mvpa2.kernels.np.Parameter

class mvpa2.kernels.np.Parameter(default, constraints=None, ro=False, index=None, value=None, name=None, doc=None, **kwargs)

This class shall serve as a representation of a parameter.

It might be useful if a little more information than the pure parameter value is required (or even only useful).

Each parameter must have a value. However additional attributes can be passed to the constructor and will be stored in the object.

Notes

BIG ASSUMPTION: stored values are not mutable, ie nobody should do

cls.parameter1[:] = ...

or we wouldn’t know that it was changed Here is a list of possible additional attributes:

step

Increment/decrement step size hint for optimization

Methods

Specify a Parameter with a default value and arbitrary number of additional attributes.

Parameters:

constraints : callable

A functor that takes any input value, performs checks or type conversions and finally returns a value that is appropriate for a parameter or raises an exception.

name : str

Name of the parameter under which it should be available in its respective collection.

doc : str

Documentation about the purpose of this parameter.

index : int or None

Index of parameter among the others. Determines order of listing in help. If None, order of instantiation determines the index.

ro : bool

Either value which will be assigned in the constructor is read-only and cannot be changed

value :

Actual value of the parameter to be assigned

Examples

-ensure the parameter to be of type float (None not allowed as value): constraints = EnsureFloat() >>> from mvpa2.base.param import Parameter >>> from mvpa2.base.constraints import (EnsureFloat, EnsureRange, ... AltConstraints, Constraints) >>> C = Parameter(23.0, constraints=EnsureFloat())

-ensure the parameter to be of type float or to be None: >>> C = Parameter(23.0, constraints=AltConstraints(EnsureFloat(), None))

-ensure the parameter to be None or to be of type float and lie in the inclusive range (7.0,44.0): >>> C = Parameter(23.0, AltConstraints(Constraints(EnsureFloat(), ... EnsureRange(min=7.0,max=44.0)), ... None))

Methods

default

equal_default

Returns True if current value is equal to default one

is_default

Returns True if current value is bound to default one

reset_value()

Reset value to the default

value