First attempt for individual hatching styles for stackplot #27158

nbarlowATI · 2023-10-20T15:56:35Z

PR summary

Closes #27146 this modifies stackplot.py to treat the hatch argument to stackplot in the same way as colors, i.e. if a list is provided, it will cycle through those hatch styles for the different data in the stack.
I believe the existing behaviour is preserved - if a single style (e.g. "x") is provided, this will be used for all elements of the stack, while if no "hatch" argument is passed, it will apply hatch style " " to all elements (though perhaps this isn't the best way of doing this?)

Using the example code in the original Issue from @pawjast :

fig, ax = plt.subplots(
    figsize=(10,6),
    facecolor="white",
    layout="constrained"
)
fig.suptitle(
    "with hatching",
    fontsize=25,
    weight="bold"
)

x = range(data.shape[1])
ax.stackplot(
    x, data,
    hatch=["x", "\\", "//", "-"]
)

ax.set_xlim(0, 9)
ax.set_ylim(0, 350)

I get the following:

Let me know if this looks OK, and if so I can make a start on updating documentation and tests.

PR checklist

"closes #0000" is in the body of the PR description to link the related issue
new and changed code is tested
Plotting related features are demonstrated in an example
New Features and API Changes are noted with a directive and release note
Documentation complies with general and docstring guidelines

github-actions

Thank you for opening your first PR into Matplotlib!

If you have not heard from us in a week or so, please leave a new comment below and that should bring it to our attention. Most of our reviewers are volunteers and sometimes things fall through the cracks.

You can also join us on gitter for real-time discussion.

For details on testing, writing docs, and our review process, please see the developer guide

We strive to be a welcoming and open project. Please follow our Code of Conduct.

rcomer · 2023-10-20T16:33:23Z

Thanks for the fast work @nbarlowATI and for checking this works for the example in the issue. I'm unsure whether passing a single string will work at the moment. Could you check what happens if you pass hatch="\\"?

This will also need some tests. Please see here for general guidance on testing. You might also like to look at the tests added at #24470, where a similar change was made for the pie method.

Please mark this as "ready for review" when you are ready (things marked as draft often get overlooked). In the meantime, if you need help, feel free to ask questions here. Or you may prefer to ask them in our incubator gitter room.

ksunden · 2023-10-20T17:46:21Z

lib/matplotlib/stackplot.py

@@ -76,6 +83,14 @@ def stackplot(axes, x, *args,
    else:
        colors = (axes._get_lines.get_next_color() for _ in y)

+    if hatch:
+        if isinstance(hatch, Iterable):


be careful because hatches can be multi-char strings, and strings are Iterable.

I think what you may want is

if isinstance(hatch, (str, None)): hatch = itertools.cycle([hatch]) else: hatch = itertools.cycle(hatch)

I think that should handle even the None case (also " " is not a valid hatch pattern and is causing warnings)

Thanks, and good point! You're right about the strings being iterables, and I think your logic is the right choice (though I had to modify slightly because None doesn't count as a type so can't be used in isinstance).

Ahh, str | None would work (but is py3.10+ only so we can't use it).

builtins.NoneType or type(None) would also work, but is used less commonly.

Type hints allow None as shorthand, but not isinstance, apparently.

story645 · 2023-10-20T19:35:34Z

It'll also need a whats new example - what you have here is fine and you can also use the what's new in #24470 as an example.

ETA: also I'm psyched, this is great and we should vectorize all the things 😄

pawjast · 2023-10-22T19:52:28Z

I was thinking perhaps this change is a good chance to introduce the following API (provided this is clear and acceptable):

ax.stackplot(hatch="x"): single str gives result as we currently have (one big hatching over all data)
ax.stackplot(hatch=["x"]): one str defined in a list will use this one str and vary the density (like what I showed in the issue) so e.g. if there are 4 datasets in the data, the hatchings should be x, xx, ... xxxx
ax.stackplot(hatch=["x", "\\", "//", "-"]): list of str, the code cycles through the list

story645 · 2023-10-22T21:08:18Z

ax.stackplot(hatch=["x"]): one str defined in a list will use this one str and vary the density (like what I showed in the issue) so e.g. if there are 4 datasets in the data, the hatchings should be x, xx, ... xxxx

I understand why you're asking for this but I'm concerned this is a bit magic cause this is not a thing we do for any variables (list of color vary by density). Also if there was support for this, it'd have to work for all the places that take a list of hatches.

And I wonder if this would be better achieved via custom hatch objects?

hatch=DensityHatch('x') which returns an iterator of increasing density - which actually may work currently given our API doesn't really care what type of iterator hatch is?

pawjast · 2023-10-23T08:04:31Z

ax.stackplot(hatch=["x"]): one str defined in a list will use this one str and vary the density (like what I showed in the issue) so e.g. if there are 4 datasets in the data, the hatchings should be x, xx, ... xxxx

I understand why you're asking for this but I'm concerned this is a bit magic cause this is not a thing we do for any variables (list of color vary by density). Also if there was support for this, it'd have to work for all the places that take a list of hatches.

And I wonder if this would be better achieved via custom hatch objects?

hatch=DensityHatch('x') which returns an iterator of increasing density - which actually may work currently given our API doesn't really care what type of iterator hatch is?

No problem. I was just throwing the idea for the debate.

I'm very much in favour of continuous improvement so even if there's a chance my proposal would be considered for implementation but hindered the current work, it's probably best to get out the current change (support for list of hatches) and then go back to the other ideas in the next iterations...

nbarlowATI · 2023-10-24T16:42:40Z

I added a test in test_axes.py that compares a stackplot with a list of hatching styles with the result of usingfill_between with the same data and hatching styles.
I also added a "what's new" entry doc/users/next_whats_new/stackplot_hatch.rst that demonstrates the use of the hatch list.
Happy to add any other tests and/or examples that people think would be useful (now I'm getting the hang of this :) )

story645 · 2023-10-24T16:48:44Z

last thing left is fixing the stubtest by adding this variable - which I think is None | List[str] | str ? - to the type stub

matplotlib/lib/matplotlib/streamplot.pyi

Line 11 in e501543

def streamplot(

story645 · 2023-10-25T13:52:40Z

lib/matplotlib/pyplot.py

@@ -3792,12 +3792,15 @@ def spy(

 # Autogenerated by boilerplate.py.  Do not edit as changes will be lost.


As the comment says, the signatures here are autogenerated

Ah yes, sorry! I think I ran boilerplate.py myself in the hope of fixing the stub github action, as I didn't know about the .pyi file... Should I revert this commit, or will the autogeneration take care of it anyway?

I think you are supposed run boilerplate.py yourself, and this comment is meant to indicate that you shouldn't manually edit the type hints here.

🤦‍♀️ I think Ruth is right given I edited this file in my PR too and we never call boilerplate.py as part of our build/install process. Sorry & I'll open up a follow up issue on documenting how this is used!

lib/matplotlib/stackplot.py

story645 · 2023-10-26T14:28:38Z

lib/matplotlib/stackplot.py

+        A sequence of hatching styles (see reference in
+        https://matplotlib.org/devdocs/gallery/shapes_and_collections/hatch_style_reference.html)


Suggested change

A sequence of hatching styles (see reference in

https://matplotlib.org/devdocs/gallery/shapes_and_collections/hatch_style_reference.html)

A sequence of hatching styles. See :ref:`sphx_gallery_shapes_and_collections_hatch_style_reference.py

Personal bias against parentheticals & we always cross-reference when possible so that the link goes to the tutorial in the same version of the docs
https://sphinx-gallery.github.io/stable/advanced.html#cross-referencing

But also double check that link by building the doc cause I may have done something wonky

OK makes sense. For me, the above ':ref:' link didn't quite work in my locally built version of the docs, but I copied something very similar (using ':doc:') from the pie chart docstring, that seems to work.

thank you for your patience w/ my "oh, another thing" 😅

No problem (I'm a big fan of Columbo!), and thanks for your help! :)

story645

Yay, it's got code and tests and typing and a whats new 🥳

Slight optional wording suggestion, but I think it's good to go. Also do you want to rebase or do you want one of us to?

doc/users/next_whats_new/stackplot_hatch.rst

nbarlowATI · 2023-10-26T17:34:56Z

Thanks! Re: rebasing, it would be great if one of you could do it - I'm away for the next few days (but I could also do it when I'm back...)

rcomer

Thanks for your work on this @nbarlowATI. I just have one minor comment.

lib/matplotlib/stackplot.py

rcomer · 2023-10-28T09:14:45Z

This will also need squashing down to one commit if you are confident to do so, otherwise we can handle that on merge.

Edit: oops I see you already discussed that.

nbarlowATI · 2023-10-31T10:23:15Z

Hi, I'm back online now, and happy to have a go at the rebase. Just to confirm, is the idea to both:

Rebase such that the commits here appear after the current state of main, i.e. git rebase upstream/main
Squash all the commits related to this change into one, using git rebase -i
?
(Or is there a single rebase command that does both of these?)
If so, I can give it a go.

rcomer · 2023-10-31T10:35:39Z

Squash all the commits related to this change into one, using git rebase -i

This one please! Since there are no conflicts between your branch and upstream/main, the ordering of your commits vs upstream/main's commits do not matter.

We have some guidance here if you need it
https://matplotlib.org/devdocs/devel/development_workflow.html#rewrite-commit-history

story645 · 2023-10-31T12:50:27Z

Don't add complication you don't need of course, but git rebase -i upstream/main is the command I always use.

nbarlowATI · 2023-10-31T14:00:35Z

Many thanks @rcomer and @story645 for all your help with this!!!
I believe the commits are now successfully squashed into one.

rcomer

Thanks @nbarlowATI and congratulations on your first PR in Matplotlib! We hope to hear from you again.

nbarlowATI mentioned this pull request Oct 20, 2023

[ENH]: Multi hatching in ax.stackplot() #27146

Closed

github-actions bot reviewed Oct 20, 2023

View reviewed changes

rcomer added the New feature label Oct 20, 2023

rcomer added this to the v3.9.0 milestone Oct 20, 2023

ksunden reviewed Oct 20, 2023

View reviewed changes

nbarlowATI marked this pull request as ready for review October 24, 2023 16:42

ksunden mentioned this pull request Oct 24, 2023

[ENH]: dashed hatching pattern #27170

Open

story645 reviewed Oct 25, 2023

View reviewed changes

story645 mentioned this pull request Oct 25, 2023

[Doc]: clarify when and how to use boilerplate.py #27190

Open

story645 reviewed Oct 25, 2023

View reviewed changes

lib/matplotlib/stackplot.py Outdated Show resolved Hide resolved

lib/matplotlib/stackplot.py Outdated Show resolved Hide resolved

story645 reviewed Oct 26, 2023

View reviewed changes

story645 approved these changes Oct 26, 2023

View reviewed changes

doc/users/next_whats_new/stackplot_hatch.rst Outdated Show resolved Hide resolved

rcomer reviewed Oct 28, 2023

View reviewed changes

lib/matplotlib/stackplot.py Outdated Show resolved Hide resolved

Individual hatching styles for stackplot, as per Issue 27146

914b213

nbarlowATI force-pushed the main branch from 9daffa3 to 914b213 Compare October 31, 2023 13:54

rcomer approved these changes Oct 31, 2023

View reviewed changes

rcomer merged commit 5452efc into matplotlib:main Oct 31, 2023
40 checks passed

mhauru mentioned this pull request Nov 3, 2023

Recurring OSS hacksessions alan-turing-institute/hut23-open-source-sa#18

Open

Jacob-Stevens-Haas mentioned this pull request Dec 7, 2023

[Bug]: Meson build fails due to qhull link issue. #27159

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

First attempt for individual hatching styles for stackplot #27158

First attempt for individual hatching styles for stackplot #27158

nbarlowATI commented Oct 20, 2023 •

edited

github-actions bot left a comment

rcomer commented Oct 20, 2023

ksunden Oct 20, 2023

nbarlowATI Oct 23, 2023

ksunden Oct 25, 2023

story645 commented Oct 20, 2023 •

edited

pawjast commented Oct 22, 2023

story645 commented Oct 22, 2023 •

edited

pawjast commented Oct 23, 2023

nbarlowATI commented Oct 24, 2023

story645 commented Oct 24, 2023

story645 Oct 25, 2023

nbarlowATI Oct 25, 2023

rcomer Oct 25, 2023

story645 Oct 25, 2023 •

edited

story645 Oct 26, 2023

nbarlowATI Oct 26, 2023 •

edited

story645 Oct 26, 2023 •

edited

nbarlowATI Oct 26, 2023

story645 left a comment

nbarlowATI commented Oct 26, 2023

rcomer left a comment

rcomer commented Oct 28, 2023 •

edited

nbarlowATI commented Oct 31, 2023

rcomer commented Oct 31, 2023

story645 commented Oct 31, 2023

nbarlowATI commented Oct 31, 2023

rcomer left a comment

		@@ -3792,12 +3792,15 @@ def spy(

		# Autogenerated by boilerplate.py. Do not edit as changes will be lost.

		A sequence of hatching styles (see reference in
		https://matplotlib.org/devdocs/gallery/shapes_and_collections/hatch_style_reference.html)

	A sequence of hatching styles (see reference in
	https://matplotlib.org/devdocs/gallery/shapes_and_collections/hatch_style_reference.html)
	A sequence of hatching styles. See :ref:`sphx_gallery_shapes_and_collections_hatch_style_reference.py

First attempt for individual hatching styles for stackplot #27158

First attempt for individual hatching styles for stackplot #27158

Conversation

nbarlowATI commented Oct 20, 2023 • edited

PR summary

PR checklist

github-actions bot left a comment

Choose a reason for hiding this comment

rcomer commented Oct 20, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

story645 commented Oct 20, 2023 • edited

pawjast commented Oct 22, 2023

story645 commented Oct 22, 2023 • edited

pawjast commented Oct 23, 2023

nbarlowATI commented Oct 24, 2023

story645 commented Oct 24, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

story645 Oct 25, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nbarlowATI Oct 26, 2023 • edited

Choose a reason for hiding this comment

story645 Oct 26, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

story645 left a comment

Choose a reason for hiding this comment

nbarlowATI commented Oct 26, 2023

rcomer left a comment

Choose a reason for hiding this comment

rcomer commented Oct 28, 2023 • edited

nbarlowATI commented Oct 31, 2023

rcomer commented Oct 31, 2023

story645 commented Oct 31, 2023

nbarlowATI commented Oct 31, 2023

rcomer left a comment

Choose a reason for hiding this comment

nbarlowATI commented Oct 20, 2023 •

edited

story645 commented Oct 20, 2023 •

edited

story645 commented Oct 22, 2023 •

edited

story645 Oct 25, 2023 •

edited

nbarlowATI Oct 26, 2023 •

edited

story645 Oct 26, 2023 •

edited

rcomer commented Oct 28, 2023 •

edited