Skip to contents

Force-directed graph layout (Fruchterman-Reingold)

Source: R/core-network.R
force_layout.Rd

A pure-R implementation of the Fruchterman-Reingold force-directed layout, so geom_sketch_node() / geom_sketch_edge() can place a network with no external graph dependency. Repulsive forces push every node apart; attractive forces pull edge-connected nodes together; a cooling schedule settles the system. Coordinates are returned rescaled to roughly [-1, 1] on both axes.

Usage

force_layout(edges, n_nodes = NULL, niter = 500L, seed = NULL)

Arguments

edges

A two-column matrix or data frame of 1-based integer node indices, one row per edge (from, to). May have zero rows (an edgeless graph, laid out on a circle).

n_nodes

Number of nodes. Defaults to the largest index in edges (so isolated high-index nodes need this set explicitly).

niter

Number of iterations. Default 500.

seed

Integer seed for the initial placement. NULL uses getOption("ggsketch.seed", 1L). The layout is otherwise deterministic.

Value

A data frame with columns x and y, one row per node in index order.

Examples

# A small ring of five nodes
e <- cbind(1:5, c(2:5, 1))
force_layout(e, seed = 1L)
#>            x           y
#> 1  1.0000000 -0.01430852
#> 2  0.2610357  1.00000000
#> 3 -0.9765955  0.64052762
#> 4 -1.0000000 -0.59521236
#> 5  0.2210999 -1.00000000