Add logging to demo scripts

This adds print statements to be able to better show
functionality of the feature store in the example

Author: accorvin

credit_model.py

@@ -101,6 +101,9 @@ def predict(self, request):
         features = request.copy()
         features.update(feature_vector)
         features_df = pd.DataFrame.from_dict(features)
+        pd.set_option('display.max_columns', 15)
+        print('Data returned from online store for request:')
+        print(features_df[['dob_ssn', 'total_debt_due', 'credit_card_due', 'mortgage_due', 'student_loan_due', 'vehicle_loan_due', 'loan_amnt']])
 
         # Apply ordinal encoding to categorical features
         self._apply_ordinal_encoding(features_df)

run.py

@@ -10,7 +10,11 @@
 
 # Train model (using Postgres for zipcode and credit history features)
 if not model.is_model_trained():
+    print('Starting model training...')
     model.train(loans)
+    print('Model has been trained successfully')
+else:
+    print('Model was already trained in a previous run')
 
 # Make online prediction (using Redis for retrieving online features)
 loan_request = {
@@ -25,6 +29,7 @@
     "loan_int_rate": [16.02],
 }
 
+print(f'Submitting a prediction to the model with data {loan_request}')
 result = model.predict(loan_request)
 
 if result == 0: