Crate ogun 

Ogun

Spatial layout generation via sequential logit dynamics on potential games.

Given a weighted graph and a 2D spatial domain, Ogun places nodes and routes edges at a controllable optimization level. The inverse temperature β governs output character — from near-random (low β) to near-optimal (high β).

Algorithm

for each agent in arrival order:
    1. EVAL_UTILITY  — score every valid grid position via potential Φ
    2. CHOOSE        — Boltzmann sample: P(p) ∝ exp(β · Φ(p))
    3. COMMIT        — place irrevocably, block footprint
    4. ROUTE         — BFS (Lee) to already-placed neighbors
score the completed layout

Quick start

use ogun::*;

let graph = Graph {
    nodes: vec![
        Node { id: NodeId(0), radius: 1 },
        Node { id: NodeId(1), radius: 1 },
    ],
    edges: vec![Edge {
        id: EdgeId(0),
        src: NodeId(0),
        dst: NodeId(1),
        weight: 1.0,
    }],
};
let space = Space { width: 20, height: 20, obstacles: vec![] };
let config = OgunConfig { seed: 42, ..OgunConfig::default() };

let layout = generate(&graph, &space, &config);
assert!(layout.score.composite > 0.0);

Structs

Edge

A weighted edge connecting two nodes.

EdgeId

Type-safe edge index.

Graph

The graph of nodes and edges to lay out.

Grid

A 2D grid stored as a flat Vec<T> in row-major order.

Layout

The result of running the Ogun algorithm.

Node

A node (agent) to be placed in the layout.

NodeId

Type-safe node index.

OgunConfig

Algorithm parameters.

Pos

Grid coordinate (column, row).

Rect

An axis-aligned rectangle used for obstacles and keep-out zones.

ScoreBreakdown

Per-metric breakdown of the composite layout score.

Space

The 2D spatial domain: boundary dimensions and obstacle list.

Functions

generate

Generate a spatial layout from a graph and spatial domain.