pydims.DimensionedArray#

class pydims.DimensionedArray(*, dims, values, unit)[source]#

Array with named dimensions and optional unit.

__init__(*, dims, values, unit)[source]#

Parameters:

dims (tuple[Hashable, ...]) – Dimension names.
values (ArrayImplementation) – Array of values.
unit (UnitImplementation | None) – Optional unit.

Methods

`__init__`(*, dims, values, unit)	type dims: `tuple`[`Hashable`, `...`]
`astype`(dtype[, copy])	rtype: `TypeVar`(`DimArr`, bound= DimensionedArray)
`expand_dims`(sizes)	Expand an array by adding new dimensions of the given sizes at the beginning.
`flatten`(*[, dims, dim])	Flatten a set of dimensions into a single dimension.
`fold`(dim, *, sizes)	Fold a dimension of an array into a new set of dimensions.
`permute_dims`(dims)	Permute the dimensions of an array.
`reshape`(**kwargs)	Not supported since it relies on a particular axis order.
`squeeze`([dim])	Remove dimensions of size 1.
`to`(*[, dtype, unit, copy])	Convert to a new dtype and/or unit.

Attributes

`array_namespace`
`dim`
`dims`
`dtype`
`ndim`
`shape`
`size`
`sizes`
`unit`
`units_namespace`
`values`

__getitem__(key)[source]#

Get a sub-array identified by key.

Parameters:: key (int | slice | dict[Hashable, int | slice] | EllipsisType) – Index or slice or dictionary of dimension names and indices.
Returns:: TypeVar(DimArr, bound= DimensionedArray) – Sub-array.

__setitem__(key, array)[source]#

Set a sub-array identified by key to the values of array.

The array to set will be automatically transposed and/or broadcast to match the dimensions of the sub-array, using the named dimensions.

Parameters:

key (int | slice | dict[Hashable, int | slice] | EllipsisType) – Index or slice or dictionary of dimension names and indices.
array (TypeVar(DimArr, bound= DimensionedArray)) – Array to set.

expand_dims(sizes)#

Expand an array by adding new dimensions of the given sizes at the beginning.

Note that in contrast to the underlying expand_dims function, this function does not add dimensions of size 1, since this library implements broadcasting differently.

Use permute_dims() to control the order of the old and new dimensions.

Parameters:

array (TypeVar(DimArr, bound= DimensionedArray)) – Array to expand.
sizes (dict[Hashable, int]) – Ordered names and sizes of the new dimensions.

Returns:

TypeVar(DimArr, bound= DimensionedArray) – Expanded array.

flatten(*, dims=None, dim=None)#

Flatten a set of dimensions into a single dimension.

Parameters:

array (TypeVar(DimArr, bound= DimensionedArray)) – Array to flatten.
dims (tuple[Hashable, ...] | None, default: None) – Dimensions to flatten.
dim (Hashable | None, default: None) – Name of the new dimension.

Returns:

TypeVar(DimArr, bound= DimensionedArray) – Flattened array.

fold(dim, *, sizes)#

Fold a dimension of an array into a new set of dimensions.

Parameters:

array (TypeVar(DimArr, bound= DimensionedArray)) – Array to fold.
dim (Hashable) – Dimension to fold.
sizes (Mapping[Hashable, int]) – Sizes of the dimensions after folding.

Returns:

TypeVar(DimArr, bound= DimensionedArray) – Folded array.

permute_dims(dims)#

Permute the dimensions of an array.

Parameters:

array (TypeVar(DimArr, bound= DimensionedArray)) – Array to permute.
dims (tuple[Hashable, ...]) – New order of dimensions.

Returns:

TypeVar(DimArr, bound= DimensionedArray) – Permuted array.

reshape(**kwargs)#

Not supported since it relies on a particular axis order.

This function of the Python Array API is not supported by PyDims because it relies on a particular axis order. This library does not support this since it can lead to errors that are hard to debug.

Raises:: NotImplementedError – Always.
Return type:: NoReturn

squeeze(dim=None)#

Remove dimensions of size 1.

Parameters:

array (TypeVar(DimArr, bound= DimensionedArray)) – Array to squeeze.
dim (Hashable | tuple[Hashable, ...] | None, default: None) – Dimensions to remove. If None, remove all dimensions of size 1.

Returns:

TypeVar(DimArr, bound= DimensionedArray) – Squeezed array.

to(*, dtype=None, unit=None, copy=True)[source]#

Convert to a new dtype and/or unit.

Parameters:

dtype (Any | None, default: None) – New dtype, None if no conversion is needed.
unit (Any | None, default: None) – New unit, None if no conversion is needed.
copy (bool, default: True) – If True, a copy of the values is made, even if no conversion is needed.

Returns:

TypeVar(DimArr, bound= DimensionedArray) – New array with the requested dtype and/or unit.