Fixes - `ConstantTimeseriesRegressor` now fills nan values in input data with zero before calling `preprocess_fit` in order to successfully (by)pass validation from `BaseTimeseriesRegressor`. Besides scikit-learn compatibility, the input data is not actually used when fitting.
3.1.8
Changes - `ConstantTimeseriesRegressor` no longer checks dtypes of input data, nor nan/inf values, as the input is only used to determine the shape of the predictions.
3.1.7
Changes - Updated `BaseTimeseriesRegressor.get_feature_names_out()` so, in case of the feature engineer is a `Pipeline`, it returns the names from the last `ColumnTransformer`, if available
3.1.6
Changes - Updated wrong types in `quantile_plot.py`
3.1.5
Changes - Properly included datasets so `load_rainbow_beach()` and `load_sewage_data()` work
3.1.4
Changes - Fixed a bug where `data_sources.weather` was not installed.