Function barabasi_game_bag 

pub fn barabasi_game_bag(
    n: u32,
    m: u32,
    outpref: bool,
    directed: bool,
    seed: u64,
) -> IgraphResult<Graph>

Generate a graph with n vertices via Barabási–Albert preferential attachment, using the original “bag” mechanism (BAG variant).

Each new vertex attaches m outgoing edges to existing vertices, chosen with probability proportional to their degree. Because draws are with replacement, the result may contain multi-edges (and self-loops when outpref = true).

• n — vertex count. n = 0 returns an empty graph.
• m — number of edges each newly-added vertex contributes. m = 0 yields an edgeless graph on n vertices.
• outpref — when true, the new vertex’s outgoing edges also bias subsequent sampling (i.e., out-degree counts toward attractiveness). Forced to true if directed = false (preferential attachment on an undirected graph naturally uses total degree).
• directed — generate a directed graph (edges point from new vertex to older).
• seed — initialises an internal SplitMix64 PRNG. Same (n, m, outpref, directed, seed) always yields the same graph.

Errors

Returns IgraphError::Internal if (n - 1) · m or the bag size overflows u64 (only possible on absurdly large inputs; (n, m) = (u32::MAX, u32::MAX) is the only realistic failure mode).

Examples

use rust_igraph::barabasi_game_bag;

// 100-vertex directed BA graph with m = 2 outgoing edges per new vertex.
// The first vertex (seed) has no outgoing edges, so the total edge
// count is exactly (n - 1) · m = 99 · 2 = 198.
let g = barabasi_game_bag(100, 2, false, true, 0xCAFEBEEF).unwrap();
assert_eq!(g.vcount(), 100);
assert_eq!(g.ecount(), 198);
assert!(g.is_directed());