Function site_percolation 

pub fn site_percolation(
    graph: &Graph,
    vertex_order: &[VertexId],
) -> IgraphResult<SitePercolation>

Site percolation: the size of the largest component as vertices (sites) of graph are activated in the order specified by vertex_order.

Each activation step:

1. Mark the vertex active.
2. For every neighbor that is already active, add the connecting edge and union the two trees.
3. Record giant_size and edge_count after the step.

vertex_order must contain valid ids (< graph.vcount()) and must not repeat — duplicates return IgraphError::InvalidArgument; out-of-range ids return IgraphError::VertexOutOfRange. The order may be a strict subset of all vertices (skip = “never activated”).

Edge direction is ignored (matches upstream — percolation is purely a connectivity model). Parallel edges count separately; self-loops on the just-activated vertex contribute two to edge_count (the loop appears twice in the all-neighbor walk for IGRAPH_LOOPS semantics).

To shuffle into a random order, generate the permutation with your own RNG. We don’t take an RNG to keep the API deterministic and dependency-free.

Counterpart of igraph_site_percolation from references/igraph/src/connectivity/percolation.c:328.

Examples

use rust_igraph::{Graph, site_percolation};

// Path 0-1-2-3, activate vertices in id order.
let mut g = Graph::with_vertices(4);
g.add_edges(vec![(0u32, 1u32), (1, 2), (2, 3)]).unwrap();
let p = site_percolation(&g, &[0, 1, 2, 3]).unwrap();
// Activating 0: no edges added (no active neighbors), giant = 1.
// Activating 1: edge (0,1) activates, giant = 2.
// Activating 2: edge (1,2) activates, giant = 3.
// Activating 3: edge (2,3) activates, giant = 4.
assert_eq!(p.giant_size, vec![1, 2, 3, 4]);
assert_eq!(p.edge_count, vec![0, 1, 2, 3]);