How to create all interaction variables between 2 groups of features with sklearn transformers?

Question

Say I have a DataFrame df with 5 columns, [\'a\', \'b\', \'c\', \'d\', \'e\'].

What\'s the most natural way, using scikit-learn transformers