Improve generality of PTO module; non-hydro DOFs

[cmichelenstrofer]

For more complex WECs the PTO kinematics might depend not only on the hydrodynamic DOFs but also on additional DOFs. The PTO module is not set up to handle this.

Initially we though of the PTO module as an 'example' but after some discussions we want to make it more general. Else it would be a lot of copy-paste.

Proposed solution:

• core only deals with the full x vector.
• we provide a decompose function that splits it into 4 different variables: hydrodynamic DOFs, non-hydro DOFs, PTO Force, others.
• The PTO module's kinematic matrix will depend on both the hydro and non-hydro dofs.

[ryancoe]

Note that if/when we make changes to the decision variable x to resolve this issue, we will need to make updates to implementation.rst and theory.rst.

Also, here's the diagram that Giorgio and I developed years ago when we were figuring out how to structure WecOptTool.