autocti.Array2D#

class autocti.Array2D(values: Union[ndarray, List, AbstractArray2D], mask: Mask2D, header: Optional[Header] = None, store_native: bool = False, skip_mask: bool = False, *args, **kwargs)[source]#

Bases: AbstractArray2D

A uniform 2D array of values, which are paired with a 2D mask of pixels which may be split into sub-pixels.

The Array2D`, like all data structures (e.g. ``Grid2D, VectorYX2D) has in-built functionality which:

  • Applies a 2D mask (a Mask2D object) to the da_ta structure’s values.

  • Maps the data structure between two data representations: slim` (all unmasked values in a 1D ndarray) and native (all unmasked values in a 2D ndarray).

  • Associates Cartesian Grid2D objects of (y,x) coordinates with the data structure (e.g. a (y,x) grid of all unmasked pixels).

  • Associates sub-grids with the data structure, which perform calculations higher resolutions which are then binned up.

Each entry of an Array2D corresponds to a value at the centre of a sub-pixel in its corresponding Mask2D. It is ordered such that pixels begin from the top-row of the corresponding mask and go right and down. The positive y-axis is upwards and positive x-axis to the right.

A detailed description of the data structure API is provided below.

SLIM DATA REPRESENTATION (sub-size=1)

Below is a visual illustration of an Array2D’s 2D mask, where a total of 10 pixels are unmasked and are included in the array.

x x x x x x x x x x
x x x x x x x x x x     This is an example ``Mask2D``, where:
x x x x x x x x x x
x x x x O O x x x x     x = `True` (Pixel is masked and excluded from the array)
x x x O O O O x x x     O = `False` (Pixel is not masked and included in the array)
x x x O O O O x x x
x x x x x x x x x x
x x x x x x x x x x
x x x x x x x x x x
x x x x x x x x x x

The mask pixel index’s are as follows (the positive / negative direction of the Grid2D objects associated with the array are also shown on the y and x axes).

<--- -ve  x  +ve -->

 x x x x x x x x x x  ^   array_2d[0] = 10
 x x x x x x x x x x  I   array_2d[1] = 20
 x x x x x x x x x x  I   array_2d[2] = 30
 x x x x 0 1 x x x x +ve  array_2d[3] = 40
 x x x 2 3 4 5 x x x  y   array_2d[4] = 50
 x x x 6 7 8 9 x x x -ve  array_2d[5] = 60
 x x x x x x x x x x  I   array_2d[6] = 70
 x x x x x x x x x x  I   array_2d[7] = 80
 x x x x x x x x x x \/   array_2d[8] = 90
 x x x x x x x x x x      array_2d[9] = 100

The Array2D in its slim data representation is an ndarray of shape [total_unmasked_pixels].

For the Mask2D above the slim representation therefore contains 10 entries and two examples of these entries are:

array[3] = the 4th unmasked pixel's value, given by value 40 above.
array[6] = the 7th unmasked pixel's value, given by value 80 above.

A Cartesian grid of (y,x) coordinates, corresponding to all slim values (e.g. unmasked pixels) is given by array_2d.derive_grid.masked.slim.

NATIVE DATA REPRESENTATION (sub_size=1)

The Array2D above, but represented as an an ndarray of shape [total_y_values, total_x_values], where all masked entries have values of 0.0.

For the following mask:

x x x x x x x x x x
x x x x x x x x x x     This is an example ``Mask2D``, where:
x x x x x x x x x x
x x x x O O x x x x     x = `True` (Pixel is masked and excluded from the array)
x x x O O O O x x x     O = `False` (Pixel is not masked and included in the array)
x x x O O O O x x x
x x x x x x x x x x
x x x x x x x x x x
x x x x x x x x x x
x x x x x x x x x x

Where the array has the following indexes (left figure) and values (right):

<--- -ve  x  +ve -->

 x x x x x x x x x x  ^   array_2d[0] = 10
 x x x x x x x x x x  I   array_2d[1] = 20
 x x x x x x x x x x  I   array_2d[2] = 30
 x x x x 0 1 x x x x +ve  array_2d[3] = 40
 x x x 2 3 4 5 x x x  y   array_2d[4] = 50
 x x x 6 7 8 9 x x x -ve  array_2d[5] = 60
 x x x x x x x x x x  I   array_2d[6] = 70
 x x x x x x x x x x  I   array_2d[7] = 80
 x x x x x x x x x x \/   array_2d[8] = 90
 x x x x x x x x x x      array_2d[9] = 100

In the above array:

- array[0,0] = 0.0 (it is masked, thus zero)
- array[0,0] = 0.0 (it is masked, thus zero)
- array[3,3] = 0.0 (it is masked, thus zero)
- array[3,3] = 0.0 (it is masked, thus zero)
- array[3,4] = 10
- array[3,5] = 20
- array[4,5] = 50

SLIM TO NATIVE MAPPING

The Array2D has functionality which maps data between the slim and native data representations.

For the example mask above, the 1D ndarray given by mask.derive_indexes.slim_to_native is:

slim_to_native[0] = [3,4]
slim_to_native[1] = [3,5]
slim_to_native[2] = [4,3]
slim_to_native[3] = [4,4]
slim_to_native[4] = [4,5]
slim_to_native[5] = [4,6]
slim_to_native[6] = [5,3]
slim_to_native[7] = [5,4]
slim_to_native[8] = [5,5]
slim_to_native[9] = [5,6]

SUB GRIDDING

If the Mask2D sub_size is > 1, the array has entries corresponding to the values at the centre of every sub-pixel of each unmasked pixel.

The sub-array indexes are ordered such that pixels begin from the first (top-left) sub-pixel in the first unmasked pixel. Indexes then go over the sub-pixels in each unmasked pixel, for every unmasked pixel.

Therefore, the shapes of the sub-array are as follows:

  • slim representation: an ndarray of shape [total_unmasked_pixels*sub_size**2].

  • native representation: an ndarray of shape [total_y_values*sub_size, total_x_values*sub_size].

Below is a visual illustration of a sub array. Indexing of each sub-pixel goes from the top-left corner. In contrast to the array above, our illustration below restricts the mask to just 2 pixels, to keep the illustration brief.

x x x x x x x x x x
x x x x x x x x x x     This is an example ``Mask2D``, where:
x x x x x x x x x x
x x x x x x x x x x     x = `True` (Pixel is masked and excluded from lens)
x 0 0 x x x x x x x     O = `False` (Pixel is not masked and included in lens)
x x x x x x x x x x
x x x x x x x x x x
x x x x x x x x x x
x x x x x x x x x x
x x x x x x x x x x

If sub_size=2, each unmasked pixel has 4 (2x2) sub-pixel values. For the example above, pixels 0 and 1 each have 4 values which map to the array_2d’s slim representation as follows:

Pixel 0 - (2x2):

       array_2d.slim[0] = value of first sub-pixel in pixel 0.
 0 1   array_2d.slim[1] = value of first sub-pixel in pixel 1.
 2 3   array_2d.slim[2] = value of first sub-pixel in pixel 2.
       array_2d.slim[3] = value of first sub-pixel in pixel 3.

Pixel 1 - (2x2):

       array_2d.slim[4] = value of first sub-pixel in pixel 0.
 4 5   array_2d.slim[5] = value of first sub-pixel in pixel 1.
 6 7   array_2d.slim[6] = value of first sub-pixel in pixel 2.
       array_2d.slim[7] = value of first sub-pixel in pixel 3.

For the native data representation we get the following mappings:

Pixel 0 - (2x2):

       array_2d.native[8, 2] = value of first sub-pixel in pixel 0.
 0 1   array_2d.native[8, 3] = value of first sub-pixel in pixel 1.
 2 3   array_2d.native[9, 2] = value of first sub-pixel in pixel 2.
       array_2d.native[9, 3] = value of first sub-pixel in pixel 3.

Pixel 1 - (2x2):

       array_2d.native[10, 4] = value of first sub-pixel in pixel 0.
 4 5   array_2d.native[10, 5] = value of first sub-pixel in pixel 1.
 6 7   array_2d.native[11, 4] = value of first sub-pixel in pixel 2.
       array_2d.native[11, 5] = value of first sub-pixel in pixel 3.

Other entries (all masked sub-pixels are zero):

       array_2d.native[0, 0] = 0.0 (it is masked, thus zero)
       array_2d.native[15, 12] = 0.0 (it is masked, thus zero)

If we used a sub_size of 3, for pixel 0 we we would create a 3x3 sub-array:

        array_2d.slim[0] = value of first sub-pixel in pixel 0.
        array_2d.slim[1] = value of first sub-pixel in pixel 1.
        array_2d.slim[2] = value of first sub-pixel in pixel 2.
0 1 2   array_2d.slim[3] = value of first sub-pixel in pixel 3.
3 4 5   array_2d.slim[4] = value of first sub-pixel in pixel 4.
6 7 8   array_2d.slim[5] = value of first sub-pixel in pixel 5.
        array_2d.slim[6] = value of first sub-pixel in pixel 6.
        array_2d.slim[7] = value of first sub-pixel in pixel 7.
        array_2d.slim[8] = value of first sub-pixel in pixel 8.

In PyAutoCTI all Array2D objects are used in their native representation without sub-gridding. Significant memory can be saved by only store this format, thus the native_binned_only config override can force this behaviour. It is recommended users do not use this option to avoid unexpected behaviour.

Parameters:

  • values – The values of the array, which can be input in the slim or native format.

  • mask – The 2D mask associated with the array, defining the pixels each array value in its slim representation is paired with.

  • store_native – If True, the ndarray is stored in its native format [total_y_pixels, total_x_pixels]. This avoids mapping large data arrays to and from the slim / native formats, which can be a computational bottleneck.

Examples

This example uses the Array2D.no_mask method to create the Array2D.

Different methods using different inputs are available and documented throughout this webpage.

import autoarray as aa

# Make Array2D from input np.ndarray with sub_size 1.

array_2d = aa.Array2D.no_mask(
    values=np.array([1.0, 2.0, 3.0, 4.0]),
    shape_native=(2, 2),
    pixel_scales=1.0,
    sub_size=1
)

# Make Array2D from input list with different shape_native and sub_size 1.

array_2d = aa.Array2D.no_mask(
    values=[1.0, 2.0, 3.0, 4.0, 5.0, 6.0],
    shape_native=(2, 3),
    pixel_scales=1.0,
    sub_size=1
)

import autoarray as aa

# Make Array2D with sub_size 2.

array_2d = aa.Array2D.no_mask(
    values=[1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0],
    shape_native=(2, 1),
    pixel_scales=1.0,
    sub_size=2,
)

# Apply 2D mask to Array2D with sub_size 2, where the
# True value masks entries (5.0, 6.0, 7.0, 8.0).

mask = aa.Mask2D(
    mask=[[False], [True]],
    pixel_scales=2.0,
    sub_size=2
)

array_2d = array_2d.apply_mask(mask=mask)

# Print certain array attributes.

print(array_2d.slim) # masked 1D data representation on sub-grid.
print(array_2d.native) # masked 2D data representation on sub-grid.
print(array_2d.slim.binned) # masked 1D data representation binned up from sub-grid.
print(array_2d.native.binned) # masked 2D data representation binned up from sub-grid.

# Output array to .fits file.

array_2d.output_to_fits(file_path="/path/for/output")

Methods

all

apply_mask

astype

copy

extent_of_zoomed_array

For an extracted zoomed array computed from the method zoomed_around_mask compute its extent in scaled coordinates.

flip_hdu_for_ds9

from_fits

Returns an Array2D by loading the array values from a .fits file.

from_primary_hdu

Returns an Array2D by from a PrimaryHDU object which has been loaded via astropy.fits

from_yx_and_values

Returns an Array2D by inputting the y and x pixel values where the array is filled and the values that fill it.

full

Returns an Array2D where all values are filled with an input fill value, analogous to np.full().

instance_flatten

Flatten an instance of an autoarray class into a tuple of its attributes (i.e.

instance_unflatten

Unflatten a tuple of attributes (i.e.

invert

max

min

no_mask

Returns an Array2D from an array via inputs in its slim or native data representation.

ones

Returns an Array2D where all values are filled with ones, analogous to np.ones().

output_to_fits

Output the array to a .fits file.

padded_before_convolution_from

When the edge pixels of a mask are unmasked and a convolution is to occur, the signal of edge pixels will be 'missing' if the grid is used to evaluate the signal via an analytic function.

reshape

resized_from

Resize the array around its centre to a new input shape.

sqrt

structure_2d_from

structure_2d_list_from

sum

trimmed_after_convolution_from

When the edge pixels of a mask are unmasked and a convolution is to occur, the signal of edge pixels will be 'missing' if the grid is used to evaluate the signal via an analytic function.

with_new_array

Copy this object but give it a new array.

zeros

Returns an Array2D where all values are filled with zeros, analogous to np.zeros().

zoomed_around_mask

Extract the 2D region of an array corresponding to the rectangle encompassing all unmasked values.

Attributes

array

binned

Convenience method to access the binned-up array in its 1D representation, which is a Grid2D stored as an ndarray of shape [total_unmasked_pixels, 2].

binned_across_columns

Bins the 2D array up to a 1D array, where each value is the mean of all unmasked values in each column.

binned_across_rows

Bins the 2D array up to a 1D array, where each value is the mean of all unmasked values in each row.

derive_grid

derive_indexes

derive_mask

dtype

geometry

hdu_for_output

The array as an HDU object, which can be output to a .fits file.

imag

in_counts

in_counts_per_second

native

Return a Array2D where the data is stored in its native representation, which is an ndarray of shape [sub_size*total_y_pixels, sub_size*total_x_pixels].

native_skip_mask

Return a Array2D where the data is stored in its native representation, which is an ndarray of shape [sub_size*total_y_pixels, sub_size*total_x_pixels].

ndim

origin

original_orientation

pixel_area

pixel_scale

pixel_scale_header

pixel_scales

readout_offsets

real

shape

shape_native

shape_slim

size

slim

Return an Array2D where the data is stored its slim representation, which is an ndarray of shape [total_unmasked_pixels * sub_size**2].

store_native

sub_shape_native

sub_shape_slim

sub_size

total_area

total_pixels

unmasked_grid

values

classmethod no_mask(values: Union[ndarray, List, AbstractArray2D], pixel_scales: Union[Tuple[float], Tuple[float, float], float], shape_native: Optional[Tuple[int, int]] = None, sub_size: int = 1, origin: Tuple[float, float] = (0.0, 0.0), header: Optional[Header] = None) → Array2D[source]#

Returns an Array2D from an array via inputs in its slim or native data representation.

From a slim 1D input the method cannot determine the 2D shape of the array and its mask. The shape_native must therefore also be input into this method. The mask is setup as a unmasked Mask2D of shape_native.

For a full description of Array2D objects, including a description of the slim and native attribute used by the API, see the Array2D class API documentation.

Parameters:

  • values – The values of the array input with shape [total_unmasked_pixels*(sub_size**2)] or shape [total_y_pixels*sub_size, total_x_pixel*sub_size].

  • pixel_scales – The (y,x) scaled units to pixel units conversion factors of every pixel. If this is input as a float, it is converted to a (float, float) structure.

  • shape_native – The 2D shape of the array in its native format, and its 2D mask (only required if input shape is in slim format).

  • sub_size – The size (sub_size x sub_size) of each unmasked pixels sub-array.

  • origin – The (y,x) scaled units origin of the mask’s coordinate system.

Examples

import autoarray as aa

# Make Array2D from input list, native format with sub_size 1
# (This array has shape_native=(2,2)).

array_2d = aa.Array2D.manual(
    array=np.array([[1.0, 2.0], [3.0, 4.0]]),
    pixel_scales=1.0.
    sub_size=1
)

import autoarray as aa

# Make Array2D from input list, slim format with sub_size 2.

array_2d = aa.Array2D.no_mask(
    values=[1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0],
    shape_native=(2, 1),
    pixel_scales=1.0,
    sub_size=2,
)
classmethod full(fill_value: float, shape_native: Tuple[int, int], pixel_scales: Union[Tuple[float], Tuple[float, float], float], sub_size: int = 1, origin: Tuple[float, float] = (0.0, 0.0), header: Optional[Header] = None) → Array2D[source]#

Returns an Array2D where all values are filled with an input fill value, analogous to np.full().

For a full description of Array2D objects, including a description of the slim and native attribute used by the API, see the Array2D class API documentation.

From this input the method cannot determine the 2D shape of the array and its mask. The shape_native must therefore also be input into this method. The mask is setup as a unmasked Mask2D of shape_native.

Parameters:

  • fill_value – The value all array elements are filled with.

  • shape_native – The 2D shape of the array in its native format, and its 2D mask.

  • pixel_scales – The (y,x) scaled units to pixel units conversion factors of every pixel. If this is input as a float, it is converted to a (float, float) structure.

  • sub_size – The size (sub_size x sub_size) of each unmasked pixels sub-array.

  • origin – The (y,x) scaled units origin of the mask’s coordinate system.

Examples

import autoarray as aa

# Make Array2D with sub_size 1.

array_2d = aa.Array2D.full(
    fill_value=2.0,
    shape_native=(2, 2),
    pixel_scales=1.0,
    sub_size=1
)

import autoarray as aa

# Make Array2D with sub_size 2.

array_2d = aa.Array2D.full(
    fill_value=2.0,
    shape_native=(2, 2),
    pixel_scales=1.0,
    sub_size=2
)
classmethod ones(shape_native: Tuple[int, int], pixel_scales: Union[Tuple[float], Tuple[float, float], float], sub_size: int = 1, origin: Tuple[float, float] = (0.0, 0.0), header: Optional[Header] = None) → Array2D[source]#

Returns an Array2D where all values are filled with ones, analogous to np.ones().

For a full description of Array2D objects, including a description of the slim and native attribute used by the API, see the Array2D class API documentation.

From this input the method cannot determine the 2D shape of the array and its mask. The shape_native must therefore also be input into this method. The mask is setup as a unmasked Mask2D of shape_native.

Parameters:

  • shape_native – The 2D shape of the array in its native format, and its 2D mask.

  • pixel_scales – The (y,x) scaled units to pixel units conversion factors of every pixel. If this is input as a float, it is converted to a (float, float) structure.

  • sub_size – The size (sub_size x sub_size) of each unmasked pixels sub-array.

  • origin – The (y,x) scaled units origin of the mask’s coordinate system.

Examples

import autoarray as aa

# Make Array2D with sub_size 1.

array_2d = aa.Array2D.ones(
    shape_native=(2, 2),
    pixel_scales=1.0,
    sub_size=1
)

import autoarray as aa

# Make Array2D with sub_size 2.

array_2d = aa.Array2D.ones(
    shape_native=(2, 2),
    pixel_scales=1.0,
    sub_size=2
)
classmethod zeros(shape_native: Tuple[int, int], pixel_scales: Union[Tuple[float], Tuple[float, float], float], sub_size: int = 1, origin: Tuple[float, float] = (0.0, 0.0), header: Optional[Header] = None) → Array2D[source]#

Returns an Array2D where all values are filled with zeros, analogous to np.zeros().

For a full description of Array2D objects, including a description of the slim and native attribute used by the API, see the Array2D class API documentation.

From this input the method cannot determine the 2D shape of the array and its mask. The shape_native must therefore also be input into this method. The mask is setup as a unmasked Mask2D of shape_native.

Parameters:

  • shape_native – The 2D shape of the array in its native format, and its 2D mask.

  • pixel_scales – The (y,x) scaled units to pixel units conversion factors of every pixel. If this is input as a float, it is converted to a (float, float) structure.

  • sub_size – The size (sub_size x sub_size) of each unmasked pixels sub-array.

  • origin – The (y,x) scaled units origin of the mask’s coordinate system.

Examples

import autoarray as aa

# Make Array2D with sub_size 1.

array_2d = aa.Array2D.zeros(
    shape_native=(2, 2),
    pixel_scales=1.0,
    sub_size=1
)

import autoarray as aa

# Make Array2D with sub_size 2.

array_2d = aa.Array2D.zeros(
    shape_native=(2, 2),
    pixel_scales=1.0,
    sub_size=2
)
classmethod from_fits(file_path: Union[Path, str], pixel_scales: Optional[Union[Tuple[float], Tuple[float, float], float]], hdu: int = 0, sub_size: int = 1, origin: Tuple[float, float] = (0.0, 0.0)) → Array2D[source]#

Returns an Array2D by loading the array values from a .fits file.

For a full description of Array2D objects, including a description of the slim and native attribute used by the API, see the Array2D class API documentation.

Parameters:

  • file_path – The path the file is loaded from, including the filename and the .fits extension, e.g. ‘/path/to/filename.fits’

  • pixel_scales – The (y,x) scaled units to pixel units conversion factors of every pixel. If this is input as a float, it is converted to a (float, float) structure.

  • hdu – The Header-Data Unit of the .fits file the array data is loaded from.

  • sub_size – The size (sub_size x sub_size) of each unmasked pixels sub-array.

  • origin – The (y,x) scaled units origin of the coordinate system.

Examples

import autoarray as aa

# Make Array2D with sub_size 1.

array_2d = aa.Array2D.from_fits(
    file_path="path/to/file.fits",
    hdu=0,
    pixel_scales=1.0,
    sub_size=1
)

import autoarray as aa

# Make Array2D with sub_size 2.
# (It is uncommon that a sub-gridded array would be loaded from
# a .fits, but the API support its).

array_2d = aa.Array2D.from_fits(
    file_path="path/to/file.fits",
    hdu=0,
    pixel_scales=1.0,
    sub_size=2
)
classmethod from_primary_hdu(primary_hdu: PrimaryHDU, sub_size: int = 1, origin: Tuple[float, float] = (0.0, 0.0)) → Array2D[source]#

Returns an Array2D by from a PrimaryHDU object which has been loaded via astropy.fits

This assumes that the header of the PrimaryHDU contains an entry named PIXSCALE which gives the pixel-scale of the array.

For a full description of Array2D objects, including a description of the slim and native attribute used by the API, see the Array2D class API documentation.

Parameters:

  • primary_hdu – The PrimaryHDU object which has already been loaded from a .fits file via astropy.fits and contains the array data and the pixel-scale in the header with an entry named PIXSCALE.

  • sub_size – The size (sub_size x sub_size) of each unmasked pixels sub-array.

  • origin – The (y,x) scaled units origin of the coordinate system.

Examples

from astropy.io import fits
import autoarray as aa

# Make Array2D with sub_size 1.

primary_hdu = fits.open("path/to/file.fits")

array_2d = aa.Array2D.from_primary_hdu(
    primary_hdu=primary_hdu,
    sub_size=1
)

import autoarray as aa

# Make Array2D with sub_size 2.
# (It is uncommon that a sub-gridded array would be loaded from
# a .fits, but the API support its).

 primary_hdu = fits.open("path/to/file.fits")

array_2d = aa.Array2D.from_primary_hdu(
    primary_hdu=primary_hdu,
    sub_size=2
)
classmethod from_yx_and_values(y: Union[ndarray, List], x: Union[ndarray, List], values: Union[ndarray, List], shape_native: Tuple[int, int], pixel_scales: Union[Tuple[float], Tuple[float, float], float], sub_size: int = 1, header: Optional[Header] = None) → Array2D[source]#

Returns an Array2D by inputting the y and x pixel values where the array is filled and the values that fill it.

For a full description of Array2D objects, including a description of the slim and native attribute used by the API, see the Array2D class API documentation.

Parameters:

  • y – The y pixel indexes where value are input, with shape [total_unmasked_pixels*sub_size].

  • x – The x pixel indexes where value are input, with shape [total_unmasked_pixels*sub_size].

  • list (values or) – The values which are used to fill in the array, with shape [total_unmasked_pixels*sub_size].

  • shape_native – The 2D shape of the array in its native format, and its 2D mask.

  • pixel_scales – The (y,x) scaled units to pixel units conversion factors of every pixel. If this is input as a float, it is converted to a (float, float) structure.

  • sub_size – The size (sub_size x sub_size) of each unmasked pixels sub-grid.

  • origin – The origin of the grid’s mask.

Examples

import autoarray as aa

# Make Array2D with sub_size 1.

array_2d = aa.Array2D.from_yx_and_values(
    y=np.array([0.5, 0.5, -0.5, -0.5]),
    x=np.array([-0.5, 0.5, -0.5, 0.5]),
    values=np.array([1.0, 2.0, 3.0, 4.0]),
    shape_native=(2, 2),
    pixel_scales=1.0,
    sub_size=1,
)

import autoarray as aa

# Make Array2D with sub_size 2.

array_2d = aa.Array2D.from_yx_and_values(
    y=np.array([1.0, 1.0. 0.5, 0.5, -0.5, -0.5, -1.0, -1.0]),
    x=np.array([-0.5, 0.5, -0.5, 0.5, -0.5, 0.5, -0.5, 0.5]),
    values=np.array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0]),
    shape_native=(2, 1),
    pixel_scales=1.0,
    sub_size=2,
)