xarray-dataclasses 0.7.0

xarray extension for typed DataArray and Dataset creation

xarray-dataclasses

xarray extension for typed DataArray and Dataset creation

Overview

xarray-dataclasses is a Python package that makes it easy to create typed DataArray and Dataset objects of xarray using the Python's dataclass.

from dataclasses import dataclass
from typing import Literal
from xarray_dataclasses import AsDataArray, Coord, Data


X = Literal["x"]
Y = Literal["y"]


@dataclass
class Image(AsDataArray):
    """Specs for a monochromatic image."""

    data: Data[tuple[X, Y], float]
    x: Coord[X, int] = 0
    y: Coord[Y, int] = 0


# create an image as DataArray
image = Image.new([[0, 1], [2, 3]], x=[0, 1], y=[0, 1])

# create an image filled with ones
ones = Image.ones((2, 2), x=[0, 1], y=[0, 1])

Features

• DataArray and Dataset objects with fixed dimensions, data type, and coordinates can easily be created.
• NumPy-like special functions such as ones() are provided as class methods.
• Compatible with the Python's dataclass.
• Compatible with static type check by Pyright.

Installation

$ pip install xarray-dataclasses

Background

xarray is useful for handling labeled multi-dimensional data, but it is a bit troublesome to create DataArray and Dataset objects with fixed dimensions, data type, or coordinates (typed DataArray and typed Dataset). For example, let us think about the following DataArray specifications for a monochromatic image.

• Dimensions of data must be ("x", "y").
• Data type of data must be float.
• Data type of dimensions must be int.
• Default value of dimensions must be 0.

Then a function to create a typed DataArray object is something like this.

import numpy as np
import xarray as xr


def create_image(data, x=0, y=0):
    """Create a monochromatic image."""
    data = np.array(data)

    if x == 0:
        x = np.full(data.shape[0], x)
    else:
        x = np.array(x)

    if y == 0:
        y = np.full(data.shape[1], y)
    else:
        y = np.array(y)

    return xr.DataArray(
        data=data.astype(float),
        dims=("x", "y"),
        coords={
            "x": ("x", x.astype(int)),
            "y": ("y", y.astype(int)),
        },
    )


image = create_image([[0, 1], [2, 3]])

The issues are

• It is not easy to figure out the specifications from the code.
• It is not easy to reuse the code, for example, to add new coordinates.

xarray-dataclasses resolves them by defining the specifications as a dataclass. As shown in the code in the overview, the specifications become much easier to read.

• The type hints have complete information for DataArray creation.
• The default values are given as class variables.
• The mix-in class AsDataArray provides class methods such as new().
• The extension of the specifications is easy by class inheritance.

Basic usage

xarray-dataclasses uses the Python's dataclass. Please learn how to use it before proceeding. Data (or data variables), coordinates, attributes, and a name of a DataArray or a Dataset object are defined as dataclass fields with the following type hints. Note that the following code is supposed in the examples below.

from dataclasses import dataclass
from typing import Literal
from xarray_dataclasses import AsDataArray, AsDataset
from xarray_dataclasses import Attr, Coord, Data, Name


X = Literal["x"]
Y = Literal["y"]

Data field

The data field is a field whose value will become the data of a DataArray object or a data variable of a Dataset object. The type hint Data[TDims, TDtype] fixes the dimensions and the data type of the object. Here are some examples of how to specify them.

Type hint	Inferred dimensions
Data[Literal[()], ...]	()
Data[Literal["x"], ...]	("x",)
Data[tuple[Literal["x"], Literal["y"]], ...]	("x", "y")

Type hint	Inferred data type
Data[..., Any]	None
Data[..., None]	None
Data[..., float]	numpy.dtype("float64")
Data[..., numpy.float128]	numpy.dtype("float128")
Data[..., Literal["datetime64[ns]"]]	numpy.dtype("<M8[ns]")

Coordinate field

The coordinate field is a field whose value will become a coordinate of a DataArray or a Dataset object. The type hint Coord[TDims, TDtype] fixes the dimensions and the data type of the object.

Attribute field

The attribute field is a field whose value will become an attribute of a DataArray or a Dataset object. The type hint Attr[T] specifies the type of the value, which is used only for static type check.

Name field

The name field is a field whose value will become the name of a DataArray object. The type hint Name[T] specifies the type of the value, which is used only for static type check.

DataArray class

The DataArray class is a dataclass that defines typed DataArray specifications. Exactly one data field is allowed in a DataArray class. The second and subsequent data fields are just ignored in DataArray creation.

@dataclass
class Image(AsDataArray):
    """Specs for a monochromatic image."""

    data: Data[tuple[X, Y], float]
    x: Coord[X, int] = 0
    y: Coord[Y, int] = 0
    units: Attr[str] = "cd / m^2"
    name: Name[str] = "luminance"

A DataArray object is created by the shorthand method new().

Image.new([[0, 1], [2, 3]], x=[0, 1], y=[0, 1])

<xarray.DataArray "luminance" (x: 2, y: 2)>
array([[0., 1.],
       [2., 3.]])
Coordinates:
  * x        (x) int64 0 1
  * y        (y) int64 0 1
Attributes:
    units:    cd / m^2

NumPy-like empty(), zeros(), ones(), full() methods are available.

Image.ones((3, 3))

<xarray.DataArray "luminance" (x: 3, y: 3)>
array([[1., 1., 1.],
       [1., 1., 1.],
       [1., 1., 1.]])
Coordinates:
  * x        (x) int64 0 0 0
  * y        (y) int64 0 0 0
Attributes:
    units:    cd / m^2

Dataset class

The Dataset class is a dataclass that defines typed Dataset specifications. Multiple data fields are allowed to define the data variables of the object.

@dataclass
class ColorImage(AsDataset):
    """Specs for a color image."""

    red: Data[tuple[X, Y], float]
    green: Data[tuple[X, Y], float]
    blue: Data[tuple[X, Y], float]
    x: Coord[X, int] = 0
    y: Coord[Y, int] = 0
    units: Attr[str] = "cd / m^2"

A Dataset object is created by the shorthand method new().

ColorImage.new(
    [[0, 0], [0, 0]],  # red
    [[1, 1], [1, 1]],  # green
    [[2, 2], [2, 2]],  # blue
)

<xarray.Dataset>
Dimensions:  (x: 2, y: 2)
Coordinates:
  * x        (x) int64 0 0
  * y        (y) int64 0 0
Data variables:
    red      (x, y) float64 0.0 0.0 0.0 0.0
    green    (x, y) float64 1.0 1.0 1.0 1.0
    blue     (x, y) float64 2.0 2.0 2.0 2.0
Attributes:
    units:    cd / m^2

Advanced usage

Coordof and Dataof type hints

xarray-dataclasses provides advanced type hints, Coordof[T] and Dataof[T]. Unlike Data and Coord, they specify a dataclass that defines a DataArray class. This is useful, for example, when users want to add metadata to dimensions for plotting.

from xarray_dataclasses import Coordof


@dataclass
class XAxis:
    """Specs for the x axis."""

    data: Data[X, int]
    long_name: Attr[str] = "x axis"
    units: Attr[str] = "pixel"


@dataclass
class YAxis:
    """Specs for the y axis."""

    data: Data[Y, int]
    long_name: Attr[str] = "y axis"
    units: Attr[str] = "pixel"


@dataclass
class Image(AsDataArray):
    """Specs for a monochromatic image."""

    data: Data[tuple[X, Y], float]
    x: Coordof[XAxis] = 0
    y: Coordof[YAxis] = 0

Custom DataArray and Dataset factories

For customization, users can use a function or a class to create an initial DataArray or Dataset object by specifying a special class attribute, __dataarray_factory__ or __dataset_factory__, respectively.

import xarray as xr


class Custom(xr.DataArray):
    """Custom DataArray."""

    __slots__ = ()

    def custom_method(self) -> None:
        print("Custom method!")


@dataclass
class Image(AsDataArray):
    """Specs for a monochromatic image."""

    data: Data[tuple[X, Y], float]
    x: Coord[X, int] = 0
    y: Coord[Y, int] = 0
    __dataarray_factory__ = Custom


image = Image.ones([3, 3])
isinstance(image, Custom) # True
image.custom_method() # Custom method!

DataArray and Dataset creation without shorthands

xarray-dataclasses provides functions, asdataarray and asdataset. This is useful, for example, users do not want to inherit the mix-in class (AsDataArray or AsDataset) in a DataArray or Dataset dataclass.

from xarray_dataclasses import asdataarray


@dataclass
class Image:
    """Specifications of images."""

    data: Data[tuple[X, Y], float]
    x: Coord[X, int] = 0
    y: Coord[Y, int] = 0


image = asdataarray(Image([[0, 1], [2, 3]], x=[0, 1], y=[0, 1]))