Sam

New features
- `sam.visualization.quantile_plot` accepts `benchmark` parameter that plots the benchmark used to calculate the model performance

Changes
- `sam.preprocessing.sam_reshape.sam_format_to_wide` now explicitly defines the arguments when calling `pd.pivot_table`
- `sam.metrics.r2_calculation.train_r2` can now use an array as a benchmark, not only a scalar average, for r2 calculation

New features
- `sam.visualization.performance_evaluation_fixed_predict_ahead` accepts `train_avg_func` parameter that provides a function to calculate the average of the train set to use for r2 calculation (default=np.nanmean)

New functions
- Name change: `sam.metrics.train_mean_r2` -> `sam.metrics.r2_calculation` to avoid circular import errors and the file now contains multiple methods
- New function: `sam.metrics.r2_calculation.train_r2` a renamed copy of `sam.metrics.r2_calculation.train_mean_r2` as any average can now be used for r2 calculation

Changes
- `sam.metrics.train_mean_r2` is now deprecated and calls `sam.metrics.train_r2`

New features
- `sam.data_sources.read_knmi` now accepts parameter `preprocessing` to transform data to more scales.

New features
- `keras_joint_mae_tilted_loss`: to fit the median in quantile regression (use average_type='median' in SamQuantileMLP)
- `plot_feature_importances`: bar plot of feature importances (e.g. computed in SamQuantileMLP.quantile_feature_importances
- `compute_quantile_ratios`: to check the proportion of data falling beneath certain quantile

Bugfixed
- eli5 uses the sklearn.metrics.scorer module, which is gone in 0.24.0, so we need <=0.24.0
- shap does not work with tensorflow 2.4.0 so we need <=2.3.1

Bugfixed
- statsmodels is no longer a dependency (dependency introduced in version 2.0.19)