check_n_features

pybear.base.check_n_features(X, n_features_in_, reset)

Set the n_features_in_ attribute, or check against it.

pybear recommends calling reset=True in fit and in the first call to partial_fit. All other methods that validate X should set reset=False.

Parameters:

XXContainer of shape (n_samples, n_features) or (n_samples,)

The input data, with a ‘shape’ attribute.

n_features_in_int | None

The number of features in the data. If this attribute exists, it is an integer. If it does not exist, it is None.

resetbool

If True:

The n_features_in_ attribute is set to X.shape[1].

If False:

If n_features_in_ exists check it is equal to X.shape[1].

If n_features_in_ does not exist the check is skipped.

Returns:

n_featuresint

The number of features in X.

Raises:

ValueError:

If X has no columns (is empty along the column axis).

If reset=False and the number of features in X does not equal

n_features_in_.

Notes

Type Aliases

NumpyTypes:

numpy.ndarray

PandasTypes:

pandas.Series | pandas.DataFrame

PolarsTypes:

polars.Series | polars.DataFrame

ScipySparseTypes:

ss.csc_matrix | ss.csc_array | ss.csr_matrix | ss.csr_array | ss.coo_matrix | ss.coo_array | ss.dia_matrix | ss.dia_array | ss.lil_matrix | ss.lil_array | ss.dok_matrix | ss.dok_array | ss.bsr_matrix | ss.bsr_array

XContainer:

NumpyTypes | PandasTypes | PolarsTypes | ScipySparseTypes

Examples

>>> from pybear.base import check_n_features
>>> import numpy as np
>>> X = np.random.randint(0, 10, (8, 5))
>>> check_n_features(X, n_features_in_=None, reset=True)
5
>>> try:
...     check_n_features(X, n_features_in_=4, reset=False)
... except Exception as e:
...     print(repr(e))
ValueError('X has 5 feature(s), but expected 4.')