AMSWorker Support for ChemicalSystem

[dormrod]

Description

This PR adds support for passing scm.base.ChemicalSystem to AMSWorker, alongside PLAMS Molecule, and adds AMSWorkerResults.get_input_system() / get_main_system() in line with AMSResults.

Changes

Main updates:

• Added ChemicalSystem handling in AMSWorker._prepare_system() via type branching (Molecule vs ChemicalSystem), including serialization of coordinates, charge, lattice, bonds, and atomicInfo
• Added lazy initialization in AMSWorkerResults for _input_molecule and _input_system, with conversion between representations using same logic as in scm.utils (not used to avoid dependency)
• Added AMSWorkerResults.get_input_system() and AMSWorkerResults.get_main_system()
• Existing AMSWorkerResults.get_input_molecule and get_main_molecule now always return copies
• Added unit tests covering the new ChemicalSystem pathways and parity with existing Molecule behaviour.
• Integration test to be added in AMSHOME

Points for Comment

• atomicInfo extraction for ChemicalSystem was a bit tricky to implement - is this the best and most reliable way of doing it?
• Kept get_main_ase_atoms() based on molecule conversion (toASE(get_main_molecule())), not ChemicalSystem.to_ase_atoms()
• Probably for after the release to be merged into trunk

[robertrueger]

Check the cs.atom_EOL_string() method. I builds exactly this string of regions + atomic properties that you need here.

[dormrod]

Great thanks, that works perfectly and is much simpler