unique_all doesn't specify that indices should be a flattened index #588
Description
second element must have the field name indices and must be an array containing the indices (first occurrences) of x that result in values. The array must have the same shape as values and must have the default array index data type.
This is under specified because it isn't clear that the indices are for a flattened x, e.g.:
>>> import numpy.array_api as xp
<stdin>:1: UserWarning: The numpy.array_api submodule is still experimental. See NEP 47.
>>> x = xp.asarray([[1, 0, 1], [0, 2, 1]])
>>> xp.unique_all(x)
UniqueAllResult(values=Array([0, 1, 2], dtype=int64), indices=Array([1, 0, 4], dtype=int64), inverse_indices=Array([[1, 0, 1],
[0, 2, 1]], dtype=int64), counts=Array([2, 3, 1], dtype=int64))