mml.core.data_preparation.data_archive

class DataArchive

Bases: object

A simple dataclass holding information about an data archive (e.g. a zipfile).

__init__(path: Path, kind: DataKind = DataKind.MIXED, md5sum: str | None = None, password: str | None = None, keep_top_level: bool = False) → None

check_hash() → None

Checks if the optional md5sum of the DataArchive matches the actual files md5sum.

keep_top_level: bool = False

kind: DataKind = 'mixed_data'

md5sum: str | None = None

password: str | None = None

path: Path

class DataKind

Bases: StrEnum

Kinds of data. Used to somehow sort into distinct top level folders. If multiple kinds are mixed, MIXED should be used as default. Usage is not enforced but may help to structure any data storage.

MIXED = 'mixed_data'

TESTING_DATA = 'testing_data'

TESTING_LABELS = 'testing_labels'

TRAINING_DATA = 'training_data'

TRAINING_LABELS = 'training_labels'

UNLABELED_DATA = 'unlabeled_data'

__new__(value)