def test_f1_cv_binary():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
f1_scores = test_suite.f1_cv(3)
assert isinstance(f1_scores, list)
assert isinstance(f1_scores[0], float)
assert len(f1_scores) == 3
def test_cross_val_roc_auc_avg_binary():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
avg = 0.1
assert test_suite.cross_val_roc_auc_avg(avg)
except ValueError:
assert True
def test_cross_val_roc_auc_anomaly_detection_binary():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
tolerance = 1
assert test_suite.cross_val_roc_auc_anomaly_detection(tolerance)
except ValueError:
assert True
def test_cross_val_f1_avg_multiclass():
df, column_names, target_name, clf, _ = generate_multiclass_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
avg = 0.1
test_suite.cross_val_f1_avg(avg, average="micro")
assert True
except:
assert False
def test_cross_val_recall_avg_binary():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
avg = 0.1
test_suite.cross_val_recall_avg(avg)
assert True
except:
assert False
def test_cross_val_f1_anomaly_detection_mutliclass():
df, column_names, target_name, clf, _ = generate_multiclass_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
tolerance = 1
test_suite.cross_val_f1_anomaly_detection(tolerance, average="micro")
assert True
except:
assert False
def test_cross_val_recall_anomaly_detection_binary():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
tolerance = 1
test_suite.cross_val_recall_anomaly_detection(tolerance)
assert True
except:
assert False
def test_spread_cross_val_f1_anomaly_detection_multiclass():
df, column_names, target_name, clf, _ = generate_multiclass_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
tolerance = 1
assert test_suite.spread_cross_val_f1_anomaly_detection(tolerance)
assert test_suite.spread_cross_val_f1_anomaly_detection(tolerance,
method="median")
assert test_suite.spread_cross_val_f1_anomaly_detection(tolerance,
method="trimean")
def test_roc_auc_cv_binary():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
roc_auc_scores = test_suite.roc_auc_cv(3)
assert isinstance(roc_auc_scores, list)
assert isinstance(roc_auc_scores[0], float)
assert len(roc_auc_scores) == 3
except ValueError:
assert True
def test_run_time_stress_test():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
sample_sizes = [i for i in range(100, 1000, 100)]
max_run_times = [100 for _ in range(len(sample_sizes))]
try:
test_suite.run_time_stress_test(sample_sizes, max_run_times)
assert True
except:
assert False
def test_precision_recall_f1_multiclass():
df, column_names, target_name, clf, _ = generate_multiclass_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
classes = list(df.target.unique())
assert test_suite.classifier_testing({klass: 0.1
for klass in classes},
{klass: 0.1
for klass in classes},
{klass: 0.1
for klass in classes})
def test_run_time_stress_test():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
performance_boundary = []
for i in range(100, 1000, 100):
performance_boundary.append({"sample_size": i, "max_run_time": 100})
try:
test_suite.run_time_stress_test(performance_boundary)
assert True
except:
assert False
def test_spread_cross_val_roc_auc_anomaly_detection_binary():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
tolerance = 1
test_suite.spread_cross_val_roc_auc_anomaly_detection(tolerance)
test_suite.spread_cross_val_roc_auc_anomaly_detection(tolerance,
method="median")
test_suite.spread_cross_val_roc_auc_anomaly_detection(tolerance,
method="trimean")
assert True
except:
assert False
def test_precision_recall_f1_binary():
df, column_names, target_name, clf, _ = generate_binary_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
classes = list(df[target_name].unique())
test_suite.classifier_testing({klass: 0.1
for klass in classes},
{klass: 0.1
for klass in classes},
{klass: 0.1
for klass in classes})
assert True
except:
assert False
def test():
df = pd.read_csv("data.csv")
column_names = ["A", "B", "C"]
target_name = "target"
clf = joblib.load("model1.joblib")
test_suite = classification_tests.ClassificationTests(clf,
df,
target_name,
column_names)
classes = list(df.target.unique())
assert test_suite.classifier_testing(
{klass: 0.9 for klass in classes},
{klass: 0.9 for klass in classes},
{klass: 0.9 for klass in classes}
)
def test_spread_cross_val_precision_anomaly_detection_multiclass():
df, column_names, target_name, clf, _ = generate_multiclass_classification_data_and_models(
)
test_suite = classification_tests.ClassificationTests(
clf, df, target_name, column_names)
try:
tolerance = 1
average = "micro"
test_suite.spread_cross_val_precision_anomaly_detection(
tolerance, average=average)
test_suite.spread_cross_val_precision_anomaly_detection(
tolerance, method="median", average=average)
test_suite.spread_cross_val_precision_anomaly_detection(
tolerance, method="trimean", average=average)
assert True
except:
assert False
