Amex Default Prediction: Test Data Preprocessing with Cupy

This code block preprocesses test data for the Amex Default Prediction competition. It reads chunks of the test CSV file 'test_data.csv' located at '../input/amex-default-prediction/test_data.csv'. For each chunk, it applies feature engineering using the 'feature_engineer' function (not shown in the provided code). The processed data is saved to disk as 'test_data_{k+1}' files in the 'PATH_TO_DATA' directory. Additionally, the code saves the customer indices of all test files as 'test_hashes_data' in the same directory. Finally, it cleans up the memory by deleting unnecessary variables and garbage collecting, ensuring efficient processing.