DimensionUniverse

class lsst.daf.butler.DimensionUniverse

Bases: object

Self-consistent set of dimensions.

A parent class that represents a complete, self-consistent set of dimensions and their relationships.

DimensionUniverse is not a class-level singleton, but all instances are tracked in a singleton map keyed by the version number and namespace in the configuration they were loaded from. Because these universes are solely responsible for constructing DimensionElement instances, these are also indirectly tracked by that singleton as well.

Parameters:

config : Config, optional

Configuration object from which dimension definitions can be extracted. Ignored if builder is provided, or if version is provided and an instance with that version already exists.

version : int, optional

Integer version for this DimensionUniverse. If not provided, a version will be obtained from builder or config.

namespace : str, optional

Namespace of this DimensionUniverse, combined with the version to provide universe safety for registries that use different dimension definitions.

builder : DimensionConstructionBuilder, optional

Builder object used to initialize a new instance. Ignored if version is provided and an instance with that version already exists. Should not have had finish called; this will be called if needed by DimensionUniverse.

Attributes Summary

namespace	The namespace associated with this universe.
skypix	All skypix systems known to this universe.
version	The version number of this universe.

Methods Summary

expandDimensionNameSet(names)	Expand a set of dimension names in-place.
extract(iterable, str]])	Construct graph from iterable.
get(name, default, None] = None)	Return the DimensionElement with the given name or a default.
getDatabaseElements()	Return set of all DatabaseDimensionElement instances in universe.
getDimensionIndex(name)	Return the position of the named dimension.
getElementIndex(name)	Return the position of the named dimension element.
getEncodeLength()	Return encoded size of graph.
getGovernorDimensions()	Return a set of all GovernorDimension instances in this universe.
getStaticDimensions()	Return a set of all static dimensions in this universe.
getStaticElements()	Return a set of all static elements in this universe.
isCompatibleWith(other)	Check compatibility between this DimensionUniverse and another
makePacker(name, dataId)	Make a dimension packer.
sorted(elements, str]], *, reverse)	Return a sorted version of the given iterable of dimension elements.

Attributes Documentation

namespace

The namespace associated with this universe.

Returns:

namespace : str

The namespace. When combined with the version can uniquely define this universe.

skypix

All skypix systems known to this universe.

(NamedValueAbstractSet [ SkyPixSystem ]).

version

The version number of this universe.

Returns:

version : int

An integer representing the version number of this universe. Uniquely defined when combined with the namespace.

Methods Documentation

expandDimensionNameSet(names: Set[str]) → None

Expand a set of dimension names in-place.

Includes recursive dependencies.

This is an advanced interface for cases where constructing a DimensionGraph (which also expands required dependencies) is impossible or undesirable.

Parameters:

names : set [ str ]

A true set of dimension names, to be expanded in-place.

extract(iterable: Iterable[Union[lsst.daf.butler.core.dimensions._elements.Dimension, str]]) → lsst.daf.butler.core.dimensions._graph.DimensionGraph

Construct graph from iterable.

Constructs a DimensionGraph from a possibly-heterogenous iterable of Dimension instances and string names thereof.

Constructing DimensionGraph directly from names or dimension instances is slightly more efficient when it is known in advance that the iterable is not heterogenous.

Parameters:

iterable: iterable of `Dimension` or `str`

Dimensions that must be included in the returned graph (their dependencies will be as well).

Returns:

graph : DimensionGraph

A DimensionGraph instance containing all given dimensions.

get(name: str, default: Optional[lsst.daf.butler.core.dimensions._elements.DimensionElement, None] = None) → Optional[lsst.daf.butler.core.dimensions._elements.DimensionElement, None]

Return the DimensionElement with the given name or a default.

Parameters:

name : str

Name of the element.

default : DimensionElement, optional

Element to return if the named one does not exist. Defaults to None.

Returns:

element : DimensionElement

The named element.

getDatabaseElements() → lsst.daf.butler.core.named.NamedValueAbstractSet[lsst.daf.butler.core.dimensions._database.DatabaseDimensionElement]

Return set of all DatabaseDimensionElement instances in universe.

This does not include GovernorDimension instances, which are backed by the database but do not inherit from DatabaseDimensionElement.

Returns:

elements : NamedValueAbstractSet [ DatabaseDimensionElement ]

A frozen set of DatabaseDimensionElement instances.

getDimensionIndex(name: str) → int

Return the position of the named dimension.

This position is in this universe’s sorting of all dimensions.

Parameters:

name : str

Name of the dimension.

Returns:

index : int

Sorting index for this dimension.

Notes

The dimension sort order for a universe is consistent with the element order (all dimensions are elements), and either can be used to sort dimensions if used consistently. But there are also some contexts in which contiguous dimension-only indices are necessary or at least desirable.

getElementIndex(name: str) → int

Return the position of the named dimension element.

The position is in this universe’s sorting of all elements.

Parameters:

name : str

Name of the element.

Returns:

index : int

Sorting index for this element.

getEncodeLength() → int

Return encoded size of graph.

Returns the size (in bytes) of the encoded size of DimensionGraph instances in this universe.

See DimensionGraph.encode and DimensionGraph.decode for more information.

getGovernorDimensions() → lsst.daf.butler.core.named.NamedValueAbstractSet[lsst.daf.butler.core.dimensions._governor.GovernorDimension]

Return a set of all GovernorDimension instances in this universe.

Returns:

governors : NamedValueAbstractSet [ GovernorDimension ]

A frozen set of GovernorDimension instances.

getStaticDimensions() → lsst.daf.butler.core.named.NamedValueAbstractSet[lsst.daf.butler.core.dimensions._elements.Dimension]

Return a set of all static dimensions in this universe.

Non-static dimensions that are created as needed may also exist, but these are guaranteed to have no direct relationships to other elements (though they may have spatial or temporal relationships).

Returns:

dimensions : NamedValueAbstractSet [ Dimension ]

A frozen set of Dimension instances.

getStaticElements() → lsst.daf.butler.core.named.NamedValueAbstractSet[lsst.daf.butler.core.dimensions._elements.DimensionElement]

Return a set of all static elements in this universe.

Non-static elements that are created as needed may also exist, but these are guaranteed to have no direct relationships to other elements (though they may have spatial or temporal relationships).

Returns:

elements : NamedValueAbstractSet [ DimensionElement ]

A frozen set of DimensionElement instances.

isCompatibleWith(other: lsst.daf.butler.core.dimensions._universe.DimensionUniverse) → bool

Check compatibility between this DimensionUniverse and another

Parameters:

other : DimensionUniverse

The other DimensionUniverse to check for compatibility

Returns:

results : bool

If the other DimensionUniverse is compatible with this one return True, else False

makePacker(name: str, dataId: DataCoordinate) → DimensionPacker

Make a dimension packer.

Constructs a DimensionPacker that can pack data ID dictionaries into unique integers.

Parameters:

name : str

Name of the packer, matching a key in the “packers” section of the dimension configuration.

dataId : DataCoordinate

Fully-expanded data ID that identifies the at least the “fixed” dimensions of the packer (i.e. those that are assumed/given, setting the space over which packed integer IDs are unique). dataId.hasRecords() must return True.

sorted(elements: Iterable[Union[E, str]], *, reverse: bool = False) → List[E]

Return a sorted version of the given iterable of dimension elements.

The universe’s sort order is topological (an element’s dependencies precede it), with an unspecified (but deterministic) approach to breaking ties.

Parameters:

elements : iterable of DimensionElement.

Elements to be sorted.

reverse : bool, optional

If True, sort in the opposite order.

Returns:

sorted : list of DimensionElement

A sorted list containing the same elements that were given.