mbi.Domain

class mbi.Domain(attributes, shape, labels=None)[source]

Bases: object

Represents a discrete domain defined by attributes and their sizes.

This class encapsulates a set of named attributes and their corresponding discrete sizes (shapes). It provides methods for common domain operations.

attributes

A tuple containing the names of the attributes in the domain.

Type:: tuple[str, …]

shape

A tuple containing the integer sizes (number of discrete values) for each corresponding attribute in the attributes tuple.

Type:: tuple[int, …]

labels

An optional tuple of tuples containing semantic information (labels) for each attribute’s values. Must be the same length as attributes, and each inner tuple must have length corresponding to the attribute’s size.

Type:: tuple[tuple[Any, …], …] | None

Supported Operations:

Projection (project): Creates a new domain with a subset of attributes.
Marginalization (marginalize): Creates a new domain excluding specified attributes.
Intersection (intersect): Creates a new domain containing only common attributes.
Merging (merge): Combines two domains into a larger one.
Size Calculation (size): Computes the total number of configurations in the domain or a subset.

Example Usage (using fromdict):

>>> domain = Domain.fromdict({'a': 2, 'b': 3})
>>> print(domain)
Domain(a: 2, b: 3)

Method generated by attrs for class Domain.

Methods

`__init__`	Method generated by attrs for class Domain.
`axes`	Return the axes tuple for the given attributes.
`canonical`	Returns attributes common to the domain and input, maintaining the domain's order.
`contains`	Checks if this domain contains all attributes present in another domain.
`fromdict`	Construct a Domain object from a dictionary of { attr : size } values.
`intersect`	Intersect this Domain object with another.
`invert`	Returns attributes present in the domain but not in the provided list.
`marginalize`	Marginalize out some attributes from the domain (opposite of project).
`merge`	Merge this Domain object with another.
`project`	Project the domain onto a subset of attributes.
`size`	Return the total size of the domain.
`supports`

Attributes

`attrs`	Alias for the attributes tuple.
`config`	size } values.
`labels_config`	labels } values.
`attributes`
`shape`
`labels`

attributes: tuple[str, ...]

shape: tuple[int, ...]

labels: tuple[tuple[Any, ...], ...] | None

property config: dict[str, int]

size } values.

Type:: Returns a dictionary of { attr

property labels_config: dict[str, tuple[Any, ...]] | None

labels } values.

Type:: Returns a dictionary of { attr

static fromdict(config: dict[str, int]) → Domain[source]

Construct a Domain object from a dictionary of { attr : size } values.

Example Usage:

>>> print(Domain.fromdict({'a': 10, 'b': 20}))
Domain(a: 10, b: 20)

Parameters:: config – a dictionary of { attr : size } values
Returns:: the Domain object

project(attributes: str | Sequence[str]) → Domain[source]

Project the domain onto a subset of attributes.

Parameters:: attributes – the attributes to project onto
Returns:: the projected Domain object

marginalize(attrs: Sequence[str]) → Domain[source]

Marginalize out some attributes from the domain (opposite of project).

Example Usage:

>>> D1 = Domain(['a','b'], [10,20])
>>> print(D1.marginalize(['a']))
Domain(b: 20)

Parameters:: attrs – the attributes to marginalize out.
Returns:: the marginalized Domain object

contains(other: Domain) → bool[source]: Checks if this domain contains all attributes present in another domain.

canonical(attrs)[source]: Returns attributes common to the domain and input, maintaining the domain’s order.

invert(attrs)[source]: Returns attributes present in the domain but not in the provided list.

intersect(other: Domain) → Domain[source]

Intersect this Domain object with another.

Example Usage:

>>> D1 = Domain(['a','b'], [10,20])
>>> D2 = Domain(['b','c'], [20,30])
>>> print(D1.intersect(D2))
Domain(b: 20)

Parameters:: other – another Domain object
Returns:: the intersection of the two domains

axes(attrs: Sequence[str]) → tuple[int, ...][source]

Return the axes tuple for the given attributes.

Parameters:: attrs – the attributes
Returns:: a tuple with the corresponding axes

merge(other: Domain) → Domain[source]

Merge this Domain object with another.

Parameters:: other – another Domain object
Returns:: a new domain object covering the full domain

Example

>>> D1 = Domain(['a','b'], [10,20])
>>> D2 = Domain(['b','c'], [20,30])
>>> print(D1.merge(D2))
Domain(a: 10, b: 20, c: 30)

Parameters:: other – another Domain object
Returns:: a new domain object covering the combined domain.

size(attributes: Sequence[str] | None = None) → int[source]

Return the total size of the domain.

Example

>>> D1 = Domain(['a','b'], [10,20])
>>> D1.size()
200
>>> D1.size(['a'])
10

Parameters:: attributes – A subset of attributes whose total size should be returned.
Returns:: the total size of the domain

property attrs: Alias for the attributes tuple.

supports(attrs: str | Sequence[str]) → bool[source]