Issue26074: Resolves different edgecolor and hatch color in bar plot #26683

Vashesh08 · 2023-09-03T11:49:28Z

PR summary

Fixes #26074
Added a new parameter hatchcolor in bar method which changes the hatch color and also added support for rcParams['hatch.color'] = "color"

Credits to @story645 for coming up with the original idea on how to resolve this.

PR checklist

"closes [ENH]: Different edgecolor and hatch color in bar plot #26074 " 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

Vashesh08 · 2023-09-03T11:51:13Z

import matplotlib.pyplot as plt
import numpy as np

width = 0.35
x = np.arange(4) + 1
y_red = [1, 3, 1, 4]
y_blue = [2, 2, 4, 1]

plt.bar(x - 0.5 * width, y_red, width, label="Height of red bars", hatch="////", facecolor=(0, 0, 0, 0), edgecolor="black", hatchcolor="red")

plt.bar(x + 0.5 * width, y_blue, width, label="Height of blue bars", hatch=r"\\", facecolor=(0, 0, 0, 0), edgecolor="black", hatchcolor="blue")

plt.xticks(x)
plt.yticks([0, 1, 2, 3, 4])
plt.legend()
plt.savefig("hatch.png")
plt.show()

The code can be edited in the following way to provide a different hatchcolor and edgecolor

rcomer · 2023-09-03T17:29:06Z

This change has caused some test failures. @Vashesh08 are you confident to investigate those yourself? You can see the full output from the tests by clicking the "Details" link by one of the checks at the bottom of this page. You can also run the tests yourself locally:
https://matplotlib.org/devdocs/devel/testing.html

jklymak

Is hatch.color only going to work for bar? I think it either needs to work for all patches or at the very least the rcparam needs to be more specific.

jklymak · 2023-09-03T18:28:37Z

lib/matplotlib/axes/_axes.py

@@ -2488,6 +2491,12 @@ def bar(self, x, height, width=0.8, bottom=None, *, align="center",
                itertools.cycle(mcolors.to_rgba_array(edgecolor)),
                # Fallback if edgecolor == "none".
                itertools.repeat('none'))
+        if hatchcolor is None:
+            hatchcolor = itertools.repeat(None)


This appears to advance the default color so by default the hatch is different from the edge color? That doesn't seem a reasonable default.

story645 · 2023-09-03T19:24:42Z

I think it either needs to work for all patches

Same here, and I think from what I can follow this PR adds a new method to patch to do so and updates the patch constructor was updated and so there shoul 8000 d be a test that the keyword is forwarded correctly for all the methods that forward patch arguments.

story645

really psyched that you took the "add it to the patch constructor" approach! Please:

check that the functions that return patch objects (like pie's wedges and histograms) can take this argument
add some tests that this argument works the way it's expected to- defaults to edgecolor or rcparam or uses the passed in color
add a versionadded note and entry to what's new - links are in the PR checklist
update the mypy stubs with the new argument and its type
Thanks!

lib/matplotlib/axes/_axes.py

Vashesh08 · 2023-09-04T12:39:25Z

This change has caused some test failures. @Vashesh08 are you confident to investigate those yourself? You can see the full output from the tests by clicking the "Details" link by one of the checks at the bottom of this page. You can also run the tests yourself locally: https://matplotlib.org/devdocs/devel/testing.html

Cool I'll try to run the tests.

Is hatch.color only going to work for bar? I think it either needs to work for all patches or at the very least the rcparam needs to be more specific.

@story645 @jklymak rcParams["hatch.color"] is by default black. I'm trying to put the update the hatch color along with the patch arguments. So I am trying to work on this but it will either default to black or I'll have to change the default of the parameter here.
https://github.com/matplotlib/matplotlib/blob/389a7e8889c25d05afab4e9ba7ce43e56cf82a99/lib/matplotlib/mpl-data/matplotlibrc#L159C1-L160C1.
Any suggestions ?

Vashesh08 · 2023-09-04T12:54:17Z

Is hatch.color only going to work for bar? I think it either needs to work for all patches or at the very least the rcparam needs to be more specific.

I've tried a couple of methods. This works for both if we use in the following way

import matplotlib as mpl
mpl.rcParams["hatch.color"]="color"

However, for custom hatchcolor in each method, as @story645 mentioned, we need to add it to each method.

Vashesh08 · 2023-09-04T13:17:56Z

really psyched that you took the "add it to the patch constructor" approach! Please:

check that the functions that return patch objects (like pie's wedges and histograms) can take this argument

add some tests that this argument works the way it's expected to- defaults to edgecolor or rcparam or uses the passed in color

add a versionadded note and entry to what's new - links are in the PR checklist

update the mypy stubs with the new argument and its type
Thanks!

Thanks super helpful. I'll try to work on these.

story645 · 2023-09-05T15:40:24Z

So please let us know what your preference is, but I want to ping @timhoffm as API consistency lead. The question here is on defaulting hatchcolor. Current behavior I think is that it defaults to edgecolor and then falls back to hatch.color but I'm not positive and is this the behavior we want? I might also just put this on the call.

jklymak · 2023-09-05T16:09:48Z

In my opinion it should default to edgecolor. I'm not even clear that an rcParam is a good idea for this parameter - default to edgecolor except if specified would be simplest.

Vashesh08 · 2023-09-05T17:07:21Z

I'm not sure about this but what I was thinking is that if there are multiple subplots, maybe then instead of specifying it individually, this parameter could be useful.

import matplotlib.pyplot as plt

# formatting
plt.rcParams["hatch.color"] = "yellow"
plt.rcParams["text.usetex"] = True
plt.rc('font', family='serif')
plt.rc("font",size=11)

first_year = [20,10,70]

legend=["First","Second"]
fig1,(ax1, ax2) = plt.subplots(1,2)
ax1.pie(first_year, hatch=['**O', 'oO', 'O.O', '.||.'])
ax1.set_title("Year 1",y=1.5,weight='bold')
fig1.legend(legend,ncol=3,loc="lower center")
data1 = [0.3, 0.4, 0.3]


# plot each pie chart in a separate subplot
ax2.pie(data1, hatch=['**O', 'oO', 'O.O', '.||.'])
ax2.set_title("Year 2",y=1.5,weight='bold')

plt.show()
plt.show()

Like In the above example, I think it should default to hatch.color (because it is set by the user).

My suggestion would be if there is no hatchcolor parameter in the function, then it should default to the hatch.color(only if it is set by the user) and maybe then fall back to the edgecolor parameter.

story645 · 2023-09-05T18:21:08Z

I'm not even clear that an rcParam is a good idea for this parameter - default to edgecolor except if specified would be simplest.

I think the rcParam is for the edgecolor=None but still want a hatch case, which I think is this example https://matplotlib.org/devdocs/gallery/lines_bars_and_markers/filled_step.html#sphx-glr-gallery-lines-bars-and-markers-filled-step-py which should be much simpler if this PR goes in?

It's also possibly all the fill methods? And possibly all the patches - trying to chase through but default seems to be edgecolor=None

then it should default to the hatch.color(only if it is set by the user) and maybe then fall back to the edgecolor parameter.

Problem is there's no way to distinguish between default black and user set black

jklymak · 2023-09-05T18:56:38Z

I guess it's OK to have an rcParam if edgecolor='none' (not None). But otherwise I think it should default to the edgecolor, and otherwise need to be specified as a kwarg.

Vashesh08 · 2023-09-06T14:37:39Z

Problem is there's no way to distinguish between default black and user set black

Yes, without changing the default,it seems difficult and changing the default would create a lot of problems. I think for the time being, defaulting to edgecolor seems to be the way forward.

I'll default it to edgecolor and then to hatch.color.

story645 · 2023-09-13T02:35:57Z

@Vashesh08 I think you didn't add/push your tests?

oscargus · 2023-09-13T10:00:04Z

If you consider the other colors (edge/face) they all store an _original_edgecolor etc so that alpha can be changed separately. Have a look at

matplotlib/lib/matplotlib/patches.py

Lines 340 to 356 in e65a4e3

    
               def _set_facecolor(self, color): 
        
                   if color is None: 
        
                       color = mpl.rcParams['patch.facecolor'] 
        
                   alpha = self._alpha if self._fill else 0 
        
                   self._facecolor = colors.to_rgba(color, alpha) 
        
                   self.stale = True 
        
               def set_facecolor(self, color): 
        
                   """ 
        
                   Set the patch face color. 
        
                   Parameters 
        
                   ---------- 
        
                   color : color or None 
        
                   """ 
        
                   self._original_facecolor = color 
        
                   self._set_facecolor(color)

I think that a similar approach should be used here so that set_alpha(0.5) will apply to the hatches as well.

This also opens up the question if set_edgecolor should set the hatch color or not? Or if one possibly can keep track of if hatchcolor is set explicitly (a non-None value) or not. Maybe it can be solved through documenting it carefully though.

Vashesh08 · 2023-10-02T05:39:34Z

I have added a workaround to set the hatchcolor directly from the Patch object, so that hatchcolor is applied on all functions which return patch object. The current behavior is explicity set hatchcolor -> defaults to edgecolor -> defaults to rcParams.

This also opens up the question if set_edgecolor should set the hatch color or not?

In my opinion, it shouldn't but changing would cause a regression for users.

Vashesh08 changed the title ~~Issue26074~~ Issue26074: Resolves different edgecolor and hatch color in bar plot Sep 3, 2023

story645 added the New feature label Sep 3, 2023

jklymak reviewed Sep 3, 2023

View reviewed changes

story645 requested changes Sep 4, 2023

View reviewed changes

lib/matplotlib/axes/_axes.py Outdated Show resolved Hide resolved

Vashesh08 requested review from jklymak and story645 September 8, 2023 08:56

github-actions bot added the status: needs rebase label Oct 1, 2023

Vashesh08 closed this Oct 4, 2023

Vashesh08 force-pushed the Issue26074 branch from 5ba82e2 to 21ce592 Compare October 4, 2023 06:45

Vashesh08 mentioned this pull request Oct 4, 2023

Resolves different edgecolor and hatch color in bar plot #26993

Closed

5 tasks

QuLogic added the status: duplicate label Oct 4, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Issue26074: Resolves different edgecolor and hatch color in bar plot #26683

Issue26074: Resolves different edgecolor and hatch color in bar plot #26683

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Issue26074: Resolves different edgecolor and hatch color in bar plot #26683

Issue26074: Resolves different edgecolor and hatch color in bar plot #26683

Uh oh!

Conversation

Uh oh!

PR summary

PR checklist

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!