We built this after having to manually compare new pipeline outputs with older ones too many times to find out why they are different. It is a simple package that, given two polars dataframes and a primary key, produces a nicely formatted summary of differences. It also has some methods to investigate the differences more deeply.