Failing rolling window

[bjlittle]

🐛 Bug Report

Unable to perform a rolling window mean over a lazy year auxiliary coordinate.

How to Reproduce

Steps to reproduce the behaviour:

>>> import iris
>>> iris.__version__
'3.12.0'
>>> import iris
>>> fname = "rnt_Eyeargm_HadGEM3-GC31-LL_piControl_r1i1p1f1_gn_18500101-30000101.nc"
>>> cube = iris.load_cube(fname)
>>> print(cube)
toa_net_downward_radiative_flux / (W m-2) (time: 1150)
    Dimension coordinates:
        time                                   x
    Auxiliary coordinates:
        year                                   x
    Scalar coordinates:
        latitude                          0.0 degrees, bound=(-90.0, 90.0) degrees
        longitude                         180.0 degrees, bound=(0.0, 360.0) degrees
    Cell methods:
        0                                 longitude: latitude: mean
    Attributes:
        <snip>
>>> cube
<iris 'Cube' of toa_net_downward_radiative_flux / (W m-2) (time: 1150)>
>>> result = cube.rolling_window("year", iris.analysis.MEAN, 11)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<snip>/lib/python3.12/site-packages/iris/cube.py", line 4922, in rolling_window
    new_bounds = new_bounds[:, (0, -1)]
                 ~~~~~~~~~~^^^^^^^^^^^^
  File "<snip>/lib/python3.12/site-packages/dask/array/core.py", line 2045, in __getitem__
    index2 = normalize_index(index, self.shape)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "<snip>/lib/python3.12/site-packages/dask/array/slicing.py", line 848, in normalize_index
    check_index(axis, i, d)
  File "<snip>/lib/python3.12/site-packages/dask/array/slicing.py", line 919, in check_index
    elif ind >= dimension or ind < -dimension:
         ^^^^^^^^^^^^^^^^
TypeError: '>=' not supported between instances of 'tuple' and 'int'
>>> 

Expected Behaviour

Making the year coordinate points eager fixes the issue i.e.,

>>> import iris
>>> fname = "rnt_Eyeargm_HadGEM3-GC31-LL_piControl_r1i1p1f1_gn_18500101-30000101.nc"
>>> cube = iris.load_cube(fname)
>>> cube.coord("year").has_lazy_points()
True
>>> _ = cube.coord("year").points         # <------ make the points eager
>>> cube.coord("year").has_lazy_points()
False
>>> result = cube.rolling_window("year", iris.analysis.MEAN, 11)
>>> print(result)
toa_net_downward_radiative_flux / (W m-2) (time: 1140)
    Dimension coordinates:
        time                                   x
    Auxiliary coordinates:
        year                                   x
    Scalar coordinates:
        latitude                          0.0 degrees, bound=(-90.0, 90.0) degrees
        longitude                         180.0 degrees, bound=(0.0, 360.0) degrees
    Cell methods:
        0                                 longitude: latitude: mean
        1                                 year: mean
    Attributes:
        <snip>

Environment

• iris 3.12
• python 3.12

[bjlittle]

@alejandrobodas Provided the example data for this issue which has been archived under ${DATADIR}/iris-issue-6480-data for the MO team account 👍