RADio-

The original RADio framework proposed rank-based divergence metrics to measure normative diversity in news recommendations. The accompanying paper showcased the behavior of these metrics on the Microsoft News Dataset and its accompanying news recommenders. To do justice to the normative underpinnings of the metrics a great deal of complicated data preprocessing was necessary, which made the code slow and hard to interpret, and therefore hard to reuse in a different application. The repository released here comprises a refactoring of the RADio code. It is stripped down to its essential components and accompanied with clear documentation, which should allow for easier understanding and adoption.