Struct Study 

pub struct Study { /* private fields */ }

A single optimization run: a direction, a sampler, a pruner, and a storage backend, tied to a named study whose trials live in that storage.

Construct one with Study::new, or use the ergonomic builder in the hyperopt facade crate. Run trials with Study::optimize (sequential) or [Study::optimize_parallel] (feature parallel).

Implementations

impl Study

pub fn new( name: impl Into<String>, direction: Direction, sampler: Box<dyn Sampler>, pruner: Box<dyn Pruner>, storage: Box<dyn Storage>, ) -> Result<Study, HyperoptError>

Assemble a study from its parts. Persists study metadata immediately so the direction survives a reload. If the study already exists in storage, its recorded direction is authoritative and is adopted here (so a resumed study can’t silently flip direction).

pub fn name(&self) -> &str

The study’s name (its key in storage).

pub fn direction(&self) -> Direction

The optimization direction.

pub fn trials(&self) -> Result<Vec<Trial>, HyperoptError>

A fresh snapshot of every trial recorded for this study.

pub fn best_trial(&self) -> Result<Option<Trial>, HyperoptError>

The best completed trial under the study direction, if any.

pub fn best_value(&self) -> Result<Option<f64>, HyperoptError>

The best objective value seen so far, if any trial has completed.

pub fn optimize<F>( &self, objective: F, n_trials: usize, ) -> Result<(), HyperoptError>

where F: FnMut(&mut TrialContext<'_>) -> Result<f64, ObjectiveError>,

Run n_trials sequentially. A trial that panics or returns ObjectiveError::Failed is marked Failed and the run continues; one that returns ObjectiveError::Pruned is marked Pruned.