etna.datasets.load_dataset#

load_dataset(name: str, download_path: Path = PosixPath('/home/runner/.etna/internal_datasets'), rebuild_dataset: bool = False, parts: str | Tuple[str, ...] = 'full') → TSDataset | List[TSDataset][source]#

Load internal dataset. Full list of available datasets you can see on internal datasets page.

Parameters:

  • name (str) – Name of the dataset.

  • download_path (Path) – The path for saving dataset locally. By default it is directory “~/.etna/internal_datasets”.

  • rebuild_dataset (bool) – Whether to rebuild the dataset from the original source. If rebuild_dataset=False and the dataset was saved locally, then it would be loaded from disk. If rebuild_dataset=True, then the dataset will be downloaded and saved locally.

  • parts (str | Tuple[str, ...]) –

    Parts of the dataset to load. Each dataset has specific parts (e.g. ("train", "test", "full") for electricity_15T dataset). By default, all datasets have “full” part, other parts may vary.

    • If parts is str, then the function will return a single TSDataset object.

    • If parts is a tuple of multiple elements, then the function will return a list of TSDataset objects.

Returns:

internal dataset

Return type:

result

Raises:

  • NotImplementedError: – if name not from available list of dataset names

  • NotImplementedError: – if part not from available list of dataset parts