def test_recommend_method_wrong_type(self):
"""
Negative test
data: array (unsupported type)
Checks that the function raises a TypeError if the data is passed as
an array.
"""
# 1. Arrange
data = [2, 4, np.nan, 1]
# 2. Act & 3. Assert
with self.assertRaises(TypeError):
recommend_method(data)
def test_recommend_method_df_wrong_column(self):
"""
Negative test
data: Correct data frame (df_breast_cancer)
column: 'z' (not a column of df_breast_cancer)
Checks that the function raises a ValueError if the columns doesn't
exist in the data.
"""
# 1. Arrange
df = generate_df_breast_cancer()
# 2. Act & 3. Assert
with self.assertRaises(ValueError):
recommend_method(df, column='z')
def test_recommend_method_col_for_series(self):
"""
Negative test
data: Correct series (ts_cat)
column: 'a' (series can't have columns=
Checks that the function raises a ValueError if a column is passed
for a series.
"""
# 1. Arrange
ser = generate_ts_cat()
# 2. Act & 3. Assert
with self.assertRaises(ValueError):
recommend_method(ser, 'a')
def test_recommend_method_df_num(self):
"""
Positive test
data: Correct dataframe (df_sales)
The data frame contains no categorical values.
Therefore, imputation using k-NN should be recommended.
"""
# 1. Arrange
ser = generate_df_sales()
# 2. Act
method = recommend_method(ser, title_only=True)
# 3. Assert
self.assertEqual(method, 'imputation using k-NN')
def test_recommend_method_series_num_no_ts(self):
"""
Positive test
data: Correct series (example_series)
The series contains numerical values but not a datetime index.
Therefore, mean substitution should be recommended.
"""
# 1. Arrange
ser = generate_example_series()
# 2. Act
method = recommend_method(ser, title_only=True)
# 3. Assert
self.assertEqual(method, 'mean substitution')
def test_recommend_method_series_cat(self):
"""
Positive test
data: Correct series (ts_cat)
The series contains categorical values.
Therefore, random sample imputation should be recommended.
"""
# 1. Arrange
ser = generate_ts_cat()
# 2. Act
method = recommend_method(ser, title_only=True)
# 3. Assert
self.assertEqual(method, 'random sample imputation')
def test_recommend_method_df_cat(self):
"""
Positive test
data: Correct dataframe (df_breast_cancer)
The data frame contains categorical values.
Therefore, most-frequent substitution should be recommended.
"""
# 1. Arrange
ser = generate_df_breast_cancer()
# 2. Act
method = recommend_method(ser, title_only=True)
# 3. Assert
self.assertEqual(method, 'most-frequent substitution')
def test_recommend_method_df_col_cat(self):
"""
Positive test
data: Correct dataframe (df_breast_cancer)
column: 'class'
The column contains categorical values.
Therefore, logistic regression imputation should be recommended.
"""
# 1. Arrange
ser = generate_df_breast_cancer()
# 2. Act
method = recommend_method(ser, 'class', title_only=True)
# 3. Assert
self.assertEqual(method, 'logistic regression imputation')
def test_recommend_method_series_timeseries(self):
"""
Positive test
data: Correct series (ts_airgap)
The series contains numerical values and has a datetime index.
Therefore, interpolation with seasonal adjustment should be
recommended.
"""
# 1. Arrange
ser = generate_ts_airgap()
# 2. Act
method = recommend_method(ser, title_only=True)
# 3. Assert
self.assertEqual(method, 'interpolation with seasonal adjustment')
def test_recommend_method_df_process_description(self):
"""
Positive test
data: Correct dataframe (df_breast_cancer)
title_only: False
Since title_only is false, the full message should be shown, in this
case, 4 lines
"""
# 1. Arrange
ser = generate_df_breast_cancer()
# 2. Act
message = recommend_method(ser)
# 3. Assert
self.assertEqual(message.count('\n') + 1, 4)
def test_recommend_method_df_col_little_na(self):
"""
Positive test
data: Correct dataframe (example_df_divcols)
column: 'd'
The column contains numerical values and the dataframe does not have a
datetime index. Less than 10% of the data is missing.
Therefore, mean substitution should be recommended.
"""
# 1. Arrange
ser = generate_example_df_divcols()
# 2. Act
method = recommend_method(ser, 'd', title_only=True)
# 3. Assert
self.assertEqual(method, 'mean substitution')
def test_recommend_method_df_col_low_corr(self):
"""
Positive test
data: Correct dataframe (example_df_divcols)
column: 'h'
The column contains numerical values and the dataframe does not have a
datetime index. More than 10% of the data is missing. The column has
los (<= 0.8) correlation with another column.
Therefore, linear regression imputation should be recommended.
"""
# 1. Arrange
ser = generate_example_df_divcols()
# 2. Act
method = recommend_method(ser, 'h', title_only=True)
# 3. Assert
self.assertEqual(method, 'imputation using k-NN')
def test_recommend_method_df_col_ts(self):
"""
Positive test
data: Correct dataframe (example_df_ts)
column: 'airgap'
The column contains numerical values and the dataframe has a
datetime index.
Therefore, interpolation with seasonal adjustment should be
recommended.
"""
# 1. Arrange
ser = generate_example_df_ts()
# 2. Act
method = recommend_method(ser, 'airgap', title_only=True)
# 3. Assert
self.assertEqual(method, 'interpolation with seasonal adjustment')