AssessmentProvider#

class scikitplot.feature_selection.AssessmentProvider(name, func, defaults=(), version='1', description='', validate=None)[source]#

A named, versioned function that answers one kind of question.

Parameters:

namestr

Registry name.

funccallable

func(context, **params) -> AssessmentOutcome.

defaultstuple of (str, value) pairs

Every accepted parameter with its default; a value’s type is the parameter’s type (None accepts any scalar).

versionstr, default=”1”

Bumped when the rule’s meaning changes.

descriptionstr, default=””

One line for listings.

validatecallable or None, default=None

validate(params) raising PolicyError for out-of-range values; called once per profile, before any column is assessed.

Parameters:

• name (str)

• func (Callable[[...], AssessmentOutcome])

• defaults (tuple[tuple[str, Any], ...])

• version (str)

• description (str)

• validate (Callable[[Mapping[str, Any]], None] | None)

defaults: tuple[tuple[str, Any], ...] = ()#

description: str = ''#

func: Callable[[...], AssessmentOutcome][source]#

name: str[source]#

resolve_params(given, *, field)[source]#

Merge given over the defaults, checking names and types.

Parameters:

givenmapping

Parameters from the policy.

fieldstr

Name used in error messages.

Returns:

dict

Complete parameters.

Raises:

PolicyError

For an unknown name, a wrong type or (through validate) a bad range.

Parameters:

• given (Mapping[str, Any])

• field (str)

Return type:

dict[str, Any]

validate: Callable[[Mapping[str, Any]], None] | None = None#

version: str = '1'#