oineus.diff.weak_alpha_filtration

oineus.diff.weak_alpha_filtration(points, packed=False, print_time=False)[source]

Build a differentiable weak-alpha filtration from a point cloud.

Parameters:

  • points – (n, d) torch.Tensor with d in {2, 3}. Differentiable.

  • packed (bool) – Use the compact bit-packed cell encoding for the Delaunay combinatorics when the vertex ids fit a 64/128-bit word. The values (and gradients) are recomputed here regardless of encoding.

  • print_time (bool) – If True, print per-stage timings.

Returns:

DiffFiltration whose values are squared longest-edge lengths.