Refactor file interpretation in CLI

[ast0815]

The CLI code is a bit of a mess right now. Would be better to refactor it so that the code that loads the data from the files in independent from the code that modifies said data. I am tempted to use pandas data frames as "universal" data container. Would make pandas a necessary dependency for the CLI, but the code would look much nicer and it would be easier to implement further features consistently.