def test_tabulardrift(tabulardrift_params):
n_cat, categories_per_feature, correction, \
update_x_ref, preprocess_x_ref, new_categories = tabulardrift_params
np.random.seed(0)
# add categorical variables
x_ref = np.random.randn(n * n_features).reshape(n, n_features).astype(
np.float32)
if n_cat > 0:
cat_cols = np.random.choice(n_features, size=n_cat, replace=False)
x_ref[:, cat_cols] = np.tile(
np.array([np.arange(n_categories)] * n_cat).T,
(n // n_categories, 1))
if categories_per_feature is None:
categories_per_feature = {f: None for f in cat_cols}
elif categories_per_feature == int:
categories_per_feature = {f: n_categories for f in cat_cols}
elif categories_per_feature == list:
categories_per_feature = {
f: list(np.arange(n_categories))
for f in cat_cols
}
else:
categories_per_feature = None
cd = TabularDrift(
x_ref=x_ref,
p_val=.05,
categories_per_feature=categories_per_feature,
preprocess_x_ref=preprocess_x_ref,
update_x_ref=update_x_ref,
correction=correction,
)
x = x_ref.copy()
if new_categories and n_cat > 0:
x[:, cat_cols] = x[:, cat_cols] + 1
preds_batch = cd.predict(x, drift_type='batch', return_p_val=True)
if new_categories and n_cat > 0:
assert preds_batch['data']['is_drift'] == 1
else:
assert preds_batch['data']['is_drift'] == 0
k = list(update_x_ref.keys())[0]
assert cd.n == x.shape[0] + x_ref.shape[0]
assert cd.x_ref.shape[0] == min(update_x_ref[k],
x.shape[0] + x_ref.shape[0])
assert preds_batch['data']['distance'].min() >= 0.
if correction == 'bonferroni':
assert preds_batch['data']['threshold'] == cd.p_val / cd.n_features
preds_feature = cd.predict(x, drift_type='feature', return_p_val=True)
assert preds_feature['data']['is_drift'].shape[0] == cd.n_features
preds_by_feature = (preds_feature['data']['p_val'] < cd.p_val).astype(int)
assert (preds_feature['data']['is_drift'] == preds_by_feature).all()
assert preds_feature['data']['threshold'] == cd.p_val
def test_tabulardrift(tabulardrift_params):
n_cat, preprocess, correction, update_X_ref, preprocess_X_ref = tabulardrift_params
n_infer = 2
preprocess_fn, preprocess_kwargs = preprocess
if isinstance(preprocess_fn, Callable):
raise NotImplementedError
else:
preprocess_fn, preprocess_kwargs = None, None
# add categorical variables
cat_cols = np.random.choice(n_features, size=n_cat, replace=False)
X_ref = np.random.randn(n * n_features).reshape(
n, n_features).astype('float32')
if n_cat > 0:
X_ref[:, cat_cols] = np.tile(
np.array([np.arange(n / n_categories)] * n_cat).T,
(n_categories, 1))
cd = TabularDrift(p_val=.05,
X_ref=X_ref,
preprocess_X_ref=preprocess_X_ref,
update_X_ref=update_X_ref,
preprocess_fn=preprocess_fn,
preprocess_kwargs=preprocess_kwargs,
correction=correction,
n_infer=n_infer)
X = X_ref.copy()
preds_batch = cd.predict(X, drift_type='batch', return_p_val=True)
assert preds_batch['data']['is_drift'] == 0
k = list(update_X_ref.keys())[0]
assert cd.n == X.shape[0] + X_ref.shape[0]
assert cd.X_ref.shape[0] == min(update_X_ref[k],
X.shape[0] + X_ref.shape[0])
assert preds_batch['data']['distance'].min() >= 0.
if correction == 'bonferroni':
assert preds_batch['data']['threshold'] == cd.p_val / cd.n_features
preds_feature = cd.predict(X, drift_type='feature', return_p_val=True)
assert preds_feature['data']['is_drift'].shape[0] == cd.n_features
preds_by_feature = (preds_feature['data']['p_val'] < cd.p_val).astype(int)
assert (preds_feature['data']['is_drift'] == preds_by_feature).all()
assert preds_feature['data']['threshold'] == cd.p_val
n_ref = 25
n_test = 25
#print(X.values)
X = X.values
X_ref, X_t0, X_t1 = X[:n_ref], X[n_ref:n_ref +
n_test], X[n_ref + n_test:n_ref + 2 * n_test]
X_ref.shape, X_t0.shape, X_t1.shape
print(X_ref.shape, X_t0.shape, X_t1.shape)
categories_per_feature = {0: None, 1: None, 2: None, 3: None}
cd = TabularDrift(X_ref,
p_val=.05,
categories_per_feature=categories_per_feature)
# filepath = 'my_path' # change to directory where detector is saved
# save_detector(cd, filepath)
# cd = load_detector(filepath)
#print(X_t0)
preds = cd.predict(X_t0)
labels = ['No!', 'Yes!']
print('Drift? {}'.format(labels[preds['data']['is_drift']]))
feature_names = [
'SepalLengthCm', 'SepalWidthCm', 'PetalLengthCm', 'PetalWidthCm'
]
#print(feature_names)
threshold=threshold,
threshold_net=threshold_net,
latent_dim=latent_dim),
KSDrift(X_ref,
p_val=p_val,
preprocess_x_ref=False,
preprocess_fn=preprocess_fn),
MMDDrift(X_ref,
p_val=p_val,
preprocess_x_ref=False,
preprocess_fn=preprocess_fn,
configure_kernel_from_x_ref=True,
n_permutations=n_permutations),
ChiSquareDrift(X_ref_cat, p_val=p_val, preprocess_x_ref=True),
TabularDrift(X_ref_mix,
p_val=p_val,
categories_per_feature={0: None},
preprocess_x_ref=True),
ClassifierDrift(X_ref,
model=model,
p_val=p_val,
n_folds=n_folds_drift,
train_size=None)
]
n_tests = len(detector)
@pytest.fixture
def select_detector(request):
return detector[request.param]