Struct SemanticDocumentGraph 

pub struct SemanticDocumentGraph { /* private fields */ }

Graph of document relationships based on semantic similarity and explicit annotations (citations, cluster membership).

Implementations

impl SemanticDocumentGraph

pub fn new(similarity_threshold: f64) -> Self

Create a new graph with the given cosine-similarity threshold for automatic linking.

pub fn add_node( &mut self, doc_id: &str, embedding: Vec<f64>, metadata: HashMap<String, String>, )

Add a document node. If a node with the same doc_id already exists it is silently replaced (edges are not touched).

pub fn remove_node(&mut self, doc_id: &str) -> bool

Remove a node and all edges incident to it. Returns true if the node existed.

pub fn add_edge( &mut self, source: &str, target: &str, kind: EdgeKind, weight: f64, ) -> Result<(), String>

Add an explicit edge between two existing nodes. Returns an error if either endpoint does not exist.

pub fn auto_link_similar(&mut self)

Compute pairwise cosine similarity for all node pairs and add Similar edges for every pair whose similarity meets or exceeds the threshold.

Existing Similar edges are not removed first — call this on a fresh graph or manually prune duplicates if needed.

pub fn neighbors(&self, doc_id: &str) -> Vec<(&DocGraphNode, f64)>

Return neighbour nodes with their edge weights.

pub fn shortest_path(&self, from: &str, to: &str) -> Option<Vec<String>>

BFS shortest path (by hop count) from from to to. Returns the sequence of doc-ids including both endpoints, or None if no path exists.

pub fn connected_components(&self) -> Vec<Vec<String>>

Return the connected components of the (undirected) graph.

pub fn node_count(&self) -> usize

Number of nodes.

pub fn edge_count(&self) -> usize

Number of edges.

pub fn degree(&self, doc_id: &str) -> usize

Degree (number of incident edges) for a given node. Returns 0 if the node does not exist.

pub fn stats(&self) -> DocumentGraphStats

Aggregate statistics.