[BUG] Atom numbers reset in each molecule in PDB2 parser about sire HOT 2 CLOSED

I've also realised that our approach doesn't preserve atom numbering for multi-molecule systems. It's a pain, do we want to be compliant with the PDB, or should we only strive for this in single molecule systems, i.e. what will be used for parameterisation? Essentially, on fixing the issue I've found that the ordering in the PDB is now ascending, so it's useless for remapping coordinates with SOMD.

from sire.

Right, I've added an option to write the PDB using the original atom numbers from the system, which makes the SOMD re-mapping work. This is activated by passing {"use_atom_numbers" : True} as a property map. These are horrible hacks on top of horrible hacks, but will hopefully be entirely redundant when SOMD2 arrives.

from sire.