The first_day_below and first_day_above indices should be made into generics

[Zeitsperre]

I think it would be a good idea to move the two first_day indices under _temperature.py to generic.py so that we can construct better indicators out of them.

Despite their docstrings being appropriate for general temperature data, they currently use tasmin in their call signature (since our regional focus largely concerns winter phenomena). For the wrapped indicators of these, it means that a dataset with tasmin among its data_vars will have that variable used by default. This can be misleading when using the indicator and indice forms.

If we were to move these over, I would elect to deprecate first_day_above and first_day_below and propose the following replacement, under generic.py:

• first_day_above: def first_day_below(da: xarray.DataArray, thresh: str = None, before_date: DayOfYearStr = "07-01", window: int = 1, freq: str = "YS",) -> xarray.DataArray
• first_day_below: def first_day_below(da: xarray.DataArray, thresh: str = None, after_date: DayOfYearStr = "07-01", window: int = 1, freq: str = "YS",) -> xarray.DataArray

under _temperature.py:

• first_day_{tn | tg | tx}_above: def first_day_above({tasmin | tas | tasmax}: xarray.DataArray, thresh: str = None, before_date: DayOfYearStr = "07-01", window: int = 1, freq: str = "YS",) -> xarray.DataArray
• first_day_{tn | tg | tx}_below: def first_day_below({tasmin | tas | tasmax}: xarray.DataArray, thresh: str = None, after_date: DayOfYearStr = "07-01", window: int = 1, freq: str = "YS",) -> xarray.DataArray

This would mean making some deprecation warnings for the existing indices/indicators, with breaking changes coming later.