fix water template for dry system

[z-gong]

when I use loch to simulate the water absorption into a dry system, which contains no water, the water_template provided causes Exception. it seems the type of the template should be _sr.mol.Molecule, rather than _sr.molecule.Molecule

[lohedges]

Many thanks for catching this! I'll merge and backport the fix to the main branch.

I'd be interested to hear how you get on with your simulations. An issue thread for a similar setup using grand implied that it wasn't possible to hydrate an empty box. (In this case it was just a completely empty box, rather than a dry system.) Presumably favourable binding with a protein / ligand would act as a hydration site.

[z-gong]

In my case, water molecules gradually filled the cavities in the dry system.

The phenomenon described in essex-lab/grand#14 is somehow simulating the vapor-liquid equilibriation. Considering the low vapor pressure of water at ambient temperature, it's actually reasonable to have less than 1 water molecule in the box.

At 300 K, Pvap​≈3.5kPa
For a 3 nm cubic box, N = PV/kT = 0.023 molecules.