named-arrays

named-arrays is an implementation of a named tensor, which assigns a name to each axis of an n-dimensional array such as a numpy array.

With a bare numpy array, the meaning of each axis lives in the programmer's head, and combining two arrays usually means inserting singleton dimensions until their shapes line up. Naming the axes removes both problems: arrays broadcast against each other by matching names, so a singleton dimension is never needed, and an operation such as a mean along the wavelength axis says exactly that.

named-arrays provides a very unapologetic implementation of a named tensor, since axes can only be accessed using their names, unlike xarray which allows for both name and index. Support for astropy.units is built in, so the values inside an array can carry a physical unit.

Installation

named-arrays is available on PyPI and can be installed using pip

pip install named-arrays

Features

The array types form a hierarchy, from a plain named tensor up to a discrete function of several variables.

ScalarArray, a named tensor with Astropy Quantity support. Analogue of xarray.Variable. Implicit variants such as ScalarLinearSpace describe an array without materializing it.
UncertainScalarArray, which carries a distribution alongside the nominal value and propagates uncertainty through every operation.
Cartesian2dVectorArray and Cartesian3dVectorArray, along with named-component variants such as SpectralPositionalVectorArray, where each component is itself any of the array types above.
Cartesian2dMatrixArray and its relatives, which are vectors of vectors supporting the usual matrix operations.
FunctionArray, a discrete function pairing inputs with outputs. Analogue of an xarray.DataArray.

Several modules extend these types to other libraries: na.plt for matplotlib, na.random and na.stats for sampling and statistics, na.regridding for resampling curvilinear grids, na.optimize for root finding and minimization, and na.transformations for rotations and translations.

Key concepts

The shape is a dictionary. shape maps each axis name to its length, and there is no positional equivalent. Anywhere the numpy API takes an axis=0, this library takes an axis="detector_x".

Arrays broadcast by matching names. Two arrays combine along the axes whose names they share, and the axes unique to either one are added to the result. An array of shape {"x": 3} plus an array of shape {"y": 2} therefore has shape {"x": 3, "y": 2}, with no reshaping and no singleton dimensions. Adding a new dimension to a calculation is a matter of giving an input an extra named axis.

Arrays are explicit or implicit. An explicit array such as ScalarArray stores its values. An implicit array such as ScalarLinearSpace stores the arguments that define it, so start, stop, and num remain available long after the array is created. Implicit arrays work in every operation an explicit array does, and .explicit materializes one on demand.

Most of the numpy API already works. These arrays implement the __array_function__ and __array_ufunc__ protocols, so np.mean, np.sqrt, and most of their siblings accept them directly, using axis names. Operations that numpy cannot express are defined in the named_arrays namespace instead.

Documentation

The full documentation, including the API reference, tutorials, and executable versions of the examples below, is hosted at named-arrays.readthedocs.io.

Examples

Broadcasting by name

The fundamental type is the ScalarArray, a composition of a numpy ndarray-like object and a tuple of axis names, which must have the same length as the number of dimensions in the array.

import numpy as np
import named_arrays as na

a = na.ScalarArray(np.array([1, 2, 3]), axes=("x",))
b = na.ScalarArray(np.array([4, 5]), axes=("y",))

Since the two arrays have different axis names, adding them together broadcasts them against each other automatically.

c = a + b

ScalarArray(
    ndarray=[[5, 6],
             [6, 7],
             [7, 8]],
    axes=('x', 'y'),
)

The result is two-dimensional, and its shape is a dictionary.

c.shape

{'x': 3, 'y': 2}

All the usual numpy reduction operations take the name of the axis to remove.

c.mean("x")

ScalarArray(
    ndarray=[6., 7.],
    axes=('y',),
)

To index the array, use a dictionary with the axis names as the keys, so the meaning of an index does not depend on the order of the axes.

c[dict(x=0)]

ScalarArray(
    ndarray=[5, 6],
    axes=('y',),
)

Implicit arrays

We recommend that you rarely create instances of ScalarArray directly. Instead, use the implicit array classes ScalarLinearSpace, ScalarLogarithmicSpace, and ScalarGeometricSpace, which mirror numpy.linspace(), numpy.logspace(), and numpy.geomspace(), with the advantage of remembering the arguments used to define them.

d = na.ScalarLinearSpace(0, 1, axis="z", num=4)

ScalarLinearSpace(start=0, stop=1, axis='z', num=4, endpoint=True, centers=False)

These implicit classes work just like a ScalarArray in any operation, and .explicit materializes one on demand.

a + d

ScalarArray(
    ndarray=[[1.        , 1.33333333, 1.66666667, 2.        ],
             [2.        , 2.33333333, 2.66666667, 3.        ],
             [3.        , 3.33333333, 3.66666667, 4.        ]],
    axes=('x', 'z'),
)

One extra axis, one plotting call

An extra named axis costs nothing, so a family of curves is a single array, and one plotting call draws all of them.

import astropy.units as u
import matplotlib.pyplot as plt

# Define the independent variable
x = na.linspace(0, 2 * np.pi, axis="x", num=101) * u.rad

# Add an axis representing three different amplitudes
amplitude = na.ScalarArray(np.array([1, 2, 3]), axes=("amplitude",))

# The result has both axes, without any reshaping
y = amplitude * np.sin(x)

fig, ax = plt.subplots(constrained_layout=True);
na.plt.plot(x, y, axis="x", ax=ax);
ax.set_xlabel(f"angle ({x.unit:latex_inline})");
ax.set_ylabel("amplitude");

plot

Uncertainty propagation

An UncertainScalarArray carries a distribution alongside its nominal value, and every operation propagates it, so the error bar at the end of a calculation needs no separate bookkeeping.

# Define a radius known to about 5%
radius = na.NormalUncertainScalarArray(
    nominal=10 * u.cm,
    width=0.5 * u.cm,
    num_distribution=11,
)

# Compute the area of the corresponding circle
area = np.pi * np.square(radius)

# The uncertainty in the radius is carried into the area
area.nominal, np.std(area.distribution, axis="_distribution")

Development

Install the package in editable mode along with its test dependencies, and run the test suite using pytest:

pip install -e .[test]
pytest

The suite is large, so continuous integration splits it into five groups using pytest-split. To run one group:

pytest --splits 5 --group 1

This project is linted using ruff, which is checked by continuous integration:

ruff check .

To build the documentation locally:

pip install -e .[doc]
sphinx-build docs docs/_build/html

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

named_arrays-2.8.0.tar.gz (284.0 kB view details)

Uploaded Aug 28, 2026 Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

named_arrays-2.8.0-py3-none-any.whl (287.1 kB view details)

Uploaded Aug 28, 2026 Python 3

File details

Details for the file named_arrays-2.8.0.tar.gz.

File metadata

Download URL: named_arrays-2.8.0.tar.gz
Upload date: Aug 28, 2026
Size: 284.0 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for named_arrays-2.8.0.tar.gz
Algorithm	Hash digest
SHA256	`559b65ba1abe0a495cc3d3961ba0e8ecbc6175e4027de74d77b7d89e90a33056`
MD5	`9733bcc29a2a2ae9ba6ef653085214db`
BLAKE2b-256	`3cff2714789f97567ee1134f5ec36ae318e24b133047d50f10f4b47247add683`

See more details on using hashes here.

Provenance

The following attestation bundles were made for named_arrays-2.8.0.tar.gz:

Publisher: publish.yml on sun-data/named-arrays

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: named_arrays-2.8.0.tar.gz
- Subject digest: 559b65ba1abe0a495cc3d3961ba0e8ecbc6175e4027de74d77b7d89e90a33056
- Sigstore transparency entry: 2628110971
- Sigstore integration time: Aug 28, 2026
Source repository:
- Permalink: sun-data/named-arrays@a01ca4fde19dcd94921c455d683293cd9994635b
- Branch / Tag: refs/tags/v2.8.0
- Owner: https://github.com/sun-data
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: publish.yml@a01ca4fde19dcd94921c455d683293cd9994635b
- Trigger Event: release