a helper library for turning structured subject data into Bunch specifications for pypeline consumption

work in progress!

right now we’re trying to get it to read a specific kind of csv and output =Bunch=es based on simple logical rules