2018

jklymak force-pushed the enh-axes-add-axes branch from b712c7b to edd3a13 Compare July 10, 2018 17:45

jklymak force-pushed the enh-axes-add-axes branch from edd3a13 to 37aefb7 Compare July 10, 2018 20:18

leejjoon mentioned this pull request Jul 18, 2018

Deprecate LocatableAxes from toolkits #10403

Merged

4 tasks

efiring reviewed Jul 18, 2018

View reviewed changes

jklymak force-pushed the enh-axes-add-axes branch from 37aefb7 to 0abd014 Compare July 18, 2018 20:18

efiring approved these changes Jul 20, 2018

View reviewed changes

jklymak changed the title ~~ENH add an inset_axes_from_bounds to the axes class~~ ENH add an inset_axes to the axes class Jul 20, 2018

jklymak force-pushed the enh-axes-add-axes branch 2 times, most recently from 82b88b7 to 35c29c2 Compare July 20, 2018 04:44

jklymak force-pushed the enh-axes-add-axes branch from 35c29c2 to 0d604bf Compare July 20, 2018 05:09

ENH: add an inset_axes to the axes class

a5b27e9

jklymak force-pushed the enh-axes-add-axes branch from 0d604bf to a5b27e9 Compare July 20, 2018 05:13

efiring merged commit bc39b1c into matplotlib:master Jul 20, 2018

This was referenced Jul 20, 2018

DOC: fix broken link in inset-locator example #11717

Merged

ENH: inset_axes anchored-artist API #11730

Closed

anntzer removed the status: needs review label Sep 14, 2018

jklymak moved this from In progress to Done in axes_grid1 toolkit to main code base.... Sep 28, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ENH add an inset_axes to the axes class #11026

ENH add an inset_axes to the axes class #11026

jklymak commented Apr 11, 2018 •

edited

ImportanceOfBeingErnest commented Apr 16, 2018

jklymak commented Apr 16, 2018

jklymak commented Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018 •

edited

jklymak Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018

anntzer Apr 19, 2018

jklymak May 16, 2018

ImportanceOfBeingErnest commented Apr 19, 2018

anntzer commented Apr 19, 2018 •

edited

ImportanceOfBeingErnest commented Apr 19, 2018

jklymak commented Apr 19, 2018 •

edited

anntzer commented Apr 19, 2018

jklymak commented Apr 19, 2018

ImportanceOfBeingErnest commented Apr 19, 2018 •

edited

anntzer commented Apr 19, 2018

jklymak commented Jun 20, 2018

jklymak commented Jul 4, 2018

jklymak commented Jul 8, 2018

jklymak commented Jul 10, 2018

jklymak commented Jul 18, 2018

efiring Jul 18, 2018

efiring left a comment

jklymak commented Jul 20, 2018

ENH add an inset_axes to the axes class #11026

ENH add an inset_axes to the axes class #11026

Conversation

jklymak commented Apr 11, 2018 • edited

PR Summary

and we are back to inset_axes....

EDIT:

Old PR descr

PR Checklist

ImportanceOfBeingErnest commented Apr 16, 2018

jklymak commented Apr 16, 2018

jklymak commented Apr 19, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

anntzer Apr 19, 2018 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ImportanceOfBeingErnest commented Apr 19, 2018

anntzer commented Apr 19, 2018 • edited

ImportanceOfBeingErnest commented Apr 19, 2018

jklymak commented Apr 19, 2018 • edited

anntzer commented Apr 19, 2018

jklymak commented Apr 19, 2018

ImportanceOfBeingErnest commented Apr 19, 2018 • edited

anntzer commented Apr 19, 2018

jklymak commented Jun 20, 2018

jklymak commented Jul 4, 2018

jklymak commented Jul 8, 2018

jklymak commented Jul 10, 2018

jklymak commented Jul 18, 2018

Choose a reason for hiding this comment

efiring left a comment

Choose a reason for hiding this comment

jklymak commented Jul 20, 2018

jklymak commented Apr 11, 2018 •

edited

and we are back to `inset_axes`....

anntzer Apr 19, 2018 •

edited

anntzer commented Apr 19, 2018 •

edited

jklymak commented Apr 19, 2018 •

edited

ImportanceOfBeingErnest commented Apr 19, 2018 •

edited