def test_clean_multi_category(self):
train_df = pd.read_csv('tests/train.csv')
test_df = pd.read_csv('tests/test.csv')
target_col = 'target'
multi_categories = [['multi_label1', 'multi_label2',
'multi_label3'], ['multi_label_independed']]
cleaned_train_df, target_series, cleaned_test_df = pb.clean(
train_df, test_df, target_col, 0.5, multi_categories)
self.assertEqual(len(cleaned_test_df.dropna()), len(test_df))
self.assertEqual(len(cleaned_train_df), len(target_series))
self.assertEqual(len(cleaned_train_df.columns),
len(cleaned_test_df.columns))
self.assertEqual(target_series.name, "y1:" + target_col)
# Check no effect for original dataframe
new_train_df = pd.read_csv('tests/train.csv')
self.assertListEqual(
new_train_df.columns.to_list(), train_df.columns.to_list())
# Update Expected CSV
# cleaned_train_df.to_csv(
# 'tests/expected_multi_category_train.csv', index=False)
# cleaned_test_df.to_csv(
# 'tests/expected_multi_category_test.csv', index=False)
expected_train_df = pd.read_csv(
'tests/expected_multi_category_train.csv')
expected_test_df = pd.read_csv(
'tests/expected_multi_category_test.csv')
self.assertListEqual(cleaned_train_df.columns.to_list(),
expected_train_df.columns.to_list())
self.assertListEqual(cleaned_test_df.columns.to_list(),
expected_test_df.columns.to_list())
def test_clean(self):
train_df = pd.read_csv('tests/train.csv')
test_df = pd.read_csv('tests/test.csv')
target_col = 'target'
cleaned_train_df, target_series, cleaned_test_df = pb.clean(
train_df, test_df, target_col, 0.5)
self.assertEqual(len(cleaned_test_df.dropna()), len(test_df))
self.assertEqual(len(cleaned_train_df), len(target_series))
self.assertEqual(len(cleaned_train_df.columns),
len(cleaned_test_df.columns))
self.assertEqual(target_series.name, "y1:" + target_col)
# Check no effect for original dataframe
new_train_df = pd.read_csv('tests/train.csv')
self.assertListEqual(
new_train_df.columns.to_list(), train_df.columns.to_list())
# Update Expected CSV
# cleaned_train_df.to_csv('tests/expected_train.csv', index=False)
# cleaned_test_df.to_csv('tests/expected_test.csv', index=False)
expected_train_df = pd.read_csv('tests/expected_train.csv')
expected_test_df = pd.read_csv('tests/expected_test.csv')
self.assertListEqual(cleaned_train_df.columns.to_list(),
expected_train_df.columns.to_list())
self.assertListEqual(cleaned_test_df.columns.to_list(),
expected_test_df.columns.to_list())
def test_clean_optuna_regressor(self):
train_df = pd.read_csv('tests/train_regression.csv')
test_df = pd.read_csv('tests/test.csv')
target_col = 'target'
cleaned_train_df, target_series, cleaned_test_df = pb.clean(
train_df, test_df, target_col)
self.assertEqual(len(cleaned_test_df.dropna()), len(test_df))
self.assertEqual(len(cleaned_train_df), len(target_series))
self.assertEqual(len(cleaned_train_df.columns),
len(cleaned_test_df.columns))
self.assertEqual(target_series.name, "y1:" + target_col)
def test_clean(self):
train_df = pd.read_csv('tests/train.csv')
test_df = pd.read_csv('tests/test.csv')
target_col = 'target'
cleaned_train_df, target_df, cleaned_test_df = pb.clean(
train_df, test_df, target_col, 0.5)
print(cleaned_train_df)
print(cleaned_test_df)
print(target_df)
self.assertEqual(len(cleaned_test_df.dropna()), len(test_df))
self.assertEqual(len(cleaned_train_df), len(target_df))
self.assertEqual(len(cleaned_train_df.columns),
len(cleaned_test_df.columns))
self.assertEqual(target_df.columns[0], "y1:" + target_col)