Split up StateTransitionManager to facilitate use of intermediate results

[grayson-helmholz]

As indicated in visualize.ipynb of #287 the user-interface should move from the almost exclusive
use of the StateTransitionManager to a collection of functions and classes/"structs" which
expose the intermediate data that should be modifiable by the user. An example are the
QNProblemSets which currently can not be fed back to the StateTransitionManager but have to
be given to the CSPSolver.
A possible rough template for the workflow is the collection of subprograms and files in the
molecular dynamics program 'GROMACS'.
The user-interface would then be put in a seperate module in order to give a similar UX compared to
the usage of the StateTransitionManager.

See also #284