Merge pull request #122 from VikramsDataScience/churn-propensity-model

Unit test cases completed. I can now commence with the construction of a CI/CD workflow

ECommerce_Churn_Propensity_Model/tests.py

@@ -1,8 +1,7 @@
 import unittest
 from pathlib import Path
+import pandas as pd
 from .src.config import Config
-from .src.eda import read_impute_data
-from .src.preprocessing import pre_processing
 
 config = Config()
 data_path = config.data_path
@@ -17,32 +16,27 @@ class Test(unittest.TestCase):
     def test_read_impute_data(self):
         """
         Execute Test Case to determine if there are NaNs in the returned
-        DataFrame produced by the read_impute_data() function.
-        IMPORTANT NOTE: Please update the 'imputed_path' object as needed when using
-        this Test Case!
-        """        
-        read_impute_test = read_impute_data(df_path=Path(content_file), 
-                                            sheet_name=1,
-                                            float_cols=float_columns, 
-                                            categorical_cols=categorical_columns,
-                                            output_path=imputed_path)
+        DataFrame produced by the read_impute_data() function from the 'eda' module.
+        """    
+        # Load imputed data to run test case
+        read_impute_test = pd.read_csv(imputed_path)
 
-        self.assertFalse(expr=read_impute_test.isna().values.any(),
+        # Perform tests and generate pass/fail print statements
+        imputation_test = self.assertFalse(expr=read_impute_test.isna().values.any(),
                          msg=f'IMPUTATION TEST CASE FAILED. THERE ARE NANS IN THE \'{imputed_path}\' DATAFRAME!')
 
-        print(f'IMPUTATION TEST CASE PASSED! THERE ARE NO NANS IN THE \'{imputed_path}\' DATAFRAME')
+        if imputation_test is None:
+            print(f'IMPUTATION TEST CASE PASSED! THERE ARE NO NANS IN THE \'{imputed_path}\' DATAFRAME')
 
     def test_preprocessed_data(self):
         """
-        Execute Test Case to determine if the preprocessed data contains:
+        Execute Test Case to determine if the preprocessed data from the 'preprocessing' module contains:
         - NaNs: Whilst running the pre_processing() the pd.cut() function's 'right' positional arg may be set to True 
         or not specified. If it's set to either True or not specified, it can create NaNs in the resulting DataFrame. 
         The 'right' arg needs to set to False.
-        - renamed columns: The same pd.cut() function will also create a ']' closed column (which is mathematically
-        very correct!). Whilst this is mathematically sound, the XGBClassifier() hates it and will raise column name 
-        errors in the downstream 'model_training' module. So the 2nd test case will test for existence of these closed columns.
-        IMPORTANT NOTE: Please update the 'preprocessed_path' object as needed when using
-        this Test Case!
+        - renamed columns: The same pd.cut() function will also create a ']' closed column (which is syntactically correct in
+        mathematics). Whilst this is mathematically sound, the XGBClassifier() hates it and will raise column name errors in the 
+        downstream 'model_training' module. So this 2nd test case will test for existence of these closed columns.
         """
         def bracket_test(df):
             """
@@ -52,22 +46,15 @@ def bracket_test(df):
                 if '[' in col or ']' in col or '{' in col or '}' in col:
                     return True
 
-        pre_process_test = pre_processing(df_path=imputed_path,
-                                        bins=[0, 12, 24, 48, 60, 72], 
-                                        onehot_cols=onehot_categoricals, 
-                                        output_path=preprocessed_path, 
-                                        bin_cols='Tenure')
+        # Load preprocessed data to run test case
+        preprocessed_df_test = pd.read_csv(preprocessed_path)
 
         # Perform tests and generate pass/fail print statements
-        nan_test = self.assertFalse(expr=pre_process_test.isna().values.any(),
-                                    msg='THERE ARE NANS IN THE INTERVAL COLUMNS! PLEASE CHECK THE PD.CUT() \'RIGHT\' ARG TO DEBUG')
-        renamed_col_test = self.assertFalse(expr=bracket_test(pre_process_test),
-                                            msg='THE INTERVAL BRACKETS ARE INCOMPATIBLE WITH XGBCLASSIFER(). PLEASE CHECK IF THE RETURNED INTERVAL COLUMNS CONTAIN BRACKETS OTHER THAN \'(\' OR \')\'')
-
-        if nan_test is None:
-            print('NAN TEST PASSED! THERE ARE NO NANS IN THE PREPROCESSED DATAFRAME')
+        renamed_col_test = self.assertFalse(expr=bracket_test(preprocessed_df_test),
+                                            msg=f'THE INTERVAL BRACKETS IN \'{preprocessed_path}\' ARE INCOMPATIBLE WITH XGBCLASSIFER(). PLEASE CHECK IF THE RETURNED INTERVAL COLUMNS CONTAIN BRACKETS OTHER THAN \'(\' OR \')\'')
+
         if renamed_col_test is None:
-            print('BRACKET SHAPE TEST PASSED! THE RETURNED INTERVAL BRACKETS ARE OF A COMPATIBLE SHAPE WITH XGBCLASSIFER()')
+            print(f'BRACKET SHAPE TEST PASSED! THE RETURNED INTERVAL BRACKETS IN \'{preprocessed_path}\' ARE OF A COMPATIBLE SHAPE WITH XGBCLASSIFER()')
 
 if __name__ == '__main__':
     unittest.main()