neuron_morphology.feature_extractor.marked_feature module

class neuron_morphology.feature_extractor.marked_feature.MarkedFeature(marks: Set[Type[neuron_morphology.feature_extractor.mark.Mark]], feature: Union[Callable[[neuron_morphology.feature_extractor.data.Data], Any], neuron_morphology.feature_extractor.marked_feature.MarkedFeature], name: Optional[str] = None, preserve_marks: bool = True)

Bases: object

Attributes:

feature

marks

name

Methods

__call__(self, \*args, \*\*kwargs)	Execute the underlying feature, passing along all arguments
add_mark(self, mark)	Assign an additional mark to this feature
deepcopy(self, \*\*kwargs)	Make a deep copy of this marked feature
ensure(feature)	If a function is not a MarkedFeature, convert it.
partial(self, \*args, \*\*kwargs)	Fix one or more parameters on this feature’s callable
specialize(self, option)	Apply a specialization option to this feature.

add_mark(self, mark: Type[neuron_morphology.feature_extractor.mark.Mark])

Assign an additional mark to this feature

deepcopy(self, **kwargs)

Make a deep copy of this marked feature

classmethod ensure(feature: 'Feature') → ~M

If a function is not a MarkedFeature, convert it.

Parameters:

feature : the feature to be converted

Returns:

Either a marked feature generated from the input, or the input

marked feature.

feature

marks

name

partial(self, *args, **kwargs)

Fix one or more parameters on this feature’s callable

specialize(self, option: Type[neuron_morphology.feature_extractor.feature_specialization.FeatureSpecialization])

Apply a specialization option to this feature. This binds parameters on the feature’s __call__ method, sets 0 or more additional marks, and namespaces the feature’s name.

Parameters:

option : The specialization option with which to specialize this

feature.

Returns:

a deep copy of this feature with updated callable, marks and name

neuron_morphology.feature_extractor.marked_feature.marked(mark: Type[neuron_morphology.feature_extractor.mark.Mark])

Decorator for adding a mark to a function.

Parameters:

mark : the mark to be applied

Examples

@marked(RequiresA) @marked(RequiresB) def some_feature_requiring_a_and_b(…):

…

neuron_morphology.feature_extractor.marked_feature.nested_specialize(feature: Union[Callable[[neuron_morphology.feature_extractor.data.Data], Any], neuron_morphology.feature_extractor.marked_feature.MarkedFeature], specialization_sets: List[Set[Type[neuron_morphology.feature_extractor.feature_specialization.FeatureSpecialization]]]) → Dict[str, neuron_morphology.feature_extractor.marked_feature.MarkedFeature]

Apply specializations hierarchically to a base feature. Generating a new collection of specialized features.

Parameters:

feature : will be used as a basis for specialization

specialization_sets : each element describes a set of specialization

options. The output will have one specialization for each element of the cartesian product of these sets.

Returns:

A dictionary mapping namespaced feature names to specialized features.

Notes

Specializations are applied from the start of the specialization_sets to the end. This means that the generated names are structures like:

“last_spec.middle_spec.first_spec.base_feature_name”

neuron_morphology.feature_extractor.marked_feature.specialize(feature: Union[Callable[[neuron_morphology.feature_extractor.data.Data], Any], neuron_morphology.feature_extractor.marked_feature.MarkedFeature], specialization_set: Set[Type[neuron_morphology.feature_extractor.feature_specialization.FeatureSpecialization]]) → Dict[str, neuron_morphology.feature_extractor.marked_feature.MarkedFeature]

Bind some of a feature’s keyword arguments, using provided specialization options.

Parameters:

feature : will be used as a basis for specialization

specialization_set : each element defines a particular specialization (i.e

a set of keyword argument values and marks) to be applied to the feature

Returns:

A dictionary mapping (namespaced) feature names to specialized features.

Note that names are formatted as “specialization_name.base_feature_name”