Fix: collapse dimension for single nested tuple MultiIndex key

[C1-BA-B1-F3]

Problem

When indexing a DataArray with a single nested tuple MultiIndex key (where one level has tuple-valued entries), the result is located correctly but the MultiIndex dimension is incorrectly preserved in the output.

nested_level_0 = pd.Index([(1, 1), (1, 1), (2, 2), (3, 3)], name="a", tupleize_cols=False)
nested_level_1 = pd.Index([1, 2, 10, 20], name="b")
nested_mi = pd.MultiIndex.from_arrays([nested_level_0, nested_level_1])

nested = xr.DataArray(np.arange(4), dims=("index",), ...)
result = nested.sel(index=((1, 1), 2))

# Before: result.dims == ('index',), result.shape == (1,)  # WRONG
# After:  result.dims == (), result.shape == ()             # CORRECT

Root Cause

In PandasMultiIndex.sel(), the _is_nested_tuple() helper detected that ((1, 1), 2) contains a sub-tuple (1, 1) and routed it to get_locs(), which returns an array. But this tuple is actually a single key for the MultiIndex (level a = (1, 1), level b = 2), and get_loc() correctly returns a scalar integer for it.

Fix

Try get_loc() first when a nested tuple is detected. If it succeeds (returns a scalar), use that result. If it raises InvalidIndexError or KeyError (e.g., for slice-in-tuple multi-value selections like (1, slice(1, 2))), fall back to get_locs().

Tests

Added a regression test test_sel_nested_tuple_key_collapses_dimension that verifies the dimension collapses for single nested tuple keys.

Closes #11341