Test failures with numpy 1.14.0

[FRidh]

________________________________ test_validate _________________________________

    def test_validate():
        assert validate(int, 1)
>       assert validate('int', 1)
E       AssertionError: assert False
E        +  where False = validate('int', 1)

datashape/tests/test_user.py:8: AssertionError
___________________________ test_nested_iteratables ____________________________

    def test_nested_iteratables():
>       assert validate('2 * 3 * int', [(1, 2, 3), (4, 5, 6)])
E       AssertionError: assert False
E        +  where False = validate('2 * 3 * int', [(1, 2, 3), (4, 5, 6)])

datashape/tests/test_user.py:19: AssertionError
_____________________________ test_validate_dicts ______________________________

    def test_validate_dicts():
>       assert validate('{x: int, y: int}', {'x': 1, 'y': 2})
E       AssertionError: assert False
E        +  where False = validate('{x: int, y: int}', {'x': 1, 'y': 2})

datashape/tests/test_user.py:26: AssertionError
________________________ test_tuples_can_be_records_too ________________________

    def test_tuples_can_be_records_too():
>       assert validate('{x: int, y: real}', (1, 2.0))
E       AssertionError: assert False
E        +  where False = validate('{x: int, y: real}', (1, 2.0))

datashape/tests/test_user.py:36: AssertionError

• datashape 0.5.4
• Python 2.7 and 3.6

[mingwandroid]

Does anyone know anything about this failure?

[kglowinski]

It has to do with changes in numpy in 1.14.0+: https://github.com/numpy/numpy/blob/master/doc/release/1.14.0-notes.rst#future-changes

In particular:

np.issubdtype will stop downcasting dtype-like arguments. It might be expected that issubdtype(np.float32, 'float64') and issubdtype(np.float32, np.float64) mean the same thing - however, there was an undocumented special case that translated the former into issubdtype(np.float32, np.floating), giving the surprising result of True.

This means that all the validate tests are now just meaningless for what they were testing with 'int'.