Add transformer functions between periods

[VEZY]

We should be able to convert e.g. a TimeStepTable with hourly time steps into daily time steps.

Something close to:

convert(TimeStepTable{Atmosphere{Dates.Day}}, ts)

But for this we first need to be able to process the TimeStepTable{Atmosphere}:

• identify if the time steps are daily (equal or more) or hourly (equal or more)
• if they are hourly (or so) and need daily, check if we have a full day length (sum of durations in the same day == 24 Hour) or if there are gaps, in which case we either make a gap-filling or return an error
• if they are daily, downscale to the required temporal scale.
• we also have to identify if time steps are homogeneous in the TimeStepTable{Atmosphere}, if not apply the above differently between them

Downscaling (e.g. daily time steps into hourly or less or more) is pretty straightforward for known default variables (e.g. T or P), but not for unknown variables, so we should define a way for the user to provide transformations for those using e.g. DataFrames.jl syntax. And same for upscaling (e.g. hour to daily), because we don't know if we have to sum or average unknown variables.

[VEZY]

A first take on upscaling ca n be found in to_daily.