As the blog post says, degenerate polygons most of the time are zero area polys, a filter points with a very small distance have solved this kind of issues for me on the past, but since your geometry is deformed you have to check if it doesn't break your plotted shape or if it doesn't break other things.
On Tue, Mar 24, 2015 at 9:44 AM, Morten Bartholdy <[email protected]> wrote: > I have a scene which crashes more or less randomly and Arnold prints an > error that there are degenerate polygons in the scene. The geometry consist > of 3600 simple cubes which are imported from an alembic file generated in > Cinema4D, so manually examining every object is not really an option. So my > question is how to fix the degenerate polygons? I found this script: > http://xsisupport.com/2013/03/13/finding-degenerate-polygons-by-area/ , > but it fails with an error: > > > This is what I ran: > > > si = Application > epsilon = 0.00001 > > # Get PolygonArea DataArray (which is a tuple) > attr = si.Selection(0).ActivePrimitive.GetICEAttributeFromName( > "PolygonArea" ) > areaData = attr.DataArray > > # > # Find the indices of the bad polys > # > bad = [ x for x,y in enumerate( areaData ) if y < epsilon] > > # Select the degenerates with a string like 'cube.poly[112,114,155]' > si.SelectGeometryComponents( 'Null2.Cube_[%s]' % ','.join(str(i) for i in > bad) ) > > > > > > - and the error: > > > Application.SelectMembers("abc_Tiles", "", "") > # ERROR : 2000 - Argument 0 (SelectionList) is invalid > # ERROR : 2001-SELE-SelectGeometryComponents - Argument 0 is invalid > Application.SelectGeometryComponents("Null2.Cube_[]") > # ERROR : Traceback (most recent call last): > # File "<Script Block >", line 14, in <module> > # si.SelectGeometryComponents( 'Null2.Cube_[%s]' % ','.join(str(i) for i > in bad) ) > # File "<COMObject Application>", line 2, in SelectGeometryComponents > # COM Error: The parameter is incorrect. (0x-7ff8ffa9) - [line 14] > > > > > > Line 14 is this one: si.SelectGeometryComponents( 'Null2.Cube_[%s]' % > ','.join(str(i) for i in bad) ) > > > > So what might be wrong with this one, and when I know which polygons are > bad, how can I fix them? I am thinking since they dont have an ICE tree > they might have no ICE PolygonArea property... > > > > > > Best > > Morten > > > > >
