Where the shuffle arg would mean that the state construction sequence does not have to be in any specific order (False would mean first sample atoms, then lattice parameters then spacegroup, for each crystal sample)
This would enable people to split the complexity of some environments and make them more testable too
Idea for v2: having a utility class to compose environments like
Where the
shufflearg would mean that the state construction sequence does not have to be in any specific order (Falsewould mean first sample atoms, then lattice parameters then spacegroup, for each crystal sample)This would enable people to split the complexity of some environments and make them more testable too