azure_ml_sdk.services.asset.dataset

Classes

DatasetManager

Class for handling datasets in Azure Machine Learning workspace.

Module Contents

class azure_ml_sdk.services.asset.dataset.DatasetManager(ml_client: azure.ai.ml.MLClient)

Bases: altametris.azure_ml_sdk.services.asset.base.BaseAssetManager

Class for handling datasets in Azure Machine Learning workspace.

register(datastore_name: str, data_path: str, dataset_name: str, dataset_version: str | None = None, dataset_description: str = '') → azure.ai.ml.entities.Data | None

Register a dataset using a data folder in a datastore.

Several cases can happen according to the input parameters of the dataset. If the data path is valid, and the version is valid, then the dataset will be registered. If the data path is invalid (not existing in the datastore), then a dataset will be registered anyway. If the same version is used, then an error will be raised. If the version is set to None, then the dataset will be registered with an incremented version.

Parameters:

  • datastore_name (str) – The name of the datastore that points to the data.

  • data_path (str) – The path of the data folder in the datastore.

  • dataset_name (str) – The name of the dataset to be registered.

  • dataset_version (str) – The version of the dataset to be registered. Defaults to None

  • dataset_description (str) – The description of the dataset to be registered. Defaults to empty string.