Support for Scipy Sparse Arrays

What happened?

Now that Scipy is moving to support sparse NDarrays, we would expect that Xarray should work with them as any other array like data.

What did you expect to happen?

Doesn't work. It seems that why trying to use a scipy sparse array as the data, Xarray wraps the the sparse array in a 0-D dense array. (there are likely more issues after this but this was the first hurdle)

With sparse array s:

print(s)
<4x4 sparse array of type '<class 'numpy.float64'>'
	with 4 stored elements in COOrdinate format>

print(xr.DataArray(s).data)
array(<4x4 sparse array of type '<class 'numpy.float64'>'
	with 4 stored elements in COOrdinate format>, dtype=object)

Minimal Complete Verifiable Example

import numpy as np
import xarray as xr
from scipy.sparse import coo_array

row  = np.array([0, 3, 1, 0])

col  = np.array([0, 3, 1, 2])

data = np.array([4, 5.4, 7, 9.2])

s= coo_array((data, (row, col)), shape=(4, 4))
da = xr.DataArray(s)
print(da._repr_html_())

MVCE confirmation

• Minimal example — the example is as focused as reasonably possible to demonstrate the underlying issue in xarray.
• Complete example — the example is self-contained, including all data and the text of any traceback.
• Verifiable example — the example copy & pastes into an IPython prompt or Binder notebook, returning the result.
• New issue — a search of GitHub Issues suggests this is not a duplicate.

Relevant log output

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Input In [4], in <cell line: 13>()
     11 s= coo_array((data, (row, col)), shape=(4, 4))
     12 da = xr.DataArray(s)
---> 13 print(da._repr_html_())

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/xarray/core/common.py:167, in AbstractArray._repr_html_(self)
    165 if OPTIONS["display_style"] == "text":
    166     return f"<pre>{escape(repr(self))}</pre>"
--> 167 return formatting_html.array_repr(self)

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/xarray/core/formatting_html.py:311, in array_repr(arr)
    303 arr_name = f"'{arr.name}'" if getattr(arr, "name", None) else ""
    305 header_components = [
    306     f"<div class='xr-obj-type'>{obj_type}</div>",
    307     f"<div class='xr-array-name'>{arr_name}</div>",
    308     format_dims(dims, indexed_dims),
    309 ]
--> 311 sections = [array_section(arr)]
    313 if hasattr(arr, "coords"):
    314     sections.append(coord_section(arr.coords))

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/xarray/core/formatting_html.py:219, in array_section(obj)
    213 collapsed = (
    214     "checked"
    215     if _get_boolean_with_default("display_expand_data", default=True)
    216     else ""
    217 )
    218 variable = getattr(obj, "variable", obj)
--> 219 preview = escape(inline_variable_array_repr(variable, max_width=70))
    220 data_repr = short_data_repr_html(obj)
    221 data_icon = _icon("icon-database")

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/xarray/core/formatting.py:274, in inline_variable_array_repr(var, max_width)
    272     return var._data._repr_inline_(max_width)
    273 if var._in_memory:
--> 274     return format_array_flat(var, max_width)
    275 dask_array_type = array_type("dask")
    276 if isinstance(var._data, dask_array_type):

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/xarray/core/formatting.py:191, in format_array_flat(array, max_width)
    188 # every item will take up at least two characters, but we always want to
    189 # print at least first and last items
    190 max_possibly_relevant = min(max(array.size, 1), max(math.ceil(max_width / 2.0), 2))
--> 191 relevant_front_items = format_items(
    192     first_n_items(array, (max_possibly_relevant + 1) // 2)
    193 )
    194 relevant_back_items = format_items(last_n_items(array, max_possibly_relevant // 2))
    195 # interleave relevant front and back items:
    196 #     [a, b, c] and [y, z] -> [a, z, b, y, c]

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/xarray/core/formatting.py:180, in format_items(x)
    177     elif np.logical_not(time_needed).all():
    178         timedelta_format = "date"
--> 180 formatted = [format_item(xi, timedelta_format) for xi in x]
    181 return formatted

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/xarray/core/formatting.py:180, in <listcomp>(.0)
    177     elif np.logical_not(time_needed).all():
    178         timedelta_format = "date"
--> 180 formatted = [format_item(xi, timedelta_format) for xi in x]
    181 return formatted

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/xarray/core/formatting.py:161, in format_item(x, timedelta_format, quote_strings)
    159     return repr(x) if quote_strings else x
    160 elif hasattr(x, "dtype") and np.issubdtype(x.dtype, np.floating):
--> 161     return f"{x.item():.4}"
    162 else:
    163     return str(x)

File ~/Scratch/.conda/envs/tessa-1/lib/python3.10/site-packages/scipy/sparse/_base.py:771, in spmatrix.__getattr__(self, attr)
    769     return self.getnnz()
    770 else:
--> 771     raise AttributeError(attr + " not found")

AttributeError: item not found

Anything else we need to know?

No response

Environment

Start by running the provided Python MVCE with scipy.sparse.coo_array and inspect where xr.DataArray(s) stores the data and where repr_html reaches formatting.py. Trace the dense 0-D wrapping and the AttributeError from formatting; done means sparse arrays remain usable as DataArray data and the example renders without this failure. No repository files or tests are named in the issue.