I have an array of points in python making a square lattice. I want to triangulate it like this:
so that the braces alternate in a checkerboard pattern from square to square. My attempts have been based on deforming it before triangulation. For example, shearing the lattice before triangulating via
xy_skew = np.dstack((xypts[:,0]+ 0.1*xypts[:,1], xypts[:,1]))[0]
tri = scipy.spatial.Delaunay(xy_skew)
TRI = tri.vertices
can give me all 'rightward' diagonals or all 'leftward' diagonals, but I haven't found a deformation that can lead to the desired triangulation.
How can I do this efficiently (for lattices of ~million points)?
If relevant, the indices of my points increase in Y first, then increase in X.
Thanks!
