- Improved user API
- Added extend method to DMVLP and WDMVLP to search for new rule when a feature state to predict from is not already matched by the model rules.
- Added BruteForce algorithm for scalability comparaison purpose with GULA
- Added explanation scoring functions in pylfit.postprocessing, evaluating a model predictions and explanation versus expected perfect model
- Extended MLJ20 evaluation with baselines and new metrics
- Regression test coverage 99% of user API accessible code