# fmt: off
import numpy as np
import pytest

from ase.outputs import Properties, all_outputs


@pytest.mark.parametrize('name', list(all_outputs))
def test_print_props(name):
    print(all_outputs[name])


natoms = 7


@pytest.fixture()
def forceprop():
    dct = dict(forces=np.zeros((natoms, 3)))
    props = Properties(dct)
    return props


def test_props(forceprop):
    print(forceprop)
    assert forceprop['forces'].shape == (natoms, 3)
    assert forceprop['natoms'] == natoms


def test_props_set_twice(forceprop):
    with pytest.raises(ValueError):
        # Cannot set same thing twice.
        # Programmatically we could allow that, but we would have to
        # invalidate implicitly set quantities like natoms.
        forceprop._setvalue('forces', np.zeros((natoms, 3)))


def test_props_set_consistent(forceprop):
    forceprop._setvalue('stresses', np.zeros((natoms, 6)))
    assert forceprop['stresses'].shape == (natoms, 6)


def test_props_set_inconsistent(forceprop):
    # Setting stresses for particular number of atoms different from
    # existing number of atoms must be an error:
    with pytest.raises(ValueError):
        forceprop._setvalue('stresses', np.zeros((natoms + 2, 6)))


@pytest.fixture()
def props():
    return Properties({})


@pytest.mark.parametrize('val', [np.zeros(7), np.zeros((2, 3))])
def test_array_badvalue(props, val):
    with pytest.raises(ValueError):
        props._setvalue('stress', val)


@pytest.mark.parametrize('val', [5])
def test_array_badtype(props, val):
    with pytest.raises(TypeError):
        props._setvalue('stress', val)


@pytest.mark.parametrize(
    'val', [np.arange(6).astype(float), np.arange(6), range(6)]
)
def test_array_good(props, val):
    props._setvalue('stress', val)
    assert props['stress'].shape == (6,)
    assert props['stress'].dtype == float


@pytest.mark.parametrize('val', [np.zeros(3), range(3), np.zeros(1), 1j, None])
def test_float_badtype(props, val):
    with pytest.raises(TypeError):
        props._setvalue('energy', val)


@pytest.mark.parametrize('val', ['hello'])
def test_float_badvalue(props, val):
    with pytest.raises(ValueError):
        props._setvalue('energy', val)


@pytest.mark.parametrize('val', [4.0, 42, np.nan, '42.0'])
def test_float_good(props, val):
    props._setvalue('energy', val)