Add a Data class representing Fermion-to-Qubit encodings

[wavefunction91]

Per #300, issues arise when encodings are mixed in an e2e applications workflow. #301 proposes a fix to harden against mismatches near-term but does not supply a long-term solution. While we only support Jordan-Wigner, Brayvi-Kitaev and Parity for now, more general encodings are possible and should be first class-citizens in the QDK/Chemistry framework.

Acceptance criteria for this feature should address the following issues:

1. A data-class representation of general fermion-to-qubit encodings should be implemented and accessible in both the C++ and python code layers. This could either be generalizing / augmenting existing data structures (e.g. Wavefunction) or standing up new data structures, or some combination of the two.
2. Both the StatePreparation and QubitMapper should be able to consume this data structure to influence their behaviour.