Function compute_persistence 

pub fn compute_persistence(
    distance_matrix: &[Vec<f64>],
    max_dim: usize,
    max_epsilon: f64,
) -> Result<Vec<PersistenceDiagram>>

Compute persistent homology from a precomputed distance matrix.

Returns one PersistenceDiagram per homological dimension (H0, H1, …, up to max_dim). The filtration is the Vietoris-Rips filtration: a simplex enters at the maximum pairwise distance among its vertices.

This is equivalent to constructing a nested family of Vietoris-Rips complexes parameterised by epsilon ∈ [0, max_epsilon].

Arguments

• distance_matrix — symmetric n×n matrix of pairwise distances
• max_dim — maximum homological dimension to compute (typically 1 or 2)
• max_epsilon — upper bound on the filtration parameter

Example

use scirs2_transform::tda_ext::compute_persistence;

let dist = vec![
    vec![0.0, 1.0, 1.4, 1.0],
    vec![1.0, 0.0, 1.0, 1.4],
    vec![1.4, 1.0, 0.0, 1.0],
    vec![1.0, 1.4, 1.0, 0.0],
];
let diagrams = compute_persistence(&dist, 1, 2.0).expect("should succeed");
assert_eq!(diagrams.len(), 2); // H0 and H1
assert!(!diagrams[0].is_empty()); // at least one H0 feature