check_is_fitted

pybear.base.check_is_fitted(estimator, attributes=None, *, msg=None, all_or_any=<built-in function all>)

Perform _is_fitted validation on an estimator/transformer.

Checks if the estimator/transformer is fitted by looking for 3 things, in the presented order, via the pybear is_fitted() function.

The estimator/transformer is fitted if it:

1. has a __pybear_is_fitted__ dunder method and it returns

   boolean True

2. has any or all attributes given by attributes, if it is

   passed; if not passed, this step is skipped

3. has an attribute that ends with an underscore and does not

   start with double underscore.

If none of these things are true, the estimator/transformer is not fitted and raises a pybear.exceptions.NotFittedError with the message given by msg or the default message if msg is not passed.

Parameters:

estimatorobject

Estimator/tranformer instance for which the validation is performed.

attributesstr | Sequence[str] | None, default=None

Attribute name(s) given as string or a list/tuple of strings Eg.: ‘coef_’ or [’coef_’, ‘estimator_’, …].

msgstr, default=None

The default error message is, f”This {name} instance is not fitted yet. Call fit with appropriate arguments before using this estimator.”

For custom messages, if {name} is present in the message string, it is substituted for the estimator name.

E.g.: f”Estimator, {name}, must be fitted before sparsifying”.

all_or_anycallable, {all, any}, default=all

Specifies whether all or any of the given attributes must exist.

Raises:

ValueError

If the passed estimator/transformer is invalid or is valid but is not instantiated.

If the value passed to attributes is invalid.

If the value passed to msg is not a string or None.

If the function passed to all_or_any is not one of the built-in Python all() or any() functions.

NotFittedError

If the estimator/transformer fails all 3 checks for being fit.

Examples

>>> import numpy as np
>>> from pybear.preprocessing import InterceptManager as IM
>>> from pybear.base import check_is_fitted
>>> from pybear.base.exceptions import NotFittedError
>>> trf = IM()
>>> try:
...     check_is_fitted(trf)
... except NotFittedError as exc:
...     print(f"Model is not fitted yet.")
Model is not fitted yet.
>>> X = np.random.randint(0, 4, (2, 2))
>>> y = np.random.randint(0, 2, (2,))
>>> trf.fit(X, y)
InterceptManager()
>>> print(check_is_fitted(trf))
None