Struct DatasetManager 

pub struct DatasetManager { /* private fields */ }

High-level facade for dataset CRUD operations.

Wraps the low-level DB traits with optional ACL enforcement, matching the Python SDK’s datasets class.

Implementations

impl DatasetManager

pub fn new(db: Arc<dyn DatasetDb>) -> Self

Create a new DatasetManager without ACL enforcement.

pub fn with_acl(self, acl_db: Arc<dyn AclDb>) -> Self

Enable ACL enforcement using the given ACL database.

pub async fn list_datasets( &self, owner_id: Uuid, ) -> Result<Vec<Dataset>, DatasetError>

List all datasets accessible to the given owner.

When ACL is configured, only datasets the owner has “read” permission on are returned. Without ACL, all datasets owned by the user are listed.

pub async fn list_data( &self, dataset_id: Uuid, owner_id: Uuid, ) -> Result<Vec<Data>, DatasetError>

List all data items in a dataset, with permission check.

Results are sorted by data_size descending (largest first), matching Python SDK behaviour.

pub async fn has_data( &self, dataset_id: Uuid, owner_id: Uuid, ) -> Result<bool, DatasetError>

Check whether a dataset contains any data items.

Enforces read permission when ACL is configured, then uses an efficient COUNT query instead of loading all records.

pub async fn get_status( &self, dataset_ids: &[Uuid], ) -> Result<HashMap<Uuid, HashMap<String, PipelineRunStatus>>, DatasetError>

Get the latest pipeline status for each dataset, across all tracked pipelines.

Returns a nested map { dataset_id → { pipeline_name → status } }. Datasets with no pipeline runs for a given pipeline are omitted from the inner map (equivalent to Python’s “not started” behaviour).

pub fn discover_datasets( directory_path: &Path, ) -> Result<Vec<String>, DatasetError>

Scan a filesystem directory for dataset-like sub-directories.

Returns the names of immediate child directories. This is a sync utility matching the Python SDK’s discover_datasets method.

pub async fn empty_dataset( &self, dataset_id: Uuid, owner_id: Uuid, delete_service: &DeleteService, ) -> Result<DeleteResult, DatasetError>

Delete all data in a dataset (and the dataset record itself).

Delegates to DeleteService with DeleteScope::Dataset.

pub async fn delete_data( &self, dataset_id: Uuid, data_id: Uuid, owner_id: Uuid, mode: DeleteMode, delete_dataset_if_empty: bool, delete_service: &DeleteService, ) -> Result<DeleteResult, DatasetError>

Delete a specific data item from a dataset.

Delegates to DeleteService with DeleteScope::Data.

pub async fn delete_all( &self, owner_id: Uuid, delete_service: &DeleteService, ) -> Result<Vec<DeleteResult>, DatasetError>

Delete all datasets for an owner.

Lists all accessible datasets and delegates each to DeleteService.

pub async fn create_dataset( &self, name: &str, owner_id: Uuid, tenant_id: Option<Uuid>, ) -> Result<Dataset, DatasetError>

Create a dataset with a deterministic ID matching Python’s formula: uuid5(NAMESPACE_OID, f"{name}{user_id}{tenant_id}").

Idempotent: if a dataset with the same deterministic ID already exists, returns the existing row.

pub async fn create_authorized_dataset( &self, name: &str, owner_id: Uuid, tenant_id: Option<Uuid>, parent_user_id: Option<Uuid>, ) -> Result<Dataset, DatasetError>

Create a dataset and grant all four ACL permissions (read, write, delete, share) to the owner.