Bug summary

When trying to see if a Polygon contains a point, the contains_point() method only seems to work if alpha is set to 0 or no options are set when defining the polygon. (mpl version = 3.8.0)

Code for reproduction

frommatplotlib.patchesimportPolygonvertices= [[0,0],[0,1],[1,1],[1,0]]# Squarepoint= [0.5,0.5]# Point inside squarepoly0=Polygon(vertices,closed=True,fill=True,color=(1,0,0,1))# Falsepoly1=Polygon(vertices,closed=True,fill=True,color='r')# Falsepoly2=Polygon(vertices,closed=True,fill=True,color=(1,0,0,0))# Truepoly3=Polygon(vertices,fill=False,alpha=0)# Truepoly4=Polygon(vertices,fill=False)# Falsepoly5=Polygon(vertices)# Trueforidx,polyinenumerate((poly0,poly1,poly2,poly3,poly4,poly5)):print(f"Polygon{idx} contains point{point}:{poly.contains_point(point)}")

Actual outcome

Polygon 0 contains point [0.5, 0.5]: False
Polygon 1 contains point [0.5, 0.5]: False
Polygon 2 contains point [0.5, 0.5]: True
Polygon 3 contains point [0.5, 0.5]: True
Polygon 4 contains point [0.5, 0.5]: False
Polygon 5 contains point [0.5, 0.5]: True

Expected outcome

Should return True for all polygons since the point is inside them all.

Additional information

Tested in mpl v3.7.1 and v3.8.0 with identical outcomes
Also fails if fill=False and alpha is not set.
Temporary fix is to set alpha to 0.

Operating system

MacOS 12.5.1 Monterrey

Matplotlib Version

3.8.0

Matplotlib Backend

MacOSX

Python version

Python 3.11.5

Jupyter version

No response

Installation

pip