`final_size()` should accept a vector of R0

This issue requests that `final_size()` should accept a vector of $R_0$ in the argument `r0`. This [stems from this Discussion](https://github.com/orgs/epiverse-trace/discussions/173) and parallels similar changes coming to {epidemics}.

Two return type options I can think of:
1. A single data.frame giving the mean final size estimate per demography group, and the upper and lower 95% CI for each group (similar to {cfr});
2. A single data.frame giving the final size estimate for each demography group, for each $R_0$ - it is left to the user to summarise the estimates (the option in the [code snippet in the Discussion](https://github.com/orgs/epiverse-trace/discussions/173)).

I think option 1 is neat and compact, but happy to implement (2) or something else. Thoughts @adamkucharski, @TimTaylor?