Allow EfficientDepthOneEvaluator to receive Paulis that are constants #58
Description
When calling the EfficientDepthOneEvaluator with a SparsePauliOp, if it includes terms that are constant (i.e. Paulis of the sort "IIIII" with some coeff), the evaluator raises an error. It would be beneficial if it could just add the constant term to the expected value computed, instead of throwing an error.
I ran into this when running RQAOA, where in some cases eliminating vertices creates new constant terms in the hamiltonian, and in that case the EfficientDepthOneEvaluator throws an error.