v0.0.3

0.0.3

Added

• tl.ranking: returns the weighted mean F0.1 per method used for the ranking, with mean_f01 plus one column per dataset, sorted descending.

Changed

• pl.ranking: added a metric_weights parameter to weight each metric class, a level='dataset' heatmap, and now weights classes per dataset before averaging across datasets, delegating the ranking math to tl.ranking.