def test_random_row_idxes_correct_size_fraction(row_lengths):
for row_length in row_lengths():
for fraction in [0.5, 0.1, 1 / np.sqrt(2)]:
row_idxes = _subsampling._get_random_row_idxes(
row_length, fraction, np.random
)
assert len(row_idxes) == int(row_length) * fraction
def test_random_row_idxes_fails_at_negative_input(row_lengths):
for row_length in row_lengths():
with pytest.raises(ValueError):
row_idxes = _subsampling._get_random_row_idxes(
row_length, -0.1, np.random)
with pytest.raises(ValueError):
row_idxes = _subsampling._get_random_row_idxes(
row_length, -1, np.random)
with pytest.raises(ValueError):
row_idxes = _subsampling._get_random_row_idxes(
row_length, 0, np.random)
with pytest.raises(ValueError):
row_idxes = _subsampling._get_random_row_idxes(
row_length, row_length + 1, np.random)
with pytest.raises(ValueError):
row_idxes = _subsampling._get_random_row_idxes(
row_length, "invalid", np.random)
def test_random_row_idxes_correct_size_integer(row_lengths):
for row_length in row_lengths():
for num in [5, 10, 1000]:
if num > row_length:
continue
row_idxes = _subsampling._get_random_row_idxes(
row_length, num, np.random)
assert len(row_idxes) == num
def test_random_row_idxes_unique(row_lengths):
for row_length in row_lengths():
row_idxes = _subsampling._get_random_row_idxes(row_length, "sqrt",
np.random)
assert len(row_idxes) == len(set(row_idxes))
def test_random_row_idxes_sqrt(row_lengths):
for row_length in row_lengths():
row_idxes = _subsampling._get_random_row_idxes(row_length, "sqrt",
np.random)
assert len(row_idxes) == int(np.sqrt(row_length))