def test_encode_numerical_attributes():
from .testdata import adults01
attr = Attribute(adults01['age'])
assert attr.bins[0] <= 19
assert attr.bins[-1] >= 56
assert len(attr.encode()) == len(attr)
from sklearn.model_selection import train_test_split
train, test = train_test_split(adults01['age'])
assert len(attr.encode(data=train)) == len(train)
def test_encode_datetime_attributes():
from pandas import DataFrame
from .testdata import adults01
frame = DataFrame(adults01)
attr = Attribute(frame['birth'])
# assert other information
assert len(attr.encode()) == len(attr)
def test_encode_categorical_attributes():
from pandas import DataFrame
from .testdata import adults01
frame = DataFrame(adults01)
attr = Attribute(frame['education'], categorical=True)
columns = ['11th', '7th-8th', '9th', 'Assoc-acdm', 'Bachelors', 'Doctorate',
'HS-grad', 'Masters', 'Some-college']
assert array_equal(attr.bins, columns)
assert array_equal(attr.encode().columns, columns)
