TST: failing sparse test on the numpy/scipy dev build (Azure)

[jorisvandenbossche]

A test started to fail the last hours on the build with numpy/scipy master.
I assume it is related to a recent scipy change, commented about that here: scipy/scipy#9920 (comment)

But it is easy to fix on our side to just update the testing code (the construction of the sparse matrix)

Example log: https://dev.azure.com/pandas-dev/pandas/_build/results?buildId=9238&view=logs&jobId=3a03f79d-0b41-5610-1aa4-b4a014d0bc70

__________________________ TestAccessor.test_from_coo __________________________
[gw0] linux -- Python 3.7.2 /home/vsts/miniconda3/envs/pandas-dev/bin/python

self = <pandas.tests.arrays.sparse.test_array.TestAccessor object at 0x7fa1bf7982b0>

    def test_from_coo(self):
        sparse = pytest.importorskip("scipy.sparse")
    
        row = [0, 3, 1, 0]
        col = [0, 3, 1, 2]
        data = [4, 5, 7, 9]
>       sp_array = sparse.coo_matrix(data, (row, col))

pandas/tests/arrays/sparse/test_array.py:1094: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
../../../miniconda3/envs/pandas-dev/lib/python3.7/site-packages/scipy/sparse/coo.py:187: in __init__
    if check_shape(shape) != self._shape:
../../../miniconda3/envs/pandas-dev/lib/python3.7/site-packages/scipy/sparse/sputils.py:279: in check_shape
    new_shape = tuple(operator.index(arg) for arg in args)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

.0 = <tuple_iterator object at 0x7fa1bf7987b8>

>   new_shape = tuple(operator.index(arg) for arg in args)
E   TypeError: 'list' object cannot be interpreted as an integer

[jorisvandenbossche]

On second thought, it was actually our test that was using sparse.coo_matrix(..) incorrectly (and which now started to be validated on the scipy side, which is a good thing :)).
Updated the PR to fix this accordingly: #25653