def inverse_transform(self, X):
for column, lbl_params in self._convert_params.items():
if "unique_values" in lbl_params and "new_columns" in lbl_params:
# convert to one hot
lbl = LabelBinarizer()
lbl.from_json(lbl_params)
X = lbl.inverse_transform(X, column) # should raise exception
else:
# convert to integer
lbl = LabelEncoder()
lbl.from_json(lbl_params)
X.loc[:, column] = lbl.inverse_transform(X.loc[:, column])
return X
def test_inverse_transform(self):
d = {"col1": ["a", "a", "c"], "col2": ["w", "e", "d"]}
df = pd.DataFrame(data=d)
lb = LabelBinarizer()
# check first column
lb.fit(df, "col1")
bb = lb.transform(df, "col1")
self.assertTrue("col1_c" in bb.columns)
self.assertTrue(np.sum(bb["col1_c"]) == 1)
bb = lb.inverse_transform(bb)
self.assertTrue("col1_c" not in bb.columns)
# check second column
lb = LabelBinarizer()
lb.fit(df, "col2")
bb = lb.transform(df, "col2")
self.assertTrue("col2_w" in bb.columns)
self.assertTrue("col2_e" in bb.columns)
self.assertTrue("col2_d" in bb.columns)
self.assertTrue(np.sum(bb["col2_w"]) == 1)
bb = lb.inverse_transform(bb)
self.assertTrue("col2_w" not in bb.columns)