def test_init_param_cost_matrix(self):
qs = VOI(cost_matrix=np.ones((2, 3)))
self.assertRaises(ValueError, qs.query, **self.kwargs)
qs = VOI(cost_matrix='string')
self.assertRaises(ValueError, qs.query, **self.kwargs)
qs = VOI(cost_matrix=np.ones((3, 3)))
self.assertRaises(ValueError, qs.query, **self.kwargs)
def test_query_param_X(self):
qs = VOI()
self.assertRaises(ValueError,
qs.query,
X_cand=self.X_cand,
clf=self.clf,
X=None,
y=self.y)
self.assertRaises(ValueError,
qs.query,
X_cand=self.X_cand,
clf=self.clf,
X='string',
y=self.y)
self.assertRaises(ValueError,
qs.query,
X_cand=self.X_cand,
clf=self.clf,
X=[],
y=self.y)
self.assertRaises(ValueError,
qs.query,
X_cand=self.X_cand,
clf=self.clf,
X=self.X[0:-1],
y=self.y)
def test_query_param_y(self):
qs = VOI()
self.assertRaises(TypeError,
qs.query,
X_cand=self.X_cand,
clf=self.clf,
X=self.X,
y=None)
self.assertRaises(ValueError,
qs.query,
X_cand=self.X_cand,
clf=self.clf,
X=self.X,
y='string')
self.assertRaises(ValueError,
qs.query,
X_cand=self.X_cand,
clf=self.clf,
X=self.X,
y=[])
self.assertRaises(ValueError,
qs.query,
X_cand=self.X_cand,
clf=self.clf,
X=self.X,
y=self.y[0:-1])
def test_query_param_clf(self):
qs = VOI()
self.assertRaises(TypeError,
qs.query,
X_cand=self.X_cand,
clf=GaussianProcessClassifier(),
X=self.X,
y=self.y)
def test_init_param_ignore_partial_fit(self):
qs = VOI(ignore_partial_fit=None)
self.assertRaises(TypeError,
qs.query,
X_cand=self.X_cand,
X=self.X,
y=self.y,
clf=self.clf)
def test_init_param_labeling_cost(self):
#qs = VOI(labeling_cost=None)
#self.assertRaises(ValueError, qs.query, **self.kwargs)
labeling_cost = np.ones((len(self.X_cand) - 1, 1))
qs = VOI(labeling_cost=labeling_cost)
self.assertRaises(ValueError, qs.query, **self.kwargs)
labeling_cost = np.ones((1, len(self.classes) - 1))
qs = VOI(labeling_cost=labeling_cost)
self.assertRaises(ValueError, qs.query, **self.kwargs)
labeling_cost = np.ones((len(self.X_cand) - 1, len(self.classes) - 1))
qs = VOI(labeling_cost=labeling_cost)
self.assertRaises(ValueError, qs.query, **self.kwargs)
qs = VOI(labeling_cost='string')
self.assertRaises(ValueError, qs.query, **self.kwargs)
def test_query_param_sample_weight_cand(self):
qs = VOI()
sample_weight_cand_list = [
'string', self.X_cand, self.X_cand[:-2],
np.empty((len(self.X_cand) - 1)),
np.empty((len(self.X_cand) + 1))
]
for sample_weight in sample_weight_cand_list:
self.assertRaises(ValueError,
qs.query,
**self.kwargs,
sample_weight=np.ones(shape=len(self.X)),
sample_weight_cand=sample_weight)
def test_query(self):
classes = [0, 1]
X_cand = np.array([[8, 1], [9, 1]])
X = np.array([[1, 2], [5, 8], [8, 4], [5, 4]])
y = np.array([MISSING_LABEL, 0, 1, MISSING_LABEL])
cost_matrix = 1 - np.eye(2)
clf_partial = SklearnClassifier(GaussianNB(),
classes=classes).fit(X, y)
clf = SklearnClassifier(estimator=GaussianProcessClassifier(),
random_state=self.random_state,
classes=classes)
qs = VOI()
# return_utilities
L = list(qs.query(**self.kwargs, return_utilities=True))
self.assertTrue(len(L) == 2)
L = list(qs.query(**self.kwargs, return_utilities=False))
self.assertTrue(len(L) == 1)
# batch_size
bs = 2
best_idx = qs.query(**self.kwargs, batch_size=bs)
self.assertEqual(bs, len(best_idx))
# query
qs.query(X_cand=X_cand, clf=clf_partial, X=X, y=y)
qs = VOI()
qs.query(X_cand=X_cand, clf=clf_partial, X=X, y=y)
class DummyClf(SkactivemlClassifier):
def fit(self, X, y, sample_weight=None):
self.classes_ = np.unique(y[labeled_indices(y)])
return self
def predict_proba(self, X):
return np.full(shape=(len(X), len(self.classes_)),
fill_value=0.5)
labeling_cost = 2.345
qs = VOI(cost_matrix=cost_matrix, labeling_cost=labeling_cost)
idxs, utils = qs.query(X_cand=X_cand,
clf=DummyClf(),
X=X,
y=y,
return_utilities=True)
np.testing.assert_array_equal(utils[0],
[-labeling_cost, -labeling_cost])
labeling_cost = np.array([2.346, 6.234])
qs = VOI(cost_matrix=cost_matrix, labeling_cost=labeling_cost)
idxs, utils = qs.query(X_cand=X_cand,
clf=DummyClf(),
X=X,
y=y,
return_utilities=True)
np.testing.assert_array_equal(utils[0], -labeling_cost)
labeling_cost = np.array([[2.346, 6.234]])
expected = [-labeling_cost.mean(), -labeling_cost.mean()]
qs = VOI(cost_matrix=cost_matrix, labeling_cost=labeling_cost)
idxs, utils = qs.query(X_cand=X_cand,
clf=DummyClf(),
X=X,
y=y,
return_utilities=True)
np.testing.assert_array_equal(utils[0], expected)
labeling_cost = np.array([[2.346, 6.234], [3.876, 3.568]])
expected = -labeling_cost.mean(axis=1)
qs = VOI(cost_matrix=cost_matrix, labeling_cost=labeling_cost)
idxs, utils = qs.query(X_cand=X_cand,
clf=DummyClf(),
X=X,
y=y,
return_utilities=True)
np.testing.assert_array_equal(utils[0], expected)