def test_matrix_geomspace_option(self):
def assert_equal(config, v1, v2, v3):
result = {'start': v1, 'stop': v2, 'num': v3}
assert config.to_dict()['geomspace'] == result
np.testing.assert_array_equal(config.to_numpy(),
np.geomspace(**result))
assert config.length == len(np.geomspace(**result))
assert config.sample() in np.geomspace(**result)
assert config.is_categorical is False
assert config.is_distribution is False
assert config.is_range is True
assert config.is_uniform is False
assert config.is_discrete is True
assert config.is_continuous is False
assert config.min == v1
assert config.max == v2
# as list
config_dict = {
'geomspace': [1, 2, 3],
}
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, *config_dict['geomspace'])
# as string
config_dict['geomspace'] = '1:10:1'
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, 1, 10, 1)
# as dict
config_dict['geomspace'] = {'start': 1.2, 'stop': 1.8, 'num': 0.1}
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
def test_matrix_values_option(self):
config_dict = {
'values': [1, 2, 3],
}
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
assert config.to_numpy() == config_dict['values']
assert config.sample() in [1, 2, 3]
assert config.length == 3
assert config.is_categorical is False
assert config.is_distribution is False
assert config.is_range is False
assert config.is_uniform is False
assert config.is_discrete is True
assert config.is_continuous is False
assert config.min == 1
assert config.max == 3
config_dict['values'] = ['ok', 'nook']
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
assert config.to_numpy() == config_dict['values']
assert config.sample() in ['ok', 'nook']
assert config.length == 2
assert config.is_categorical is True
assert config.is_distribution is False
assert config.is_range is False
assert config.is_uniform is False
assert config.is_discrete is True
assert config.is_continuous is False
assert config.min is None
assert config.max is None
config_dict['values'] = [[1, 2], [2, 4]]
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
assert config.to_numpy() == config_dict['values']
assert config.sample() in [[1, 2], [2, 4]]
assert config.length == 2
assert config.is_categorical is True
assert config.is_distribution is False
assert config.is_range is False
assert config.is_uniform is False
assert config.is_discrete is True
assert config.is_continuous is False
assert config.min is None
assert config.max is None
def test_matrix_uniform_option(self):
def assert_equal(config, v1, v2, v3=None):
result = {'low': v1, 'high': v2}
if v3:
result['size'] = v3
assert config.to_dict()['uniform'] == result
with self.assertRaises(ValidationError):
config.to_numpy()
with self.assertRaises(ValidationError):
config.to_numpy()
with self.assertRaises(ValidationError):
config.length
assert v1 <= config.sample() <= v2
assert config.is_categorical is False
assert config.is_distribution is True
assert config.is_range is False
assert config.is_uniform is True
assert config.is_discrete is False
assert config.is_continuous is True
assert config.min == v1
assert config.max == v2
# as list
config_dict = {
'uniform': [0, 1],
}
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, *config_dict['uniform'])
# as string
config_dict['uniform'] = '0:1'
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, 0, 1)
# as dict
config_dict['uniform'] = {'low': 0, 'high': 1}
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
def test_matrix_pvalues_option(self):
config_dict = {
'pvalues': [(1, 0.1), (2, 0.3), (3, 6)],
}
with self.assertRaises(ValidationError):
MatrixConfig.from_dict(config_dict)
config_dict = {
'pvalues': [(1, 0.1), (2, 0.3), (3, 0.8)],
}
with self.assertRaises(ValidationError):
MatrixConfig.from_dict(config_dict)
config_dict = {
'pvalues': [(1, 0.1), (2, 0.3), (3, -0.6)],
}
with self.assertRaises(ValidationError):
MatrixConfig.from_dict(config_dict)
config_dict['pvalues'] = ['ok', 'nook']
with self.assertRaises(ValidationError):
MatrixConfig.from_dict(config_dict)
# Pass for correct config
config_dict = {
'pvalues': [(1, 0.1), (2, 0.1), (3, 0.8)],
}
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
with self.assertRaises(ValidationError):
config.to_numpy()
assert config.sample() in [1, 2, 3]
assert config.length == 3
assert config.is_categorical is False
assert config.is_distribution is True
assert config.is_range is False
assert config.is_uniform is False
assert config.is_discrete is True
assert config.is_continuous is False
assert config.min is None
assert config.max is None
def test_matrix_qlognormal_option(self):
def assert_equal(config, v1, v2, q, v3=None):
result = {'loc': v1, 'scale': v2, 'q': q}
if v3:
result['size'] = v3
assert config.to_dict()['qlognormal'] == result
with self.assertRaises(ValidationError):
config.to_numpy()
with self.assertRaises(ValidationError):
config.length
assert isinstance(config.sample(), float)
assert config.is_categorical is False
assert config.is_distribution is True
assert config.is_range is False
assert config.is_uniform is False
assert config.is_discrete is False
assert config.is_continuous is True
assert config.min is None
assert config.max is None
# as list
config_dict = {
'qlognormal': [0, 1, 0.1],
}
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, *config_dict['qlognormal'])
# as string
config_dict['qlognormal'] = '0:1:0.1'
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, 0, 1, 0.1)
# as dict
config_dict['qlognormal'] = {'loc': 0, 'scale': 1, 'q': 0.1}
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
def test_matrix_logspace_option(self):
def assert_equal(config, v1, v2, v3, v4=None):
result = {'start': v1, 'stop': v2, 'num': v3}
if v4:
result['base'] = v4
assert config.to_dict()['logspace'] == result
np.testing.assert_array_equal(config.to_numpy(),
np.logspace(**result))
assert config.length == len(np.logspace(**result))
assert config.sample() in np.logspace(**result)
assert config.is_categorical is False
assert config.is_distribution is False
assert config.is_range is True
assert config.is_uniform is False
assert config.is_discrete is True
assert config.is_continuous is False
assert config.min == v1
assert config.max == v2
# as list
config_dict = {
'logspace': [1, 2, 3],
}
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, *config_dict['logspace'])
# with base
config_dict['logspace'] = [1, 2, 3, 2]
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, *config_dict['logspace'])
# as string
config_dict['logspace'] = '0:10:1'
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, 0, 10, 1)
# with base
config_dict['logspace'] = '0:10:1:2'
config = MatrixConfig.from_dict(config_dict)
assert_equal(config, 0, 10, 1, 2)
# as dict
config_dict['logspace'] = {'start': 1.2, 'stop': 1.8, 'num': 0.1}
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
# with base
config_dict['logspace'] = {
'start': 1.2,
'stop': 1.8,
'num': 0.1,
'base': 2
}
config = MatrixConfig.from_dict(config_dict)
assert config.to_dict() == config_dict
def validate_bo_matrix(matrix):
if not matrix:
return None
matrix_data = {}
for key, value in six.iteritems(matrix):
if not isinstance(value, MatrixConfig):
matrix_data[key] = MatrixConfig.from_dict(value)
else:
matrix_data[key] = value
if matrix_data[key].is_distribution and not matrix_data[key].is_uniform:
raise ValidationError(
'`{}` defines a non uniform distribution, '
'and it cannot be used with bayesian optimization.'.format(
key))
return matrix_data
def test_matrix_accept_only_one_option(self):
config_dict = {'values': [1, 2, 3], 'linspace': '1:2:1'}
with self.assertRaises(ValidationError):
MatrixConfig.from_dict(config_dict)
def test_matrix_requires_at_least_one_value(self):
config_dict = {}
with self.assertRaises(ValidationError):
MatrixConfig.from_dict(config_dict)