visions.relations.relations.TypeRelation

class visions.relations.relations.TypeRelation(related_type, inferential, transformer, relationship=<function default_relation>, type=None)

Relationship encoder between implementations of visions.types.type.VisionsBaseType

Defines a one to one relationship between two visions.types.type.VisionsBaseType implementations, A and B, with respect to an underlying data series. In order to define a relationship we need two methods:

• is_relationship, determines whether a series of type B can be alternatively represented as type A.

• transform, provides a mechanism to convert the series from B -> A.

For example, the series pd.Series([1.0, 2.0, 3.0]) is encoded as a sequence of floats but in reality they are all integers.

Examples

>>> from visions.types import Integer, Float
>>> x = pd.Series([1.0, 2.0, 3.0])
>>> state = dict()
>>> relation = TypeRelation(Integer, Float)
>>> relation.is_relation(x, state)
True

>>> relation.transform(x, state)
pd.Series([1, 2, 3])

Method generated by attrs for class TypeRelation.

__init__(related_type, inferential, transformer, relationship=<function default_relation>, type=None)

Method generated by attrs for class TypeRelation.

Methods

__init__(related_type, inferential, transformer)	Method generated by attrs for class TypeRelation.
is_relation(series[, state])	rtype bool
transform(series[, state])	rtype ~T

Attributes

inferential
transformer
relationship