Distance#310
Conversation
There was a problem hiding this comment.
Yes, I think this is right (and better too, there could have been issues with the old code if someone initialized a polygon with the vertices out of order). Plus, tests are still passing
| ) | ||
|
|
||
| np.testing.assert_allclose(distances, true_distances) | ||
| np.testing.assert_allclose(displacements[1:7], true_displacements) |
There was a problem hiding this comment.
@rynoliphant do you have the other displacement values written out?
There was a problem hiding this comment.
The reason the other displacement values weren't written out was because there are multiple possible displacements to the surface (besides point [4,3,3], which I think actually has only one valid displacement). Point [3,3,3] is in the center of the spheropolyhedron and has 6 valid displacements (one to each face), point [4,4,4] sits on a vertex of the inner cube and has infinite valid displacements, point [4,4,3] sits on an edge of the inner cube and has infinite valid displacements, and point [4,3,3] sits on a face of the inner cube and has 1 valid displacement (which I can add, it should be [radius, 0,0] )
There was a problem hiding this comment.
Noted -- what does the internal code do in this case? Is it left up to whichever is closer in floating point?
2 participants
Description
Motivation and Context
Types of Changes
1The change breaks (or has the potential to break) existing functionality.
Checklist: