Struct SubgraphSummarizer 

pub struct SubgraphSummarizer;

Produces cluster-based summaries of KG subgraphs.

Implementations

impl SubgraphSummarizer

pub fn new() -> Self

Create a new summarizer (stateless).

pub fn summarize( &self, graph: &KgSubgraph, max_clusters: usize, ) -> Vec<SummaryCluster>

Summarise graph into at most max_clusters SummaryClusters.

Algorithm

1. Group nodes by node_type.
2. If there are more types than max_clusters, merge the smallest groups into an "Other" cluster so that only max_clusters remain.
3. For each cluster, count internal edges and pick the most-connected node (highest undirected degree within the whole graph) as the representative.

pub fn extract_key_relations( &self, graph: &KgSubgraph, top_n: usize, ) -> Vec<(String, usize)>

Return the top-top_n relation types by frequency of occurrence.

Ties are broken by relation name (alphabetical ascending).

pub fn node_degree(&self, graph: &KgSubgraph, node_id: &str) -> usize

Compute the undirected degree of node_id in graph.

Both outgoing and incoming edges are counted (a self-loop counts twice).

pub fn generate_text_summary(&self, clusters: &[SummaryCluster]) -> String

Generate a human-readable summary paragraph from a set of clusters.

The paragraph lists each cluster’s representative node, member count, and internal edge count, ending with the total number of clusters.

Trait Implementations

impl Default for SubgraphSummarizer

fn default() -> Self

Returns the “default value” for a type.