Revise time handling

[lkluft]

This PR:

• Improves the internal time handling by using datetime.datetime and datetime.timedelta objects.
• Adds a TimestepAdjuster which allows the user to adjust the timestep during runtime.

[lkluft]

I guess it speeds up the runs to equilibrium quite a lot!

It seems like you cannot chouse the maximum timestep too high (above 2 days is increasingly unstable). But it is really nice that you can have a 1h timestep during the adjustments, especially in high forcing scenarios.

On the other hand, I also think the option of a fixed timestep is important for other experiments (eg studying the diurnal cycle), so it's good that you've kept this.

I totally agree. That's why I kept the old default.

I would like to have a way of seeing the timestep though, at least for the case where it is changing!

To be honest, I have that in my personal branch because I want the information too :D I will also add this log message to the PR ;)

[lkluft]

Thanks for the review! If you are fine with the changes I will force-push a rebased version of the branch and merge the PR :)

[SallyDa]

Yes, looks good, thank you :)