Struct SqliteStorage 

pub struct SqliteStorage { /* private fields */ }

SQLite-backed storage: trials are persisted to a file so studies survive process restarts and can be resumed. Mirrors Optuna’s RDB storage pattern.

Each trial is stored as a JSON document keyed by (study_name, number), which keeps the schema stable while faithfully round-tripping the define-by-run parameter set and every intermediate report.

The connection is guarded by a Mutex, so the backend is Send + Sync and usable under parallel execution. Several independent processes can also open the same file (SQLite handles the file-level locking) for lightweight multi-process coordination — a practical middle ground short of full distributed execution.

Implementations

impl SqliteStorage

pub fn open(path: impl AsRef<Path>) -> Result<Self, StorageError>

Open (creating if absent) a study database at path.

pub fn open_in_memory() -> Result<Self, StorageError>

Open an anonymous in-memory SQLite database (useful for tests).

Trait Implementations

impl Storage for SqliteStorage

fn save_trial( &self, study_name: &str, trial: &Trial, ) -> Result<(), StorageError>

Insert or update a trial for the given study.

fn load_trials(&self, study_name: &str) -> Result<Vec<Trial>, StorageError>

Load all trials for a study, in trial-number order. Returns an empty vec for a study that exists but has no trials yet.

fn save_study_metadata(&self, meta: &StudyMetadata) -> Result<(), StorageError>

Insert or update study-level metadata.

fn load_study_metadata( &self, study_name: &str, ) -> Result<Option<StudyMetadata>, StorageError>

Fetch study-level metadata, or None if the study is unknown.