Consider allowing reinfections in models

[pratikunterwegs]

This discussion is taken from previous issue #227 and comes from the {epidemics}-derived package {noromod} co-developed by @bolthikaru. The aim would be to allow {epidemics} models to be used for long-term scenario modelling for time-scales where waning immunity and re-infections become important.

Some changes needed to affected models:

1. Allow immunity waning and $R \rightarrow S, V \rightarrow S$ transitions;
2. Vacamole has two vaccination compartments: decide how to handle multiple levels of protection. Waning immunity could be represented as a transition from $V_2 \rightarrow V_1 \rightarrow S$.

For helper functions:

• The epidemic_size() and new_infections() functions would have to be replaced with explicit tracking of new infections (which can be summed for epidemic size). Currently functions assume no reinfections.

Fully appreciate that there are very many combinations of infection and vaccination provided immunity that make for a challenging immunity/susceptibility landscape, so I'm happy to hear feedback on this one.