ctf4science.data_module.load_dataset#

ctf4science.data_module.load_dataset(dataset_name: str, pair_id: int, transpose=False) tuple[list, ndarray | None]#

Load train and initialization data for a given dataset and pair ID.

Training matrices are loaded from the dataset train/ directory according to the pair config; optional initialization matrix is loaded if present.

Parameters:
dataset_namestr

Name of the dataset (e.g. 'ODE_Lorenz', 'PDE_KS').

pair_idint

ID of the train-test pair.

transposebool, optional

If True, transpose each loaded matrix. Default is False.

Returns:
train_datalist of ndarray

List of training data arrays (one per train entry in the pair).

init_datandarray or None

Initialization data if the pair defines it, otherwise None.

Raises:
ValueError

If pair_id does not exist, has no training matrices, or a data file does not contain exactly one main variable.

FileNotFoundError

If any train or initialization data file is not found.