Module datasets 

Loading, discovery, caching, and subsetting for converted dataset bundles.

The bundle format is specified in the repository’s dataset_spec.md. DatasetLoader is strictly local and never performs network access or source-specific conversion. DatasetRegistry adds explicit registry access, verified downloads, and managed caching for callers that opt in.

Structs

Dataset

A validated dataset loaded from a bundle conforming to dataset_spec.md.

DatasetArchive

Immutable compressed bundle information from a registry entry.

DatasetCitation

Preferred citation metadata from a registry entry.

DatasetLicense

License metadata from a registry entry.

DatasetListing

A registry entry annotated with its current managed-cache status.

DatasetLoader

Loads an already-converted dataset bundle from a local directory.

DatasetManifest

The dataset.json entry point defined by dataset_spec.md.

DatasetRegistry

Registry client with verified downloads and a managed local cache.

DatasetRegistryDocument

Strict versioned dataset registry document.

DatasetRegistryEntry

One current immutable dataset version advertised by a registry.

DatasetSource

Original-source provenance from a registry entry.

DatasetSubset

DatasetSubsetBuilder

Rect

Enums

FrameSelector

Constants

DATASET_CACHE_DIR_ENV

Environment variable overriding the platform dataset cache directory.

DATASET_FORMAT_VERSION

Current version of the language-neutral dataset bundle format.

DATASET_REGISTRY_URL_ENV

Environment variable overriding DEFAULT_DATASET_REGISTRY_URL.

DATASET_REGISTRY_VERSION

Registry document version understood by this release.

DEFAULT_DATASET_REGISTRY_URL

Default registry source used when no API or environment override is set.

Functions

open_dataset

Opens a registry dataset using environment and platform defaults.