neurom.features

NeuroM, lightweight and fast.

Examples

Obtain some morphometrics >>> import neurom >>> from neurom import features >>> m = neurom.load_morphology(“tests/data/swc/Neuron.swc”) >>> ap_seg_len = features.get(‘segment_lengths’, m, neurite_type=neurom.APICAL_DENDRITE) >>> ax_sec_len = features.get(‘section_lengths’, m, neurite_type=neurom.AXON)

Functions

feature(shape, namespace[, name])

Feature decorator to automatically register the feature in the appropriate namespace.

get(feature_name, obj, **kwargs)

Obtain a feature from a set of morphology objects.

Classes

NameSpace(value)

The level of morphology abstraction that feature applies to.
class neurom.features.NameSpace(value)

The level of morphology abstraction that feature applies to.

neurom.features.feature(shape, namespace: NameSpace, name=None)

Feature decorator to automatically register the feature in the appropriate namespace.

This decorator also ensures that the results of the features are casted to built-in types.

Parameters:

  • shape (tuple) – the expected shape of the feature values

  • namespace (str) – a namespace, see NameSpace

  • name (str) – name of the feature, used to access the feature via neurom.features.get().

neurom.features.get(feature_name, obj, **kwargs)

Obtain a feature from a set of morphology objects.

Features can be either Neurite, Morphology or Population features. For Neurite features see neurom.features.neurite. For Morphology features see neurom.features.morphology. For Population features see neurom.features.population.

Parameters:

  • feature_name (str) – feature to extract

  • obj – a morphology, a morphology population or a neurite tree

  • kwargs – parameters to forward to underlying worker functions

Returns:

feature value as a list or a single number.

Return type:

List|float