This is an explicit CISD calculation, which produced the same result as an analogous RAS calculation.
   -> Therefore it looks like the phase adjustment in read_civfl.py is correct

The results are also consistent with the cioverlap code by J. Pittner
