def test_multilabel_binarizer_non_integer_labels():
tuple_classes = _to_object_array([(1,), (2,), (3,)])
inputs = [
([('2', '3'), ('1',), ('1', '2')], ['1', '2', '3']),
([('b', 'c'), ('a',), ('a', 'b')], ['a', 'b', 'c']),
([((2,), (3,)), ((1,),), ((1,), (2,))], tuple_classes),
]
indicator_mat = np.array([[0, 1, 1],
[1, 0, 0],
[1, 1, 0]])
for inp, classes in inputs:
# fit_transform()
mlb = MultiLabelBinarizer()
inp = np.array(inp, dtype=object)
assert_array_equal(mlb.fit_transform(inp), indicator_mat)
assert_array_equal(mlb.classes_, classes)
indicator_mat_inv = np.array(mlb.inverse_transform(indicator_mat),
dtype=object)
assert_array_equal(indicator_mat_inv, inp)
# fit().transform()
mlb = MultiLabelBinarizer()
assert_array_equal(mlb.fit(inp).transform(inp), indicator_mat)
assert_array_equal(mlb.classes_, classes)
indicator_mat_inv = np.array(mlb.inverse_transform(indicator_mat),
dtype=object)
assert_array_equal(indicator_mat_inv, inp)
mlb = MultiLabelBinarizer()
with pytest.raises(TypeError):
mlb.fit_transform([({}), ({}, {'a': 'b'})])
def test_multilabel_binarizer_non_unique():
inp = [(1, 1, 1, 0)]
indicator_mat = np.array([[1, 1]])
mlb = MultiLabelBinarizer()
assert_array_equal(mlb.fit_transform(inp), indicator_mat)
def test_multilabel_binarizer_given_classes():
inp = [(2, 3), (1,), (1, 2)]
indicator_mat = np.array([[0, 1, 1],
[1, 0, 0],
[1, 0, 1]])
# fit_transform()
mlb = MultiLabelBinarizer(classes=[1, 3, 2])
assert_array_equal(mlb.fit_transform(inp), indicator_mat)
assert_array_equal(mlb.classes_, [1, 3, 2])
# fit().transform()
mlb = MultiLabelBinarizer(classes=[1, 3, 2])
assert_array_equal(mlb.fit(inp).transform(inp), indicator_mat)
assert_array_equal(mlb.classes_, [1, 3, 2])
# ensure works with extra class
mlb = MultiLabelBinarizer(classes=[4, 1, 3, 2])
assert_array_equal(mlb.fit_transform(inp),
np.hstack(([[0], [0], [0]], indicator_mat)))
assert_array_equal(mlb.classes_, [4, 1, 3, 2])
# ensure fit is no-op as iterable is not consumed
inp = iter(inp)
mlb = MultiLabelBinarizer(classes=[1, 3, 2])
assert_array_equal(mlb.fit(inp).transform(inp), indicator_mat)
# ensure a ValueError is thrown if given duplicate classes
err_msg = "The classes argument contains duplicate classes. Remove " \
"these duplicates before passing them to MultiLabelBinarizer."
mlb = MultiLabelBinarizer(classes=[1, 3, 2, 3])
with pytest.raises(ValueError, match=err_msg):
mlb.fit(inp)
def test_multilabel_binarizer_inverse_validation():
inp = [(1, 1, 1, 0)]
mlb = MultiLabelBinarizer()
mlb.fit_transform(inp)
# Not binary
with pytest.raises(ValueError):
mlb.inverse_transform(np.array([[1, 3]]))
# The following binary cases are fine, however
mlb.inverse_transform(np.array([[0, 0]]))
mlb.inverse_transform(np.array([[1, 1]]))
mlb.inverse_transform(np.array([[1, 0]]))
# Wrong shape
with pytest.raises(ValueError):
mlb.inverse_transform(np.array([[1]]))
with pytest.raises(ValueError):
mlb.inverse_transform(np.array([[1, 1, 1]]))
def test_multilabel_binarizer_empty_sample():
mlb = MultiLabelBinarizer()
y = [[1, 2], [1], []]
Y = np.array([[1, 1], [1, 0], [0, 0]])
assert_array_equal(mlb.fit_transform(y), Y)
def test_multilabel_binarizer_non_integer_labels():
tuple_classes = _to_object_array([(1,), (2,), (3,)])
inputs = [
([("2", "3"), ("1",), ("1", "2")], ["1", "2", "3"]),
([("b", "c"), ("a",), ("a", "b")], ["a", "b", "c"]),
([((2,), (3,)), ((1,),), ((1,), (2,))], tuple_classes),
]
indicator_mat = np.array([[0, 1, 1], [1, 0, 0], [1, 1, 0]])
for inp, classes in inputs:
# fit_transform()
mlb = MultiLabelBinarizer()
inp = np.array(inp, dtype=object)
assert_array_equal(mlb.fit_transform(inp), indicator_mat)
assert_array_equal(mlb.classes_, classes)
indicator_mat_inv = np.array(mlb.inverse_transform(indicator_mat), dtype=object)
assert_array_equal(indicator_mat_inv, inp)
# fit().transform()
mlb = MultiLabelBinarizer()
assert_array_equal(mlb.fit(inp).transform(inp), indicator_mat)
assert_array_equal(mlb.classes_, classes)
indicator_mat_inv = np.array(mlb.inverse_transform(indicator_mat), dtype=object)
assert_array_equal(indicator_mat_inv, inp)
mlb = MultiLabelBinarizer()
with pytest.raises(TypeError):
mlb.fit_transform([({}), ({}, {"a": "b"})])
RM = pd.read_csv("../input_data/labeled_data_RM.csv")
GK = pd.read_csv("../input_data/labeled_data_GK.csv")
truth['labelVecs'] = ''
truth['labels'] = truth['Clusters'].str.lower()
truth['labels'] = truth['labels'].str.rstrip(';')
truth['labels'] = truth['labels'].str.split(';')
data = [prepareData(d) for d in [torsten, Mike, CK, RM, GK]]
data.append(truth)
clusters = set()
for d in data:
clusters.update([item for sublist in d['labels'] for item in sublist])
enc = MultiLabelBinarizer()
enc.fit([clusters])
for result in data:
for i, doc in result.iterrows():
doc['labelVecs'] = enc.transform([doc['labels']])[0]
doc['labelVecs'] = doc['labelsVecs'].str.wrap(500)
pd.DataFrame(enc.classes_, columns=['Label']).to_csv('labels.csv', index=False)
truth.to_csv('truth.csv', index=False)
torsten.to_csv('labeler1.csv', index=False)
Mike.to_csv('labeler2.csv', index=False)
CK.to_csv('labeler3.csv', index=False)
RM.to_csv('labeler4.csv', index=False)
GK.to_csv('labeler5.csv', index=False)
