Struct DefaultAnalytics

pub struct DefaultAnalytics {
    pelt_config: PeltConfig,
}

Default analytics backend using pure-Rust implementations.

When the torch-backend feature is enabled and a model is loaded, prediction uses the Neural ODE. Otherwise falls back to linear extrapolation.

Fields

pelt_config: PeltConfig

Implementations

impl DefaultAnalytics

pub fn new() -> Self

Create with default configuration.

pub fn with_pelt_config(pelt_config: PeltConfig) -> Self

Create with custom PELT configuration.

pub fn has_neural_ode(&self) -> bool

Whether a Neural ODE model is loaded.

Trait Implementations

impl AnalyticsBackend for DefaultAnalytics

fn predict( &self, trajectory: &[TemporalPoint], target_timestamp: i64, ) -> Result<TemporalPoint, AnalyticsError>

Predict a future vector state using the learned trajectory model.

fn detect_changepoints( &self, trajectory: &[TemporalPoint], method: CpdMethod, ) -> Result<Vec<ChangePoint>, AnalyticsError>

Detect change points in a trajectory.

fn velocity( &self, trajectory: &[TemporalPoint], timestamp: i64, ) -> Result<Vec<f32>, AnalyticsError>

Compute the velocity vector at a given timestamp.

impl Default for DefaultAnalytics

fn default() -> Self

Returns the “default value” for a type.