def test_QM_expected_fail():
'''input not of type Distribtion'''
parameters_dict = {
'a': Normal(0., 1., name='a'),
'b': 1. # input should be a Distribution
}
with pytest.raises(ValueError):
QM(parameters_dict)
def test_lognormal_code():
mu = Normal(1.0, 1.2, name='mu')
sigma = Lognormal(2.1, 2.2, name='sigma')
model = LognormalQM(mu, sigma)
code = model.code
with open(FILLED_TEMPLATES_PATH / 'os_lognormal.stan') as f:
code_hard_coded = f.read()
assert code == code_hard_coded
def test_normal_code():
mu = Normal(0., 1., name='mu')
sigma = Gamma(1., 1.2, name='sigma')
model = NormalQM(mu, sigma)
code = model.code
with open(FILLED_TEMPLATES_PATH / 'os_normal.stan') as f:
code_hard_coded = f.read()
assert code == code_hard_coded
def test_lognormal_check_domain_expected_fail():
mu = Normal(1.0, 1.2, name='mu')
sigma = Lognormal(2.1, 2.2, name='sigma')
model = LognormalQM(mu, sigma)
N, q, X = 1000, [0.25, 0.5, 0.75], [-0.1, 1.0, 1.4] #-0.1 is invalid
with pytest.raises(ValueError):
model.sampling(N, q, X)
with pytest.raises(ValueError):
model.optimizing(N, q, X)
def test_normalQM_print():
mu = Normal(0., 1., name='mu')
sigma = Gamma(1., 1., name='sigma')
model = NormalQM(mu, sigma)
assert str(
model
) == 'NormalQM(Normal(mu=0.0, sigma=1.0, name="mu"), Gamma(alpha=1.0, beta=1.0, name="sigma"))'
assert repr(
model
) == 'NormalQM(Normal(mu=0.0, sigma=1.0, name="mu"), Gamma(alpha=1.0, beta=1.0, name="sigma"))'
def test_normalQM_template_replacements():
mu = Normal(0., 1., name='loc')
sigma = Gamma(1., 1., name='scale')
model = NormalQM(mu, sigma)
replacements = model._template_replacements()
assert replacements['parametersnames'] == 'loc, scale'
assert replacements['parameters'] == 'real loc;\n real<lower=0> scale;'
assert replacements['priors'] == \
'loc ~ normal(0.0, 1.0);\n scale ~ gamma(1.0, 1.0);'
assert replacements['cdf'] == 'normal_cdf'
assert replacements['lpdf'] == 'normal_lpdf'
assert replacements['rng'] == 'normal_rng'
def lognormal_compiled_model():
mu = Normal(1., 1., name='mu')
sigma = Lognormal(1., 1., name='sigma')
model = LognormalQM(mu, sigma)
model.compile()
return model
def normal_compiled_model():
mu = Normal(0., 1., name='mu')
sigma = Gamma(1., 1.2, name='sigma')
model = NormalQM(mu, sigma)
model.compile()
return model