bjorkegeek / polytri Goto Github PK

(As reported in an email from Charles Bradshaw)

I have been using your python polygon triangulate algorithm successfully for some time.

Mostly it works well, but if I present this polygon:

[ 1.0000, -0.5000,  1.0000]
[ 0.7500, -0.4330,  1.0000]
[ 0.5670, -0.2500,  1.0000]
[ 0.5000,  0.0000,  1.0000]
[ 0.5670,  0.2500,  1.0000]
[ 0.7500,  0.4330,  1.0000]
[ 1.0000,  0.5000,  1.0000]
[ 1.0000,  1.0000,  1.0000]
[-1.0000,  1.0000,  1.0000]
[-1.0000, -1.0000,  1.0000]
[ 1.0000, -1.0000,  1.0000]

I see the ValueError("Triangulation dot < 1E-20") , as the value of dot is zero. (I changed the value to 1E-20 to get a previous example to work)

I can make triangulate work by rotating the vertices so as to present some other point first. For example this:

[ 0.5000,  0.0000,  1.0000]
[ 0.5670,  0.2500,  1.0000]
[ 0.7500,  0.4330,  1.0000]
[ 1.0000,  0.5000,  1.0000]
[ 1.0000,  1.0000,  1.0000]
[-1.0000,  1.0000,  1.0000]
[-1.0000, -1.0000,  1.0000]
[ 1.0000, -1.0000,  1.0000]
[ 1.0000, -0.5000,  1.0000]
[ 0.7500, -0.4330,  1.0000]
[ 0.5670, -0.2500,  1.0000]

Produces a result. However it is wrong. Please see the attached showing the result of the latter data set. In the image the yellow lines and green points is the above polygon. The blue arrows show the winding order. The white lines are the triangles returned for your code. Some falling outside the original polygon!