working ccd imaginary time propagation

[halvarsu]

Straight from the hat.

[Schoyen]

If you remove the unused import of the RungeKutta4 integrator, black the files and import ITDCCD in coupled_cluster/__init__.py, we'll merge :)

[Schoyen]

Nice! Do a final merge with master and black, and we're there.

[halvarsu]

This has been broken by the refactoring of our time-dependent solvers. However, with the new structure, a very simple implementation of imaginary time propagation would be to have a property of the time-dependent solver, time_direction or something similar, which the __call__-method uses to decide the correct factor of i for the different terms.

I'll close this PR and open a new one (which will contain considerably less code) if ya'll agree:)

[Schoyen]

That sounds like a good idea!