Module attributed_graph 

Attributed graphs with typed node and edge features

This module provides AttributedGraph<N, E> – a graph where every node carries data of type N and every edge carries data of type E. It differs from the attribute-map approach in crate::attributes in that the feature types are statically known, enabling zero-cost extraction of dense feature matrices via a user-supplied projection function.

Key types

Type	Purpose
AttributedGraph<N,E>	Core graph container
AttributedGraphBuilder<N,E>	Ergonomic builder
NeighborInfo<N,E>	Neighbour + data bundle returned by attributed_neighbors

Free functions

• node_feature_matrix – project node data into an Array2<f64>
• edge_feature_matrix – project edge data into an Array2<f64>
• attributed_neighbors – enumerate neighbours with their data
• dijkstra_attributed – generalised Dijkstra using a caller-supplied cost function

Structs

AttributedGraph

A directed graph whose nodes carry data of type N and whose edges carry data of type E.

AttributedGraphBuilder

Builder for AttributedGraph that provides a fluent API.

NeighborInfo

A neighbour together with its node data and the connecting edge data.

NodeId

Stable integer node identifier.

Functions

attributed_neighbors

Return the out-neighbours of node together with their node data and the connecting edge data.

dijkstra_attributed

Generalised Dijkstra shortest-path algorithm on an attributed graph.

edge_feature_matrix

Construct a dense edge-feature matrix from an attributed graph.

filter_nodes

Filter nodes by a predicate on their data, returning matching NodeIds.

in_degrees

Compute the in-degree of each node.

node_feature_matrix

Construct a dense node-feature matrix from an attributed graph.

out_degrees

Compute the out-degree of each node.