Struct CDCRResolver 

pub struct CDCRResolver { /* private fields */ }

Cross-Document Coreference Resolver.

Clusters entity mentions across multiple documents into unified clusters representing the same real-world entities.

Algorithm

1. Blocking (LSH): Generate candidate pairs from all mentions
2. Comparison: Compute similarity for candidate pairs (uses ClusterEncoder if available)
3. Clustering: Agglomerative clustering with single-link

Scalability

With LSH blocking, CDCR scales to millions of mentions:

• Without blocking: O(n²) comparisons
• With blocking: O(n × average_block_size)

Cluster Encoder Integration

When a ClusterEncoder is provided in the config, CDCR uses learned cluster embeddings for similarity scoring instead of string similarity. This enables more accurate cross-document linking based on semantic similarity rather than surface form matching.

Implementations

impl CDCRResolver

pub fn new() -> Self

Create a new resolver.

pub fn with_config(config: CDCRConfig) -> Self

Create with configuration.

pub fn with_cluster_encoder(self, encoder: Arc<dyn ClusterEncoder>) -> Self

Set cluster encoder for learned similarity scoring.

When a cluster encoder is provided, CDCR will use learned embeddings for similarity computation instead of string similarity. This enables more accurate cross-document entity linking.

pub fn resolve(&self, documents: &[Document]) -> Vec<CrossDocCluster>

Resolve cross-document coreference.

Trait Implementations

impl Clone for CDCRResolver

fn clone(&self) -> CDCRResolver

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CDCRResolver

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

Formats the value using the given formatter.

impl Default for CDCRResolver

fn default() -> CDCRResolver

Returns the “default value” for a type.