_winsorize can (and should) generate uncertainties

[MichaelTiemannOSC]

In the _winsorize function (data_warehouse.py):

        lower=nominal_intensities.quantile(q=self.projection_controls.LOWER_PERCENTILE, axis='index', numeric_only=False)
        upper=nominal_intensities.quantile(q=self.projection_controls.UPPER_PERCENTILE, axis='index', numeric_only=False)
        winsorized: pd.DataFrame = nominal_intensities.clip(
            lower=lower,
            upper=upper,
            axis='columns'
        )
        if ITR.HAS_UNCERTAINTIES:
            # FIXME: the clipping process can properly introduce uncertainties.  The low and high values that are clipped could be                                                                      
            # replaced by the clipped values +/- the lower and upper percentile values respectively.