In our case if we want to share some of the logic in a reducer/action-creator or something else, we write "helper" functions that handle the heavy lifting, and

In our case if we want to share some of the logic in a reducer/action-creator or something else, we write "helper" functions that handle the heavy lifting, and call them from the action-creator. This way we can easily test and reuse the actual "hard logic" for those parts, don't need to worry about trying to reuse the actual Redux stuff, and can add/modify the input or result from the helper functions slightly if needed per app. Yes, it's more boilerplate, but it's served us well.