[Bug]: set_rticks makes polar autoscale move the origin away from zero

Bug summary

On polar axes, autoscaling (which only applies on the radial direction) normally avoids moving the origin away from r=0 (unless there's negative radius data). However, this behavior is lost if set_rticks has been called first.

(split out of #29516)

Code for reproduction

import matplotlib.pyplot as plt

ax = plt.figure().add_subplot(projection="polar")
ax.plot([0, 1], [2, 3])  # theta, r
print(ax.get_ylim())  # -> rlim = (0, 3.05), OK: lower r=0 bound forced

ax = plt.figure().add_subplot(projection="polar")
ax.set_rticks([1, 2, 3, 4])
print(ax.get_ylim())  # -> rlim = (0, 4), OK
ax.plot([0, 1], [2, 3])  # theta, r
print(ax.get_ylim())  # -> rlim = (1.95, 3.05), oops: lower r=0 bound lost

Actual outcome

See comments in code.

Expected outcome

The last autoscaling should also set rlims to (0, 3.05).

Additional information

No response

Operating system

any

Matplotlib Version

3.11.0.dev342+g3b54c6abab

Matplotlib Backend

agg

Python version

3.13

Jupyter version

no

Installation

git checkout

Fix matplotlib#29528 set_rticks makes autoscale move the origin away …

I think the problem is that set_rticks replaces the RadialLocator with the FixedLocator. The view_limits method on the RadialLocator is what usually keeps the origin at or below zero.

matplotlib/lib/matplotlib/projections/polar.py

Lines 506 to 511 in a9dc9ac

    
           def view_limits(self, vmin, vmax): 
        
               vmin, vmax = self.base.view_limits(vmin, vmax) 
        
               if self._zero_in_bounds() and vmax > vmin: 
        
                   # this allows inverted r/y-lims 
        
                   vmin = min(0, vmin) 
        
               return mtransforms.nonsingular(vmin, vmax)

Do we need to implement a RadialFixedLocator?

Do we need to implement a RadialFixedLocator?

The RadialLocator is a wrapper around a base locator. Would it work to set RadialLocator.base to the FixedLocator?

@bruswaine

Oh, maybe we just need to call self.yaxis._wrap_locator_formatter within set_rticks

matplotlib/lib/matplotlib/projections/polar.py

Lines 722 to 725 in a9dc9ac

    
           def _wrap_locator_formatter(self): 
        
               self.set_major_locator(RadialLocator(self.get_major_locator(), 
        
                                                    self.axes)) 
        
               self.isDefault_majloc = True

@bruswaine do you want to give that a try?

@rcomer

@rcomer yes. I have already altered my solution to one that uses the RadialLocator, very similar to what you proposed and it seems to work. Should i open a new PR or just commit to the branch in my fork? Sorry im not sure how it works. Thanks for the help

@bruswaine

@bruswaine great - it seems you are ahead of me! It's fine to commit to your existing branch.

Fix matplotlib#29528 set_rticks makes autoscale move the origin away …

Ensure polar plot radial lower limit remains at 0 after set_rticks + …

Navigation Menu

Repository navigation

[Bug]: set_rticks makes polar autoscale move the origin away from zero #29528

Bug summary

Code for reproduction

Actual outcome

Expected outcome

Additional information

Operating system

Matplotlib Version

Matplotlib Backend

Python version

Jupyter version

Installation

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

	def view_limits(self, vmin, vmax):
	vmin, vmax = self.base.view_limits(vmin, vmax)
	if self._zero_in_bounds() and vmax > vmin:
	# this allows inverted r/y-lims
	vmin = min(0, vmin)
	return mtransforms.nonsingular(vmin, vmax)

	def _wrap_locator_formatter(self):
	self.set_major_locator(RadialLocator(self.get_major_locator(),
	self.axes))
	self.isDefault_majloc = True

[Bug]: set_rticks makes polar autoscale move the origin away from zero #29528

Description

Bug summary

Code for reproduction

Actual outcome

Expected outcome

Additional information

Operating system

Matplotlib Version

Matplotlib Backend

Python version

Jupyter version

Installation

Activity

rcomer commented on Mar 24, 2025

timhoffm commented on Mar 24, 2025

rcomer commented on Mar 25, 2025

bruswaine commented on Mar 25, 2025

rcomer commented on Mar 25, 2025

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions