Struct SemanticContentRouter 

pub struct SemanticContentRouter { /* private fields */ }

Routes content queries to the most semantically relevant nodes.

Maintains a registry of TopicEmbeddings and, for each query, computes cosine similarity against every registered embedding, filters by min_similarity, ranks by combined_score, and caps results at max_candidates.

Example

use ipfrs_semantic::content_router::{
    RouterConfig, SemanticContentRouter, TopicEmbedding,
};

let config = RouterConfig { min_similarity: 0.5, max_candidates: 5 };
let mut router = SemanticContentRouter::new(config);

let emb = TopicEmbedding::new("node-1", "science", vec![1.0, 0.0, 0.0], 100);
router.register_topic(emb);

let decision = router.route(&[0.9, 0.1, 0.0]);
assert!(decision.best_node.is_some());

Implementations

impl SemanticContentRouter

pub fn new(config: RouterConfig) -> Self

Creates a new router with the given configuration.

pub fn register_topic(&mut self, topic: TopicEmbedding)

Registers a topic embedding.

Multiple topics from the same node are allowed.

pub fn update_load(&mut self, node_id: &str, topic: &str, load: usize)

Updates the current load for a specific (node, topic) pair.

If no matching entry is found, this is a no-op.

pub fn route(&mut self, query_embedding: &[f32]) -> RoutingDecision

Routes a query embedding to the most suitable nodes.

Steps:

1. Compute cosine similarity to every registered topic embedding.
2. Filter candidates whose similarity is below min_similarity.
3. Sort remaining candidates by combined_score (descending).
4. Retain at most max_candidates results.
5. Update statistics and return the RoutingDecision.

pub fn unregister_node(&mut self, node_id: &str)

Removes all registered topic embeddings for the given node.

pub fn stats(&self) -> &RouterStats

Returns a reference to the accumulated routing statistics.

Trait Implementations

impl Debug for SemanticContentRouter

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.