def test_can_save_and_load(tmpdir) -> None:
PROTO_FILE_NAME = str(tmpdir.join("test.proto"))
JSON_FILE_NAME = str(tmpdir.join("test.json"))
DOT_FILE_NAME = str(tmpdir.join("test.dot"))
gamma = Gamma(1.0, 1.0)
gamma.set_value(2.5)
# %%SNIPPET_START%% PythonSaveSnippet
net = BayesNet(gamma.iter_connected_graph())
metadata = {"Author": "Documentation Team"}
protobuf_saver = ProtobufSaver(net)
protobuf_saver.save(PROTO_FILE_NAME, True, metadata)
json_saver = JsonSaver(net)
json_saver.save(JSON_FILE_NAME, True, metadata)
dot_saver = DotSaver(net)
dot_saver.save(DOT_FILE_NAME, True, metadata)
# %%SNIPPET_END%% PythonSaveSnippet
# %%SNIPPET_START%% PythonLoadSnippet
protobuf_loader = ProtobufLoader()
new_net_from_proto = protobuf_loader.load(PROTO_FILE_NAME)
json_loader = JsonLoader()
new_net_from_json = json_loader.load(JSON_FILE_NAME)
def test_sample_throws_if_vertices_in_sample_from_are_missing_labels() -> None:
sigma = Gamma(1., 1)
vertex = Gaussian(0., sigma, label="gaussian")
assert sigma.get_label() is None
net = BayesNet([sigma, vertex])
with pytest.raises(ValueError,
match=r"Vertices in sample_from must be labelled."):
samples = sample(net=net, sample_from=net.iter_latent_vertices())
def test_probe_for_non_zero_probability_from_bayes_net() -> None:
gamma = Gamma(1., 1.)
poisson = Poisson(gamma)
net = BayesNet([poisson, gamma])
assert not gamma.has_value()
assert not poisson.has_value()
net.probe_for_non_zero_probability(100, KeanuRandom())
assert gamma.has_value()
assert poisson.has_value()
def net() -> BayesNet:
with Model() as m:
m.gamma = Gamma(1., 1.)
m.exp = Exponential(1.)
m.cauchy = Cauchy(m.gamma, m.exp)
return m.to_bayes_net()
def tensor_net() -> BayesNet:
with Model() as m:
m.gamma = Gamma(
np.array([1., 1., 1., 1.]).reshape((2, 2)),
np.array([2., 2., 2., 2.]).reshape((2, 2)))
m.exp = Exponential(np.array([1., 1., 1., 1.]).reshape((2, 2)))
m.add = m.gamma + m.exp
return m.to_bayes_net()
def test_can_get_vertices_from_bayes_net(get_method: str, latent: bool,
observed: bool, continuous: bool,
discrete: bool) -> None:
gamma = Gamma(1., 1.)
gamma.observe(0.5)
poisson = Poisson(gamma)
cauchy = Cauchy(gamma, 1.)
assert gamma.is_observed()
assert not poisson.is_observed()
assert not cauchy.is_observed()
net = BayesNet([gamma, poisson, cauchy])
vertex_ids = [vertex.get_id() for vertex in getattr(net, get_method)()]
if observed and continuous:
assert gamma.get_id() in vertex_ids
if latent and discrete:
assert poisson.get_id() in vertex_ids
if latent and continuous:
assert cauchy.get_id() in vertex_ids
assert len(vertex_ids) == (observed and continuous) + (
latent and discrete) + (latent and continuous)
def test_to_bayes_net_excludes_non_vertices() -> None:
with Model() as m:
m.not_a_vertex = 1
m.vertex = Gamma(0.5, 0.1)
net = m.to_bayes_net()
assert isinstance(net, BayesNet)
net_vertex_ids = [vertex.get_id() for vertex in net.iter_latent_or_observed_vertices()]
assert len(net_vertex_ids) == 1
assert m.vertex.get_id() in net_vertex_ids
def test_can_save_and_load(tmpdir) -> None:
PROTO_FILE = str(tmpdir.join("test.proto"))
JSON_FILE = str(tmpdir.join("test.json"))
DOT_FILE = str(tmpdir.join("test.dot"))
gamma = Gamma(1.0, 1.0)
gamma.set_value(2.5)
net = BayesNet(gamma.get_connected_graph())
metadata = {"Team": "GraphOS"}
protobuf_saver = ProtobufSaver(net)
protobuf_saver.save(PROTO_FILE, True, metadata)
json_saver = JsonSaver(net)
json_saver.save(JSON_FILE, True, metadata)
dot_saver = DotSaver(net)
dot_saver.save(DOT_FILE, True, metadata)
check_dot_file(DOT_FILE)
protobuf_loader = ProtobufLoader()
json_loader = JsonLoader()
new_net_from_proto = protobuf_loader.load(PROTO_FILE)
check_loaded_net(new_net_from_proto)
new_net_from_json = json_loader.load(JSON_FILE)
check_loaded_net(new_net_from_json)
def test_to_bayes_net() -> None:
with Model() as m:
m.mu = Exponential(1.)
m.sigma = Gamma(0.5, 0.1)
m.gaussian = Gaussian(m.mu, m.sigma)
net = m.to_bayes_net()
assert isinstance(net, BayesNet)
net_vertex_ids = [
vertex.get_id() for vertex in net.get_latent_or_observed_vertices()
]
assert len(net_vertex_ids) == 3
assert m.mu.get_id() in net_vertex_ids
assert m.sigma.get_id() in net_vertex_ids
assert m.gaussian.get_id() in net_vertex_ids
def net() -> BayesNet:
with Model() as m:
m.gamma = Gamma(1., 1.)
m.gaussian = Gaussian(0., m.gamma)
return m.to_bayes_net()
def net() -> BayesNet:
gamma = Gamma(1., 1.)
exp = Exponential(1.)
cauchy = Cauchy(gamma, exp)
return BayesNet(cauchy.get_connected_graph())
