mvpa2.algorithms.searchlight_hyperalignment.Sphere

class mvpa2.algorithms.searchlight_hyperalignment.Sphere(radius, element_sizes=None, distance_func=None)

N-Dimensional hypersphere.

Use this if you want to obtain all the neighbors within a given radius from a point in a space with arbitrary number of dimensions assuming that the space is discrete.

No validation of producing coordinates within any extent is done.

Examples

Create a Sphere of diameter 1 and obtain all coordinates within range for the coordinate (1,1,1).

>>> s = Sphere(1)
>>> s((2, 1))
[(1, 1), (2, 0), (2, 1), (2, 2), (3, 1)]
>>> s((1, ))
[(0,), (1,), (2,)]

If elements in discrete space have different sizes across dimensions, it might be preferable to specify element_sizes parameter.

>>> s = Sphere(2, element_sizes=(1.5, 2.5))
>>> s((2, 1))
[(1, 1), (2, 1), (3, 1)]

>>> s = Sphere(1, element_sizes=(1.5, 0.4))
>>> s((2, 1))
[(2, -1), (2, 0), (2, 1), (2, 2), (2, 3)]

Attributes

distance_func
element_sizes
radius

Methods

__call__(coordinate)	Get all coordinates within diameter
train(dataset)

Initialize the Sphere

Parameters:

radius : float

Radius of the ‘sphere’. If no element_sizes provided – radius would be effectively in number of voxels (if operating on MRI data).

element_sizes : None or iterable of floats

Sizes of elements in each dimension. If None, it is equivalent to 1s in all dimensions.

distance_func : None or lambda

Distance function to use (choose one from mvpa2.clfs.distance). If None, cartesian_distance to be used.

Attributes

distance_func
element_sizes
radius

Methods

__call__(coordinate)	Get all coordinates within diameter
train(dataset)

distance_func

element_sizes

radius

train(dataset)