Struct Track 

pub struct Track {
    pub id: TrackId,
    pub signals: Vec<SignalRef>,
    pub entity_type: Option<TypeLabel>,
    pub canonical_surface: String,
    pub identity_id: Option<IdentityId>,
    pub cluster_confidence: f32,
    pub embedding: Option<Vec<f32>>,
}

A track: a cluster of signals referring to the same entity within a document.

Terminology Mapping

Vision	NLP
Tracklet	CorefChain
Object track	Entity cluster
Re-identification	Coreference resolution

Design Philosophy

Tracks are the bridge between raw signals and global identities. They answer: “which signals in THIS document refer to the same entity?”

Key properties:

• Document-scoped: A track only exists within one document
• Homogeneous type: All signals in a track should have compatible types
• Representative: The track has a “canonical” signal (usually the first proper mention)

Fields

id: TrackId

Unique identifier within the document

signals: Vec<SignalRef>

Signal references in this track (document order)

entity_type: Option<TypeLabel>

Entity type (consensus from signals).

This is a TypeLabel to support both core taxonomy types and domain-specific labels.

canonical_surface: String

Canonical surface form (the “best” name for this entity)

identity_id: Option<IdentityId>

Link to global identity (Level 3), if resolved

cluster_confidence: f32

Confidence that signals are correctly clustered

embedding: Option<Vec<f32>>

Optional embedding for track-level representation (aggregated from signal embeddings)

Implementations

impl Track

pub fn new( id: impl Into<TrackId>, canonical_surface: impl Into<String>, ) -> Track

Create a new track.

pub fn add_signal(&mut self, signal_id: impl Into<SignalId>, position: u32)

Add a signal to this track.

pub fn len(&self) -> usize

Get the number of mentions in this track.

pub fn is_empty(&self) -> bool

Check if this track is empty.

pub fn is_singleton(&self) -> bool

Check if this is a singleton (single mention).

pub const fn id(&self) -> TrackId

Get the track’s unique identifier.

pub fn signals(&self) -> &[SignalRef]

Get the signal references in this track.

pub fn canonical_surface(&self) -> &str

Get the canonical surface form.

pub const fn identity_id(&self) -> Option<IdentityId>

Get the linked identity ID, if any.

pub const fn cluster_confidence(&self) -> f32

Get the cluster confidence score.

pub fn set_cluster_confidence(&mut self, confidence: f32)

Set the cluster confidence score.

pub fn set_identity_id(&mut self, identity_id: IdentityId)

Link this track to a global identity (mutable setter).

pub fn clear_identity_id(&mut self)

Unlink this track from its identity.

pub fn with_identity(self, identity_id: IdentityId) -> Track

Link this track to a global identity.

pub fn with_type(self, entity_type: impl Into<String>) -> Track

Set the entity type from a string.

For new code, prefer Self::with_type_label which provides type safety.

pub fn with_type_label(self, label: TypeLabel) -> Track

Set the entity type using a type-safe label.

This is the preferred method for new code as it provides type safety and integrates with the core EntityType taxonomy.

Example

use anno_core::core::grounded::Track;
use anno_core::core::types::TypeLabel;
use anno_core::EntityType;

let track = Track::new(0, "Marie Curie")
    .with_type_label(TypeLabel::Core(EntityType::Person));

pub fn type_label(&self) -> Option<TypeLabel>

Get the entity type as a type-safe label.

This converts the internal string representation to a TypeLabel, attempting to parse it as a core EntityType first.

pub fn with_embedding(self, embedding: Vec<f32>) -> Track

Set the embedding for this track.

pub fn compute_spread(&self, doc: &GroundedDocument) -> Option<usize>

Get the spread (distance from first to last mention).

Requires document to resolve signal positions.

pub fn collect_variations(&self, doc: &GroundedDocument) -> Vec<String>

Collect all surface form variations from signals.

Requires document to resolve signal surfaces.

pub fn confidence_stats( &self, doc: &GroundedDocument, ) -> Option<(f32, f32, f32)>

Get confidence statistics across all signals.

Returns (min, max, mean) confidence values.

pub fn compute_stats( &self, doc: &GroundedDocument, text_len: usize, ) -> TrackStats

Compute aggregate statistics for this track.

Returns a TrackStats struct with comprehensive aggregate features.

Trait Implementations

impl Clone for Track

fn clone(&self) -> Track

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Track

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for Track

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for Track

fn eq(&self, other: &Track) -> 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 Track

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 StructuralPartialEq for Track