wrong chunk sizes in html repr with nonuniform chunks

[dcherian]

What happened:

The HTML repr is using the first element in a chunks tuple;

What you expected to happen:

it should be using whatever dask does in this case

Minimal Complete Verifiable Example:

import xarray as xr
import dask


test = xr.DataArray(
    dask.array.zeros(
        (12, 901, 1001),
        chunks=(
            (1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1),
            (1, 899, 1),
            (1, 199, 1, 199, 1, 199, 1, 199, 1, 199, 1),
        ),
    )
)
test.to_dataset(name="a")

EDIT: The text repr has the same issue

<xarray.Dataset>
Dimensions:  (dim_0: 12, dim_1: 901, dim_2: 1001)
Dimensions without coordinates: dim_0, dim_1, dim_2
Data variables:
    a        (dim_0, dim_1, dim_2) float64 dask.array<chunksize=(1, 1, 1), meta=np.ndarray>

[keewis]

that's because we're actively picking the first element of each chunk:

xarray/xarray/core/formatting.py

Line 233 in 9c85dd5

chunksize = tuple(c[0] for c in array.chunks)

Maybe it's time to finalize _repr_inline_ and try to get dask to implement it? Or, if that doesn't work, we could also monkeypatch it onto dask arrays, similar to what pint-xarray is doing with pint quantities.

[dcherian]

I agree. We should outsource this to dask

[keewis]

for that to work I guess we first have to improve the documentation on _repr_inline_ (see #4248 (comment))