Now that we have a somewhat stable pathway for the behavior that we DO want, we should add some tests to ensure that the same numbers (MM ddEs, RMSDs, etc) come out of the analysis step. So these tests could start with a QCFractal record (the result of a QM optimization), and ensure that minimization with a particular FF returns a consistent result.
Now that we have a somewhat stable pathway for the behavior that we DO want, we should add some tests to ensure that the same numbers (MM ddEs, RMSDs, etc) come out of the analysis step. So these tests could start with a QCFractal record (the result of a QM optimization), and ensure that minimization with a particular FF returns a consistent result.