Does imblearn make_pipeline() exclude SMOTE oversampling on test set?

Question

I am currently running multiple classification models for a binary data problem. I know that for good practice we should only oversample training set and leave test set alon