RFE Protocol uses stateA / stateB from the mappings instead of the provided ChemicalSystems

[IAlibay]

We're using SMCs from the mappings intead of the Components, which can be rather dangerous.

We need to:

1. Make sure the mapping SMCs match the ChemicalSystem Components
2. Use the ChemicalSystem SMCs

[IAlibay]

Looking at https://github.com/OpenFreeEnergy/openfe/blob/35a999907e6c3567e8173e8ef3ad11cf0b5785f4/openfe/protocols/openmm_rfe/equil_rfe_methods.py#L245C28-L245C34 it looks like at least once the Protocol runs this isn't an issue, because we check that the alchemical component exists in the mapping (so it passes an equality check).

[atravitz]

@IAlibay has this actually been resolved during the Big Refactor?