def test_given_categories(self):
cats = ["a", "b", "c", "d"]
ce = dpp.Categorizer(categories={"A": (cats, True)})
trn = ce.fit_transform(raw)
assert trn["A"].dtype == "category"
tm.assert_index_equal(trn["A"].cat.categories, pd.Index(cats))
assert all(trn["A"].cat.categories == cats)
assert trn["A"].cat.ordered
def test_given_categories(self):
cats = ['a', 'b', 'c', 'd']
ce = dpp.Categorizer(categories={'A': (cats, True)})
trn = ce.fit_transform(raw)
assert trn['A'].dtype == 'category'
tm.assert_index_equal(trn['A'].cat.categories, pd.Index(cats))
assert all(trn['A'].cat.categories == cats)
assert trn['A'].cat.ordered
def test_dask(self):
a = dd.from_pandas(raw, npartitions=2)
ce = dpp.Categorizer()
trn = ce.fit_transform(a)
assert is_categorical_dtype(trn['A'])
assert is_categorical_dtype(trn['B'])
assert is_categorical_dtype(trn['C'])
assert trn['D'].dtype == int
tm.assert_index_equal(ce.columns_, pd.Index(['A', 'B', 'C']))
def test_dask(self):
a = dd.from_pandas(raw, npartitions=2)
ce = dpp.Categorizer()
trn = ce.fit_transform(a)
assert is_categorical_dtype(trn["A"])
assert is_categorical_dtype(trn["B"])
assert is_categorical_dtype(trn["C"])
assert trn["D"].dtype == int
tm.assert_index_equal(ce.columns_, pd.Index(["A", "B", "C"]))
def test_ce(self):
ce = dpp.Categorizer()
original = raw.copy()
trn = ce.fit_transform(raw)
assert is_categorical_dtype(trn['A'])
assert is_categorical_dtype(trn['B'])
assert is_categorical_dtype(trn['C'])
assert trn['D'].dtype == int
tm.assert_index_equal(ce.columns_, pd.Index(['A', 'B', 'C']))
tm.assert_frame_equal(raw, original)
def test_ce(self):
ce = dpp.Categorizer()
original = raw.copy()
trn = ce.fit_transform(raw)
assert is_categorical_dtype(trn["A"])
assert is_categorical_dtype(trn["B"])
assert is_categorical_dtype(trn["C"])
assert trn["D"].dtype == int
tm.assert_index_equal(ce.columns_, pd.Index(["A", "B", "C"]))
tm.assert_frame_equal(raw, original)
def test_raises(self):
ce = dpp.Categorizer()
X = np.array([[0, 0], [1, 1]])
with pytest.raises(TypeError):
ce.fit(X)
X = da.from_array(X, chunks=(2, 2))
with pytest.raises(TypeError):
ce.fit(X)
with pytest.raises(NotFittedError):
ce.transform(raw)
def test_categorical_dtype(self):
ce = dpp.Categorizer()
ce.fit(raw)
assert (hash(ce.categories_['A']) == hash(
pd.api.types.CategoricalDtype(['a', 'b', 'c'], False)))
def test_non_categorical_dtype(self):
ce = dpp.Categorizer()
ce.fit(raw)
idx, ordered = ce.categories_['A']
tm.assert_index_equal(idx, pd.Index(['a', 'b', 'c']))
assert ordered is False
def test_columns(self):
ce = dpp.Categorizer(columns=['A'])
trn = ce.fit_transform(raw)
assert is_categorical_dtype(trn['A'])
assert is_object_dtype(trn['B'])
def test_categorical_dtype(self):
ce = dpp.Categorizer()
ce.fit(raw)
assert hash(ce.categories_["A"]) == hash(
pd.api.types.CategoricalDtype(["a", "b", "c"], False))
def test_non_categorical_dtype(self):
ce = dpp.Categorizer()
ce.fit(raw)
idx, ordered = ce.categories_["A"]
tm.assert_index_equal(idx, pd.Index(["a", "b", "c"]))
assert ordered is False
