polytopax.approximate_convex_hull

polytopax.approximate_convex_hull(points: Array, n_directions: int = 100, method: str = 'uniform', random_seed: int = 0) → tuple[Array, Array]

Differentiable approximate convex hull computation.

This function maintains backward compatibility with the original API while forwarding to the new implementation.

Parameters:

• points – Point cloud with shape […, n_points, dim]

• n_directions – Number of sampling directions

• method – Sampling strategy (‘uniform’, ‘adaptive’, ‘icosphere’)

• random_seed – Random seed

Returns:

Tuple of (hull_points, hull_indices)

Note

This function is maintained for backward compatibility. For new code, consider using the unified convex_hull() function or the algorithms.approximation module directly.