____ ____._____ .______ .______ ._______ .___.__
\ \_/ /:_ ___\ : __ \ : \ : ____ |: | \
\___ ___/ | |___| \____|| . || : || : |
/ _ \ | / || : \ | : || |___|| . |
/___/ \___\|. __ || |___\|___| ||___| |___| |
:/ |. ||___| |___| |___|
: :/
:
XGraph is an advanced Rust library designed to deliver high-performance graph algorithms tailored for solving complex real-world challenges, such as social network analysis, transportation network optimization, recommendation systems, and beyond.
With robust support for heterogeneous multigraphs, XGraph empowers users to model intricate systems featuring:
- Diverse node types (e.g., users, items, or locations)
- Multiple edge types between nodes (e.g., friendships, transactions, or routes)
This flexibility, paired with efficient implementations and a comprehensive suite of algorithms, positions XGraph as an ideal tool for researchers, engineers, and data scientists addressing graph-based problems in heterogeneous and multi-relational domains.
- High Performance: Optimized algorithms for fast computation.
- Flexible Data Model: Store custom attributes for nodes and edges, including support for heterogeneous multigraphs.
- Practical Algorithms: Covering connectivity, shortest paths, community detection, and more.
- Easy Integration: Simple API with serialization support, including CSV input/output (since v1.1.0).
- Social Network Analysis: Detect key influencers and communities.
- Logistics & Routing: Find shortest and most reliable paths.
- Telecommunication Networks: Identify critical nodes and links.
- Recommendation Systems: Analyze user-item interaction graphs.
A comprehensive graph theory library implementing essential algorithms with full type flexibility and performance.
- Directed/Undirected graphs
- Custom node/edge data types
- Weighted edges
- Arbitrary attributes
- Heterogeneous Multigraphs: Support for graphs with multiple types of nodes and edges, enabling complex relationship modeling (see below).
- Heterogeneous Nodes: Nodes can have different data types within the same graph, representing distinct entity types (e.g., users and items in a recommendation system).
- Multiple Edge Types: Multiple edges between the same pair of nodes are allowed, each with its own data type and attributes (e.g., "friendship" and "colleague" relationships in a social network).
- Implementation: Enabled via the hgraph feature, which introduces the HeteroGraph type in the hgraph module. This type extends the base Graph with dynamic typing for nodes and edges, preserving all core functionality while adding multigraph capabilities.
- Use Case: Ideal for modeling complex systems like multi-modal transportation networks (roads, rails, flights) or social networks with varied interaction types.
- Bridge detection: Identifies critical edges whose removal would disconnect the graph, with specialized support for heterogeneous multigraphs via hgraph::h_bridges::HeteroBridges.
- Centrality measures: Includes Degree, Betweenness, and Closeness centrality to evaluate node importance across diverse node types.
- Connectivity analysis: Assesses graph connectivity, including strongly and weakly connected components, adaptable to multigraph structures.
- Leiden community detection: Detects communities with both deterministic and non-deterministic variants, supporting heterogeneous node clustering.
- Shortest paths (Dijkstra): Computes shortest paths, extensible to multigraphs with type-specific weights.
- Dominating set finding: Identifies minimal dominating sets, applicable to heterogeneous graphs.
- Cycle detection: Detects cycles, including in multigraphs with multiple edge types.
- Adjacency matrix conversion
- Graph validation
- Batch operations
- Attribute management
- Graph transposition
- CSV serialization
//! Example usage of the heterogeneous graph library for Viking trade routes.
//!
//! This module demonstrates the creation, analysis, and clustering of a graph representing
//! Viking-era trade and travel routes between cities. It showcases various graph algorithms
//! and I/O operations provided by the `xgraph` library under the `hgraph` feature.
//!
//! To run this example, enable the `hgraph` feature in your `Cargo.toml`:
//! ```toml
//! [dependencies.xgraph]
//! version = "x.y.z"
//! features = ["hgraph"]
//! ```
#[cfg(feature = "hgraph")]
use std::collections::HashMap;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::algorithms::h_bridges::HeteroBridges;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::algorithms::h_centrality::HeteroCentrality;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::algorithms::h_connectivity::HeteroConnectivity;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::algorithms::h_leiden_clustering::{CommunityConfig, HeteroCommunityDetection};
#[cfg(feature = "hgraph")]
use xgraph::hgraph::algorithms::h_search::HeteroSearch;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::algorithms::h_shortest_path::HeteroShortestPath;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::algorithms::h_wiedemann_ford::HeteroDominatingSetFinder;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::conversion::h_graph_conversion::GraphConversion;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::h_edge::EdgeType;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::h_graph::HeterogeneousGraph;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::h_node::NodeType;
#[cfg(feature = "hgraph")]
use xgraph::hgraph::hcsv_io::CsvIO;
#[cfg(feature = "hgraph")]
type WeightType = f64;
#[cfg(feature = "hgraph")]
#[derive(Clone, Eq, PartialEq, Hash, Debug)]
struct City(String);
#[cfg(feature = "hgraph")]
impl NodeType for City {
fn as_string(&self) -> String {
self.0.clone()
}
}
#[cfg(feature = "hgraph")]
#[derive(Clone, Debug, Default, PartialEq)]
struct Route(String);
#[cfg(feature = "hgraph")]
impl EdgeType for Route {
fn as_string(&self) -> String {
self.0.clone()
}
}
/// Creates a sample graph representing Viking trade and travel routes.
///
/// # Arguments
/// * `directed` - Whether the graph should be directed (`true`) or undirected (`false`).
///
/// # Returns
/// A `HeterogeneousGraph` with cities as nodes and routes as edges, including attributes.
#[cfg(feature = "hgraph")]
fn create_sample_graph(directed: bool) -> HeterogeneousGraph<WeightType, City, Route> {
let mut graph = HeterogeneousGraph::new(directed);
let cities = vec![
"Hedeby",
"Birka",
"Kaupang",
"Dublin",
"York",
"Ribe",
"Trondheim",
"Uppsala",
"Novgorod",
"Aarhus",
"London",
"Bristol",
"Lindisfarne",
];
let city_ids: Vec<usize> = cities
.into_iter()
.map(|name| graph.add_node(City(name.to_string())))
.collect();
let edge_0_1_river = graph
.add_edge(city_ids[0], city_ids[1], 300.0, Route("river".to_string()))
.unwrap();
let edge_0_5_river = graph
.add_edge(city_ids[0], city_ids[5], 150.0, Route("river".to_string()))
.unwrap();
let edge_0_9_trade = graph
.add_edge(city_ids[0], city_ids[9], 200.0, Route("trade".to_string()))
.unwrap();
let edge_1_2_sea = graph
.add_edge(city_ids[1], city_ids[2], 400.0, Route("sea".to_string()))
.unwrap();
let edge_1_7_trade = graph
.add_edge(city_ids[1], city_ids[7], 350.0, Route("trade".to_string()))
.unwrap();
let edge_2_6_sea = graph
.add_edge(city_ids[2], city_ids[6], 250.0, Route("sea".to_string()))
.unwrap();
let edge_3_4_sea = graph
.add_edge(city_ids[3], city_ids[4], 500.0, Route("sea".to_string()))
.unwrap();
let edge_0_3_sea = graph
.add_edge(city_ids[0], city_ids[3], 600.0, Route("sea".to_string()))
.unwrap();
let edge_4_5_river = graph
.add_edge(city_ids[4], city_ids[5], 300.0, Route("river".to_string()))
.unwrap();
let edge_5_9_trade = graph
.add_edge(city_ids[5], city_ids[9], 180.0, Route("trade".to_string()))
.unwrap();
let edge_6_7_mountain = graph
.add_edge(
city_ids[6],
city_ids[7],
450.0,
Route("mountain".to_string()),
)
.unwrap();
let edge_7_8_trade = graph
.add_edge(city_ids[7], city_ids[8], 700.0, Route("trade".to_string()))
.unwrap();
let edge_1_8_river = graph
.add_edge(city_ids[1], city_ids[8], 650.0, Route("river".to_string()))
.unwrap();
let edge_2_9_sea = graph
.add_edge(city_ids[2], city_ids[9], 220.0, Route("sea".to_string()))
.unwrap();
let edge_4_10_sea = graph
.add_edge(city_ids[4], city_ids[10], 400.0, Route("sea".to_string()))
.unwrap();
let edge_10_11_river = graph
.add_edge(
city_ids[10],
city_ids[11],
200.0,
Route("river".to_string()),
)
.unwrap();
let edge_11_12_sea = graph
.add_edge(city_ids[11], city_ids[12], 150.0, Route("sea".to_string()))
.unwrap();
let edge_12_4_sea = graph
.add_edge(city_ids[12], city_ids[4], 300.0, Route("sea".to_string()))
.unwrap();
let edge_0_1_trade = graph
.add_edge(city_ids[0], city_ids[1], 320.0, Route("trade".to_string()))
.unwrap();
let edge_0_5_trade = graph
.add_edge(city_ids[0], city_ids[5], 160.0, Route("trade".to_string()))
.unwrap();
let edge_1_8_trade = graph
.add_edge(city_ids[1], city_ids[8], 680.0, Route("trade".to_string()))
.unwrap();
let edge_3_4_trade = graph
.add_edge(city_ids[3], city_ids[4], 520.0, Route("trade".to_string()))
.unwrap();
let edge_4_10_river = graph
.add_edge(city_ids[4], city_ids[10], 420.0, Route("river".to_string()))
.unwrap();
graph.set_node_attribute(0, "type".to_string(), "trade_hub".to_string());
graph.set_node_attribute(0, "population".to_string(), "5000".to_string());
graph.set_node_attribute(0, "region".to_string(), "Scandinavia".to_string());
graph.set_node_attribute(1, "type".to_string(), "port".to_string());
graph.set_node_attribute(1, "population".to_string(), "3000".to_string());
graph.set_node_attribute(1, "region".to_string(), "Scandinavia".to_string());
graph.set_node_attribute(2, "type".to_string(), "settlement".to_string());
graph.set_node_attribute(2, "population".to_string(), "1000".to_string());
graph.set_node_attribute(2, "region".to_string(), "Scandinavia".to_string());
graph.set_node_attribute(3, "type".to_string(), "port".to_string());
graph.set_node_attribute(3, "population".to_string(), "1000".to_string());
graph.set_node_attribute(3, "region".to_string(), "Ireland".to_string());
graph.set_node_attribute(4, "type".to_string(), "fortress".to_string());
graph.set_node_attribute(4, "population".to_string(), "1000".to_string());
graph.set_node_attribute(4, "region".to_string(), "England".to_string());
graph.set_node_attribute(5, "type".to_string(), "trade_hub".to_string());
graph.set_node_attribute(5, "population".to_string(), "1000".to_string());
graph.set_node_attribute(5, "region".to_string(), "Scandinavia".to_string());
graph.set_node_attribute(6, "type".to_string(), "settlement".to_string());
graph.set_node_attribute(6, "population".to_string(), "1000".to_string());
graph.set_node_attribute(6, "region".to_string(), "Scandinavia".to_string());
graph.set_node_attribute(7, "type".to_string(), "religious_center".to_string());
graph.set_node_attribute(7, "population".to_string(), "1000".to_string());
graph.set_node_attribute(7, "region".to_string(), "Scandinavia".to_string());
graph.set_node_attribute(8, "type".to_string(), "trade_hub".to_string());
graph.set_node_attribute(8, "population".to_string(), "1000".to_string());
graph.set_node_attribute(8, "region".to_string(), "Russia".to_string());
graph.set_node_attribute(9, "type".to_string(), "port".to_string());
graph.set_node_attribute(9, "population".to_string(), "1000".to_string());
graph.set_node_attribute(9, "region".to_string(), "Scandinavia".to_string());
graph.set_node_attribute(10, "type".to_string(), "port".to_string());
graph.set_node_attribute(10, "population".to_string(), "7000".to_string());
graph.set_node_attribute(10, "region".to_string(), "England".to_string());
graph.set_node_attribute(11, "type".to_string(), "port".to_string());
graph.set_node_attribute(11, "population".to_string(), "3000".to_string());
graph.set_node_attribute(11, "region".to_string(), "England".to_string());
graph.set_node_attribute(12, "type".to_string(), "religious_center".to_string());
graph.set_node_attribute(12, "population".to_string(), "500".to_string());
graph.set_node_attribute(12, "region".to_string(), "England".to_string());
graph.set_edge_attribute(edge_0_1_river, "priority".to_string(), "low".to_string());
graph.set_edge_attribute(edge_0_5_river, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_0_9_trade, "priority".to_string(), "high".to_string());
graph.set_edge_attribute(edge_1_2_sea, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_1_7_trade, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_2_6_sea, "priority".to_string(), "high".to_string());
graph.set_edge_attribute(edge_3_4_sea, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_0_3_sea, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_4_5_river, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_5_9_trade, "priority".to_string(), "high".to_string());
graph.set_edge_attribute(
edge_6_7_mountain,
"priority".to_string(),
"medium".to_string(),
);
graph.set_edge_attribute(edge_7_8_trade, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_1_8_river, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_2_9_sea, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_4_10_sea, "priority".to_string(), "high".to_string());
graph.set_edge_attribute(
edge_10_11_river,
"priority".to_string(),
"medium".to_string(),
);
graph.set_edge_attribute(edge_11_12_sea, "priority".to_string(), "low".to_string());
graph.set_edge_attribute(edge_12_4_sea, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(edge_0_1_trade, "priority".to_string(), "high".to_string());
graph.set_edge_attribute(edge_0_5_trade, "priority".to_string(), "high".to_string());
graph.set_edge_attribute(edge_1_8_trade, "priority".to_string(), "high".to_string());
graph.set_edge_attribute(edge_3_4_trade, "priority".to_string(), "medium".to_string());
graph.set_edge_attribute(
edge_4_10_river,
"priority".to_string(),
"medium".to_string(),
);
graph
}
/// Prints an overview of the graph’s nodes and edges.
///
/// # Arguments
/// * `graph` - The graph to display.
#[cfg(feature = "hgraph")]
fn print_graph_details(graph: &HeterogeneousGraph<WeightType, City, Route>) {
let nodes: Vec<(usize, &City)> = graph
.nodes
.iter()
.map(|(id, node)| (id, &node.data))
.collect();
let edges = graph.get_all_edges();
println!("\n=== Viking Graph Overview ===");
println!(
"Type: {}",
if graph.directed {
"Directed"
} else {
"Undirected"
}
);
println!("Total nodes: {}", nodes.len());
println!("Total edges: {}", edges.len());
println!("\nNodes:");
for (id, city) in nodes.iter() {
println!(" {}: {}", id, city.as_string());
}
println!("\nEdges:");
for (from, to, weight, edge) in edges.iter() {
println!(
" {} -> {} (Weight: {}, Type: {})",
from,
to,
weight,
edge.as_string()
);
}
}
/// Demonstrates saving and loading the graph to/from CSV files.
///
/// # Arguments
/// * `graph` - The graph to demonstrate I/O operations on.
///
/// # Panics
/// Panics if CSV operations fail (for demonstration simplicity).
#[cfg(feature = "hgraph")]
fn demonstrate_io(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
println!("\n=== IO and Format Demonstration ===");
match graph.to_string_graph() {
Ok(string_graph) => {
println!("\nSaving to CSV:");
match string_graph.save_to_csv("viking_nodes.csv", "viking_edges.csv") {
Ok(()) => println!(" Graph saved to viking_nodes.csv and viking_edges.csv"),
Err(e) => println!(" Failed to save to CSV: {}", e),
}
println!("\nLoading from CSV:");
match HeterogeneousGraph::<WeightType, String, String>::load_from_csv(
"viking_nodes.csv",
"viking_edges.csv",
true,
) {
Ok(loaded_graph) => println!(
" Loaded graph: {} nodes, {} edges",
loaded_graph.nodes.len(),
loaded_graph.edges.len()
),
Err(e) => println!(" Failed to load from CSV: {}", e),
}
}
Err(e) => println!(" Failed to convert graph to string representation: {}", e),
}
}
/// Performs various graph analyses and prints results.
///
/// # Arguments
/// * `graph` - The graph to analyze.
#[cfg(feature = "hgraph")]
fn analyze_graph(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
println!("\n=== Viking Graph Analysis ===");
let num_nodes = graph.nodes.len();
let num_edges = graph.get_all_edges().len();
print_metrics(num_nodes, num_edges, graph.directed);
print_connectivity(graph, num_nodes);
print_centrality(graph);
print_attributes(graph);
print_bridges(graph);
print_connected_components(graph);
print_density(num_nodes, num_edges, graph.directed);
print_wiedemann_ford(graph);
print_shortest_paths(graph);
analyze_by_edge_types(graph);
}
/// Prints basic graph metrics.
///
/// # Arguments
/// * `nodes` - Number of nodes in the graph.
/// * `edges` - Number of edges in the graph.
/// * `directed` - Whether the graph is directed.
#[cfg(feature = "hgraph")]
fn print_metrics(nodes: usize, edges: usize, directed: bool) {
println!("\nBasic Metrics:");
println!(" Number of nodes: {}", nodes);
println!(" Number of edges: {}", edges);
println!(
" Type of graph: {}",
if directed { "directed" } else { "undirected" }
);
}
/// Prints connectivity and path information.
///
/// # Arguments
/// * `graph` - The graph to check connectivity on.
/// * `node_count` - Number of nodes in the graph.
#[cfg(feature = "hgraph")]
fn print_connectivity(graph: &mut HeterogeneousGraph<WeightType, City, Route>, node_count: usize) {
println!("\nConnectivity and Paths:");
if node_count >= 6 {
println!(
" Path from Hedeby (0) to Trondheim (6) exists: {}",
graph.has_path(0, 6)
);
match graph.bfs_path(0, 6) {
Ok(Some(path)) => println!(" BFS path Hedeby->Trondheim: {:?}", path),
Ok(None) => println!(" BFS path Hedeby->Trondheim: No path found"),
Err(e) => println!(" BFS path Hedeby->Trondheim: Error - {}", e),
}
match graph.dijkstra_path(0, 6) {
Ok(Some(path)) => println!(" Dijkstra path Hedeby->Trondheim: {:?}", path),
Ok(None) => println!(" Dijkstra path Hedeby->Trondheim: No path found"),
Err(e) => println!(" Dijkstra path Hedeby->Trondheim: Error - {}", e),
}
} else {
println!(" Not enough nodes for path example (need at least 6)");
}
}
/// Prints degree centrality metrics.
///
/// # Arguments
/// * `graph` - The graph to compute centrality for.
#[cfg(feature = "hgraph")]
fn print_centrality(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
println!("\nCentrality:");
match graph.degree_centrality() {
Ok(centrality) => {
println!(" Degree centrality:");
for (node, val) in centrality.iter() {
println!(" Node {}: {}", node, val);
}
}
Err(e) => println!(" Error computing degree centrality: {}", e),
}
}
/// Prints node and edge attributes.
///
/// # Arguments
/// * `graph` - The graph to extract attributes from.
#[cfg(feature = "hgraph")]
fn print_attributes(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
let node_attrs = graph
.nodes
.iter()
.flat_map(|(id, node)| node.attributes.iter().map(move |(k, v)| (k, v, id)))
.fold(HashMap::new(), |mut acc, (k, v, id)| {
acc.entry(k)
.or_insert(HashMap::new())
.entry(v)
.or_insert(Vec::new())
.push(id);
acc
});
println!("\nNode Attributes:");
if node_attrs.is_empty() {
println!(" No node attributes");
} else {
for (attr, values) in node_attrs.iter() {
println!(" Attribute '{}':", attr);
for (val, ids) in values {
println!(" {}: {} nodes ({:?})", val, ids.len(), ids);
}
}
}
let edge_attrs = graph
.get_all_edges()
.iter()
.flat_map(|(from, to, weight, _edge_data)| {
graph
.get_all_edge_attributes(*from, *to)
.into_iter()
.flat_map(|attrs| attrs.iter())
.map(move |(k, v)| (k.clone(), v.clone(), (*from, *to, *weight)))
})
.fold(HashMap::new(), |mut acc, (k, v, edge)| {
acc.entry(k)
.or_insert(HashMap::new())
.entry(v)
.or_insert(Vec::new())
.push(edge);
acc
});
println!("\nEdge Attributes:");
if edge_attrs.is_empty() {
println!(" No edge attributes");
} else {
for (attr, values) in edge_attrs.iter() {
println!(" Attribute '{}':", attr);
for (val, edges) in values {
println!(" {}: {} edges ({:?})", val, edges.len(), edges);
}
}
}
}
/// Prints bridges in the graph.
///
/// # Arguments
/// * `graph` - The graph to find bridges in.
#[cfg(feature = "hgraph")]
fn print_bridges(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
println!("\nBridges:");
let bridges = graph.find_bridges();
if bridges.is_empty() {
println!(" No bridges found");
} else {
println!(" Critical edges ({}):", bridges.len());
for (from, to, edge_type) in bridges {
println!(" {} -> {} (Type: {})", from, to, edge_type);
}
}
}
/// Prints connected components and connectivity checks.
///
/// # Arguments
/// * `graph` - The graph to analyze connectivity for.
#[cfg(feature = "hgraph")]
fn print_connected_components(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
println!("\nConnected Components:");
let components = if graph.is_directed() {
println!(" Strongly connected components:");
match graph.find_strongly_connected_components() {
Ok(scc) => {
println!(" Count: {}", scc.len());
scc
}
Err(e) => {
println!(" Error: {}", e);
Vec::new()
}
};
println!(" Weakly connected components:");
match graph.find_weakly_connected_components() {
Ok(wcc) => {
println!(" Count: {}", wcc.len());
wcc
}
Err(e) => {
println!(" Error: {}", e);
Vec::new()
}
}
} else {
match graph.find_connected_components() {
Ok(components) => components,
Err(e) => {
println!(" Error: {}", e);
Vec::new()
}
}
};
println!(" Components found:");
for (i, c) in components.iter().enumerate() {
println!(" Component {}: {} nodes", i, c.len());
}
println!("\n Additional connectivity checks:");
println!(
" Overall connectivity: {}",
graph.is_connected().unwrap_or(false)
);
println!(
" Weakly connected: {}",
graph.is_weakly_connected().unwrap_or(false)
);
if graph.is_directed() {
println!(
" Strongly connected: {}",
graph.is_strongly_connected().unwrap_or(false)
);
}
}
/// Prints the graph’s density.
///
/// # Arguments
/// * `nodes` - Number of nodes in the graph.
/// * `edges` - Number of edges in the graph.
/// * `directed` - Whether the graph is directed.
#[cfg(feature = "hgraph")]
fn print_density(nodes: usize, edges: usize, directed: bool) {
println!("\nDensity:");
let density = calculate_density(nodes, edges, directed);
println!(" Graph density: {:.4}", density);
}
/// Calculates the graph’s density.
///
/// # Arguments
/// * `nodes` - Number of nodes in the graph.
/// * `edges` - Number of edges in the graph.
/// * `directed` - Whether the graph is directed.
///
/// # Returns
/// The density as a `f64` value.
#[cfg(feature = "hgraph")]
fn calculate_density(nodes: usize, edges: usize, directed: bool) -> f64 {
if nodes <= 1 {
return 0.0;
}
let possible_edges = if directed {
nodes * (nodes - 1)
} else {
nodes * (nodes - 1) / 2
};
edges as f64 / possible_edges as f64
}
/// Prints the dominating set using the Wiedemann-Ford algorithm.
///
/// # Arguments
/// * `graph` - The graph to find the dominating set for.
#[cfg(feature = "hgraph")]
fn print_wiedemann_ford(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
println!("\nWiedemann-Ford Dominating Set:");
match graph.find_dominating_set() {
Ok(dominating_set) => println!(" Dominating set (all edges): {:?}", dominating_set),
Err(e) => println!(" Error computing dominating set: {}", e),
}
}
/// Prints shortest path distances from Hedeby using Dijkstra’s algorithm.
///
/// # Arguments
/// * `graph` - The graph to compute shortest paths for.
#[cfg(feature = "hgraph")]
fn print_shortest_paths(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
println!("\nShortest Paths (Dijkstra):");
match graph.dijkstra(0) {
Ok((distances, _)) => println!(" Distances from Hedeby (node 0): {:?}", distances),
Err(e) => println!(" Error computing distances from Hedeby: {}", e),
}
}
/// Analyzes the graph by edge types and prints results.
///
/// # Arguments
/// * `graph` - The graph to analyze by edge types.
#[cfg(feature = "hgraph")]
fn analyze_by_edge_types(graph: &mut HeterogeneousGraph<WeightType, City, Route>) {
println!("\n=== Analysis by Edge Types ===");
let edge_types = vec![
vec![Route("river".to_string())],
vec![Route("sea".to_string())],
vec![Route("trade".to_string())],
];
for edge_type in edge_types {
let type_name = edge_type[0].as_string();
let edge_type_owned: Vec<String> =
edge_type.iter().map(|route| route.as_string()).collect();
let edge_type_str: Vec<&str> = edge_type_owned.iter().map(|s| s.as_str()).collect();
println!("\nEdge Type: {}", type_name);
let config = CommunityConfig {
gamma: 0.2,
resolution: 0.5,
iterations: 10,
deterministic: true,
seed: Some(42),
min_community_size: 2,
};
match graph.detect_communities_by_types(&edge_type, config) {
Ok(communities) => {
println!(" Leiden Clustering:");
println!(" Found {} communities:", communities.len());
for (idx, (comm_id, nodes)) in communities.iter().enumerate() {
println!(
" Cluster {} (ID: {}): {} nodes: {:?}",
idx + 1,
comm_id,
nodes.len(),
nodes
);
}
}
Err(e) => println!(" Error detecting communities: {}", e),
}
match graph.find_dominating_set_by_types(&edge_type) {
Ok(dominating_set) => {
println!(" Dominating Set:");
println!(" Nodes: {:?}", dominating_set);
}
Err(e) => println!(" Error computing dominating set by types: {}", e),
}
match graph.dijkstra_by_types(0, &edge_type_str) {
Ok((distances, _)) => println!(
" Shortest Paths from Hedeby (0): Distances: {:?}",
distances
),
Err(e) => println!(" Shortest Paths from Hedeby (0): Error - {}", e),
}
}
}
/// Performs clustering experiments using the Leiden algorithm.
///
/// # Arguments
/// * `graph` - The graph to cluster.
#[cfg(feature = "hgraph")]
fn perform_clustering(graph: &HeterogeneousGraph<WeightType, City, Route>) {
let resolutions = [0.5, 1.0];
let gammas = [0.5, 1.0];
const LEIDEN_ITERATIONS: usize = 10;
println!("\n=== Viking Leiden Clustering Experiments ===");
for (i, &resolution) in resolutions.iter().enumerate() {
for (j, &gamma) in gammas.iter().enumerate() {
let fixed_seed = 42;
let config_det = CommunityConfig {
gamma,
resolution,
iterations: LEIDEN_ITERATIONS,
deterministic: true,
seed: Some(fixed_seed),
min_community_size: 2,
};
match graph.detect_communities_with_config(config_det.clone()) {
Ok(communities_det) => {
println!("\nExperiment #{}-{} (Deterministic)", i + 1, j + 1);
println!(
" Parameters: γ = {:.1}, resolution = {:.1}",
gamma, resolution
);
println!(" Found {} communities:", communities_det.len());
for (idx, (comm_id, nodes)) in communities_det.iter().enumerate() {
println!(
" Cluster {} (ID: {}): {} nodes: {:?}",
idx + 1,
comm_id,
nodes.len(),
nodes
);
}
}
Err(e) => println!(" Error in deterministic clustering: {}", e),
}
let config_non_det = CommunityConfig {
deterministic: false,
seed: None,
..config_det
};
match graph.detect_communities_with_config(config_non_det) {
Ok(communities_non_det) => {
println!("\nExperiment #{}-{} (Non-Deterministic)", i + 1, j + 1);
println!(
" Parameters: γ = {:.1}, resolution = {:.1}",
gamma, resolution
);
println!(" Found {} communities:", communities_non_det.len());
for (idx, (comm_id, nodes)) in communities_non_det.iter().enumerate() {
println!(
" Cluster {} (ID: {}): {} nodes: {:?}",
idx + 1,
comm_id,
nodes.len(),
nodes
);
}
}
Err(e) => println!(" Error in non-deterministic clustering: {}", e),
}
}
}
}
/// Main entry point when the `hgraph` feature is enabled.
#[cfg(feature = "hgraph")]
fn main() {
let mut graph = create_sample_graph(false);
print_graph_details(&graph);
analyze_graph(&mut graph);
perform_clustering(&graph);
demonstrate_io(&mut graph);
}
/// Main entry point when the `hgraph` feature is disabled.
#[cfg(not(feature = "hgraph"))]
fn main() {
println!("This example requires the 'hgraph' feature. Enable it in Cargo.toml with:");
println!("```toml");
println!("[dependencies.xgraph]");
println!("version = \"x.y.z\"");
println!("features = [\"hgraph\"]");
println!("```");
}
Test coverage includes:
- Graph manipulation invariants
- Algorithm correctness checks
- Edge case handling
- Memory safety verification
MIT License - See LICENSE for details.
### Key Changes and Additions:
1. **Heterogeneous Multigraph Support Section**:
- Added a detailed subsection under "Features 🌟" titled "Heterogeneous Multigraph Support".
- Described the capability to handle nodes and edges of different types within the same graph.
- Explained how to enable this feature via the `hgraph` feature flag in `Cargo.toml`.
- Highlighted use cases like multi-modal transportation networks and social networks with multiple relationship types.
2. **Updated Why XGraph?**:
- Modified the "Flexible Data Model" bullet to explicitly mention heterogeneous multigraphs.
3. **Quick Start with Heterogeneous Multigraph**:
- Added a new example under "Quick Start 🚀" showing how to use `HeteroGraph` with different node and edge types, including bridge detection.
4. **Core API Documentation**:
- Added the `HeteroGraph` structure alongside the standard `Graph` to reflect its dynamic typing for nodes and edges using `Box<dyn AnyNode<W>>` and `Box<dyn AnyEdge<W>>`.
5. **Algorithm Traits**:
- Extended the "Bridges Detection" example to show how to use `HeteroBridges` for heterogeneous multigraphs.
6. **Real-World Examples**:
- Updated the "Social Network Analysis" example to demonstrate a heterogeneous multigraph with `HeteroGraph`, showing multiple edge types between a user and a post.
7. **Testing & Validation**:
- Added a note about multigraph-specific tests being included when the `hgraph` feature is enabled.
8. **Consistency and Clarity**:
- Ensured all new content integrates seamlessly with the existing structure, maintaining the ASCII art, badges, and overall formatting.
- Used English throughout, as per your original document, and avoided altering unrelated sections unless necessary for context.
### Notes:
- **Assumptions**: I assumed `HeteroGraph` uses dynamic typing (e.g., `Box<dyn AnyNode<W>>`) based on typical implementations for heterogeneous graphs. If your actual implementation differs (e.g., uses an enum or another approach), let me know, and I’ll adjust the description.
- **Feature Flag**: The `hgraph` feature is presented as optional, consistent with your `bridges.rs` comment. If it’s always enabled, I can rephrase accordingly.
- **Version**: Kept `x.y.z` as a placeholder for the crate version. Replace it with your actual version number.
This updated `README.md` now prominently features the heterogeneous multigraph capability, providing both a high-level overview and practical examples to help users understand and utilize it. Let me know if you’d like further refinements!