Struct SemanticCluster 

pub struct SemanticCluster {
    pub cluster_id: usize,
    pub centroid: Vec<f32>,
    pub member_count: u64,
    pub total_drift: f32,
}

A single semantic cluster with a centroid maintained via incremental updates.

Fields

cluster_id: usize

Cluster index (0-based).

centroid: Vec<f32>

Current centroid vector (mean embedding direction).

member_count: u64

Number of documents that have been assigned to and updated this cluster.

total_drift: f32

Cumulative Euclidean shift of the centroid across all updates.

Implementations

impl SemanticCluster

pub fn euclidean_distance(&self, embedding: &[f32]) -> f32

Compute Euclidean distance between the centroid and embedding.

Returns 0.0 if either vector is empty or if the dimensions differ.

pub fn update_centroid(&mut self, new_embedding: &[f32], learning_rate: f32)

Update the centroid toward new_embedding using an exponential moving average.

Formula: centroid[i] = (1 - learning_rate) * centroid[i] + learning_rate * new_embedding[i]

Computes the drift (Euclidean shift of the centroid) and adds it to total_drift. Increments member_count by one.

This is a no-op if the dimensions of new_embedding do not match the centroid.

Trait Implementations

impl Clone for SemanticCluster

fn clone(&self) -> SemanticCluster

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for SemanticCluster

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

Formats the value using the given formatter.