load_MAREoS#
- uniharmony.datasets.load_MAREoS(effects: list[str] | str | None = None, effect_types: list[str] | str | None = None, effect_examples: list[str] | str | None = None, as_numpy: bool = True, data_dir: Path | None = None, force_download: bool = False) dict[str, dict[str, DataFrame | ndarray]]#
Load multiple MAREoS datasets.
- Parameters:
- effectslist of str, str or None, optional (default None)
List of effects to load. If None, loads all [“eos”, “true”]
- effect_typeslist of str, str or None, optional (default None)
List of effect types to load. If None, loads all [“simple”, “interaction”]
- effect_exampleslist of str, str or None, optional (default None)
List of examples to load. If None, loads all [“1”, “2”].
- as_numpybool, optional (default True)
If True, return
numpy.ndarray, elsepandas.DataFrame.- data_dirPath | None, optional (default None)
Directory containing MAREoS data files. If None, downloads to cache.
- force_downloadbool, optional (default False)
Force to download again the dataset in case of corrupt files.
- Returns:
- dict of str and dict
Nested dictionary where keys are dataset names containing:
“X”: Feature matrix
“y”: Target labels
“sites”: Site labels
“covs”: Covariates
“folds”: Cross-validation folds
- Raises:
- ValueError
If any parameter contains invalid values.
Examples
>>> datasets = load_MAREoS() >>> len(datasets) 8 >>> datasets = load_MAREoS(effects=["eos"], effect_types=["simple"]) >>> len(datasets) 2 >>> list(datasets.keys()) ['eos_simple1', 'eos_simple2']
