- **Enhancement**
- Now transformers can create a new column instead of replace the input
- **Bug Fix**
- Make requirements more flexible to cover the latest releases
- split_evaluator_extractor now supports eval_name parameter
- Fixed `drop_first_column` behaviour in onehot categorizer
- **New**
- Add learner to calibrate predictions based on a fairness metric
- **Documentation**
- Fixed docstrings for `reverse_time_learning_curve_splitter` and `feature_importance_backward_selection`