Sam

New functions
- `sam.models.SamQuantileMLP.quantile_feature_importances`: now has argument sum_time_components that summarizes feature importances for different features generated for a single component (i.e. in onehot encoding).

Changes
- `sam.feature_engineering.automatic_rolling_engineering`: `estimator_type` argument can now also be 'bayeslin', which should be used if one hot components are used

Bugfixes
- `sam.feature_engineering.automatic_rolling_engineering`: constant features are no longer deleted (broke one hot features)

Bugfixes
- `sam.models.SamQuantileMLP`: When using y_scaler, name of rescaled y-series is set correctly.

Changes
- `sam.models.SamQuantileMLP`: Now accepts a keyword argument `r2_callback_report` to add the new custom r2 callback.

New functions
- `sam.metrics.custom_callbacks`: Added a custom callback that computes r2 with `sam.metrics.train_mean_r2` for each epoch

Bugfixes
- `sam.validation.create_validation_pipe`: the imputation part is now correctly applied only to the `cols` columns in the df
- `sam.metrics.train_mean_r2`: now only adds non-nan values in np.arrays (previously would return nan R2)

Changes
- `sam.visualization.quantile_plot`: now accepts custom outliers with 'outlier' argument

Bugfixes
- `sam.visualization.quantile_plot`: now correctly shifts y_hat with predict_ahead

New functions
- New function: `sam.metrics.train_mean_r2` that evaluates r2 based on the train set mean
- New function: `sam.visualization.performance_evaluation_fixed_predict_ahead` that evaluates model performance with certain predict ahead.

Changes
- `sam.feature_engineering.automatic_rolling_engineering` now has new argument 'onehots'. The argument 'add_time_features' is now removed, as 'cyclicals' and 'onehots' now together make up both timefeatures