nvalchemi.dynamics.FitPolicy#

class nvalchemi.dynamics.FitPolicy(*args, **kwargs)[source]#

Decide whether the batch being drawn still fits once a candidate joins it.

OrderedStructureSampler.draw() calls the policy with the atom and edge totals the drawn structures would hold with the candidate included. A policy is therefore a stateless predicate over running totals. WithinBudget bounds the totals. A policy that bounds a memory estimate, or any other quantity, is another class with the same call signature.