Based on feedback from the PR from SERVER-117697

I think this is okay for now but I would like to revisit whether PrepareExecutionHelper and its instances should really be exposing internal engine choice and how to design around it. It feels like an abstraction leakage since previously, the callers wouldn't know the specifics of the engine type and could treat instances of PrepareExecutionHelper uniformly. Ideally, we are able to make the new plan ranking code templated/more engine agnostic in the future to avoid needing this. I don't think we need a ticket per-se but more of a point to bring up in future discussions around refactoring in this area of code.