In the answer, i'm not sure how do derive the density of Y given Z. Also, why are we integrating with respect to y but not to z? I know by definition that if we're searching for expectation of Y given Z then the integral would be with respect to Y. But here we have a function of both Z and Y