Struct CorefChain 

pub struct CorefChain {
    pub mentions: Vec<Mention>,
    pub cluster_id: Option<CanonicalId>,
    pub entity_type: Option<String>,
}

A coreference chain: mentions that all refer to the same entity.

// "John went to the store. He bought milk."
//  ^^^^                    ^^
let john = Mention::new("John", 0, 4);
let he = Mention::new("He", 25, 27);

let chain = CorefChain::new(vec![john, he]);
assert_eq!(chain.len(), 2);
assert!(!chain.is_singleton());

Note

This type is for evaluation and intermediate processing. For production pipelines, use Track which integrates with the Signal/Track/Identity hierarchy.

Fields

mentions: Vec<Mention>

Mentions in document order (sorted by start position).

cluster_id: Option<CanonicalId>

Cluster ID from the source data, if any.

entity_type: Option<String>

Entity type shared by all mentions (e.g., “PERSON”).

Implementations

impl CorefChain

pub fn new(mentions: Vec<Mention>) -> CorefChain

Build a chain from mentions. Sorts by position automatically.

let chain = CorefChain::new(vec![
    Mention::new("she", 50, 53),
    Mention::new("Dr. Smith", 0, 9),  // out of order
]);
assert_eq!(chain.mentions[0].text, "Dr. Smith"); // sorted

pub fn with_id( mentions: Vec<Mention>, cluster_id: impl Into<CanonicalId>, ) -> CorefChain

Build a chain with an explicit cluster ID.

pub fn singleton(mention: Mention) -> CorefChain

A chain with exactly one mention (entity mentioned only once).

pub fn len(&self) -> usize

Number of mentions. A chain with 3 mentions has 2 implicit “links”.

pub fn is_empty(&self) -> bool

True if chain has no mentions. Shouldn’t happen in valid data.

pub fn is_singleton(&self) -> bool

True if chain has exactly one mention (singleton entity).

pub fn links(&self) -> Vec<(&Mention, &Mention)>

All pairwise links. For MUC: n mentions = n*(n-1)/2 links.

let chain = CorefChain::new(vec![
    Mention::new("A", 0, 1),
    Mention::new("B", 2, 3),
    Mention::new("C", 4, 5),
]);
assert_eq!(chain.links().len(), 3); // A-B, A-C, B-C

pub fn link_count(&self) -> usize

Number of coreference links.

For a chain of n mentions: n*(n-1)/2 pairs, but only n-1 links needed to connect all mentions (spanning tree).

pub fn all_pairs(&self) -> Vec<(&Mention, &Mention)>

Get all pairwise mention combinations (for B³, CEAF).

pub fn contains_span(&self, start: usize, end: usize) -> bool

Check if chain contains a mention with given span.

pub fn first(&self) -> Option<&Mention>

Get first mention (usually the most salient/representative).

pub fn mention_spans(&self) -> HashSet<(usize, usize)>

Get set of mention span IDs for set operations.

pub fn canonical_mention(&self) -> Option<&Mention>

Get the canonical (representative) mention for this chain.

Prefers proper nouns over other mention types, then longest mention. Falls back to first mention if no proper noun exists.

pub fn canonical_id(&self) -> Option<CanonicalId>

Get the canonical ID for this chain (cluster_id if set).

Trait Implementations

impl Clone for CorefChain

fn clone(&self) -> CorefChain

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CorefChain

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for CorefChain

fn deserialize<__D>( __deserializer: __D, ) -> Result<CorefChain, <__D as Deserializer<'de>>::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Display for CorefChain

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

Formats the value using the given formatter.

impl PartialEq for CorefChain

fn eq(&self, other: &CorefChain) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for CorefChain

fn serialize<__S>( &self, __serializer: __S, ) -> Result<<__S as Serializer>::Ok, <__S as Serializer>::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl Eq for CorefChain

impl StructuralPartialEq for CorefChain