Errorous behavior in Vispy.Visuals.Markers.set_data

[c40zAtGitHub]

In line 606-608 of markers.py, the following statement is written:

if (edge_width is not None) + (edge_width_rel is not None) != 1: raise ValueError('exactly one of edge_width and edge_width_rel ' 'must be non-None')

while in line 583-585, the method is defined as

def set_data(self, pos=None, size=10., edge_width=1., edge_width_rel=None, edge_color='black', face_color='white', symbol='o'):

An error occurs when only edge_width_rel is supplied to the set_data method while the edge_width is untouched. The sum in the if statement in this case will be 2, causing the ValueError raised instead of normal executation of the method (which is expected since the edge_width_rel is supplied).

Current issue can be dodged by explicitly supplying edge_width=None in the input.

Apparent logic for the ValueError on line 607 to be raised should be
if (edge_width is None) and (edge_width_rel is None):

Personally, I guess that the usage of edge_width should be surpressed when edge_width_rel is supplied. This can be achieved by changing code in line 610-617 to something like
if edge_width_rel is not None: edge_width_rel = np.asarray(edge_width_rel) if np.any(edge_width_rel < 0): raise ValueError('edge_width_rel cannot be negative') else: edge_width = np.asarray(edge_width) if np.any(edge_width < 0): raise ValueError('edge_width cannot be negative')

[djhoese]

Looks like some of this logic has been there for 9 years or so. In the last couple years @brisvag had updated some of it. I think your suggestion to essentially swap the i edge_width_rel is not None and edge_width if statements makes sense. Would you mind submitting a pull request and then we can get more feedback from other vispy maintainers.

[c40zAtGitHub]

Hi, David. Thanks for the reply. I have setup the pull request: #2565. It has a slight deviation from what I had written originally, but I believe the problem is still there.

…

________________________________ From: David Hoese ***@***.***> Sent: Thursday, December 28, 2023 10:27 AM To: vispy/vispy ***@***.***> Cc: c40zAtGitHub ***@***.***>; Author ***@***.***> Subject: Re: [vispy/vispy] Errorous behavior in Vispy.Visuals.Markers.set_data (Issue #2564) Looks like some of this logic has been there for 9 years or so. In the last couple years @brisvag<https://github.com/brisvag> had updated some of it. I think your suggestion to essentially swap the i edge_width_rel is not None and edge_width if statements makes sense. Would you mind submitting a pull request and then we can get more feedback from other vispy maintainers. — Reply to this email directly, view it on GitHub<#2564 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AFIQ3LXNXS3Y54ZEEBEWO3DYLWFXPAVCNFSM6AAAAABBFSJ5YWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNZRGI3TINZQGA>. You are receiving this because you authored the thread.Message ID: ***@***.***>