def test_missing_buckets(self):
np = pytest.importorskip("numpy")
one = data_types.DiscreteHistogram(
buckets=["a", "b"],
data=[0, 1],
)
two = data_types.DiscreteHistogram(
buckets=["c", "a"],
data=[1, 0],
)
d = one.dist(two)
np.isclose(d, 0)
def test_creation(self, client, strs):
key = strs[0]
attr = data_types.DiscreteHistogram(
buckets=["yes", "no"],
data=[10, 20],
)
experiment_run = client.create_experiment_run(attrs={key: attr}, )
assert experiment_run.get_attribute(key) == attr
def test_creation(self, client, strs):
key = strs[0]
attr = data_types.DiscreteHistogram(
buckets=["yes", "no"],
data=[10, 20],
)
registered_model = client.create_registered_model()
model_version = registered_model.create_version(attrs={key: attr}, )
assert model_version.get_attribute(key) == attr
def test_discrete_histogram(self):
attr = data_types.DiscreteHistogram(
buckets=["yes", "no"],
data=[10, 20],
)
assert attr._as_dict() == {
"type": "verta.discreteHistogram.v1",
"discreteHistogram": {
"buckets": ["yes", "no"],
"data": [10, 20],
},
}
def test_discrete_histogram_numpy(self):
np = pytest.importorskip("numpy")
attr = data_types.DiscreteHistogram(
buckets=np.array(["yes", "no"]),
data=np.array([10, 20]),
)
assert attr._as_dict() == {
"type": "verta.discreteHistogram.v1",
"discreteHistogram": {
"buckets": ["yes", "no"],
"data": [10, 20],
},
}
def test_summary_labels(self, client):
pytest.importorskip("scipy")
summaries = client.operations.summaries
monitored_entity = client.operations.get_or_create_monitored_entity()
summary_name = "summary_v2_{}".format(generate_default_name())
summary = summaries.create(summary_name, data_types.DiscreteHistogram,
monitored_entity)
assert isinstance(summary, Summary)
summaries_for_monitored_entity = SummaryQuery(
monitored_entities=[monitored_entity])
retrieved_summaries = summaries.find(summaries_for_monitored_entity)
assert len(retrieved_summaries) > 0
for s in retrieved_summaries:
assert isinstance(s, Summary)
now = time_utils.now()
yesterday = now - timedelta(days=1)
discrete_histogram = data_types.DiscreteHistogram(
buckets=["hotdog", "not hotdog"], data=[100, 20])
labels = {"env": "test", "color": "blue"}
summary_sample = summary.log_sample(
discrete_histogram,
labels=labels,
time_window_start=yesterday,
time_window_end=now,
)
assert isinstance(summary_sample, SummarySample)
float_histogram = data_types.FloatHistogram(
bucket_limits=[1, 13, 25, 37, 49, 61],
data=[15, 53, 91, 34, 7],
)
labels2 = {"env": "test", "color": "red"}
with pytest.raises(TypeError):
summary_sample_2 = summary.log_sample(
float_histogram,
labels=labels2,
time_window_start=yesterday,
time_window_end=now,
)
labels = client.operations.labels
retrieved_label_keys = labels.find_keys(
summary_query=summaries_for_monitored_entity)
assert len(retrieved_label_keys) > 0
if retrieved_label_keys:
retrieved_labels = labels.find_values(
summary_query=summaries_for_monitored_entity,
keys=retrieved_label_keys)
for key in retrieved_label_keys:
assert key in retrieved_labels
all_samples_for_summary = summary.find_samples()
assert len(all_samples_for_summary) == 1
blue_samples = summary.find_samples(
SummarySampleQuery(labels={"color": ["blue"]}), )
assert len(blue_samples) == 1
