Crate sevensense_analysis 

sevensense-analysis

Analysis bounded context for 7sense bioacoustic analysis platform.

This crate provides clustering, motif detection, sequence analysis, and anomaly detection capabilities for bioacoustic embeddings.

Features

• Clustering: HDBSCAN and K-means clustering for grouping similar vocalizations
• Prototype Extraction: Identify representative embeddings (exemplars) for each cluster
• Motif Detection: Discover recurring patterns in vocalization sequences
• Sequence Analysis: Markov chain analysis, transition matrices, entropy computation
• Anomaly Detection: Identify unusual or novel vocalizations

Architecture

This crate follows Domain-Driven Design (DDD) with hexagonal architecture:

• domain/ - Core domain entities, value objects, and repository traits
• application/ - Application services orchestrating domain operations
• infrastructure/ - Concrete implementations (HDBSCAN, Markov chains, etc.)

Example

use sevensense_analysis::{
    application::ClusteringService,
    domain::{ClusteringConfig, ClusteringMethod},
};

let service = ClusteringService::new(ClusteringConfig::default());
let embeddings = vec![/* ... */];
let clusters = service.run_hdbscan(&embeddings).await?;

Re-exports

pub use domain::entities::Anomaly;

pub use domain::entities::AnomalyType;

pub use domain::entities::Cluster;

pub use domain::entities::ClusterId;

pub use domain::entities::EmbeddingId;

pub use domain::entities::Motif;

pub use domain::entities::MotifOccurrence;

pub use domain::entities::Prototype;

pub use domain::entities::RecordingId;

pub use domain::entities::SegmentId;

pub use domain::entities::SequenceAnalysis;

pub use domain::repository::ClusterRepository;

pub use domain::repository::MotifRepository;

pub use domain::repository::SequenceRepository;

pub use domain::events::AnalysisEvent;

pub use domain::events::ClusterAssigned;

pub use domain::events::ClustersDiscovered;

pub use domain::events::MotifDetected;

pub use domain::events::SequenceAnalyzed;

pub use domain::value_objects::ClusteringConfig;

pub use domain::value_objects::ClusteringMethod;

pub use domain::value_objects::ClusteringParameters;

pub use domain::value_objects::MotifConfig;

pub use domain::value_objects::SequenceMetrics;

pub use domain::value_objects::TransitionMatrix;

pub use application::services::AnomalyDetectionService;

pub use application::services::ClusteringService;

pub use application::services::MotifDetectionService;

pub use application::services::SequenceAnalysisService;

pub use metrics::ClusteringMetrics;

pub use metrics::SequenceEntropy;

pub use metrics::SilhouetteScore;

pub use metrics::VMeasure;

Modules

application

Application layer for the Analysis bounded context.

domain

Domain layer for the Analysis bounded context.

infrastructure

Infrastructure layer for the Analysis bounded context.

metrics

Clustering quality metrics and evaluation.

prelude

Prelude module for convenient imports

Constants

VERSION

Crate version information