Estimation of Policy-Relevant Causal Effects in the Presence of Interference under the Difference-in-Differences Framework with an Application to the Philadelphia Beverage Tax