Adding structure to foi_index in fit_seromodel

[ntorresd]

In #200 @ben18785 raised the following concern about how we manage FOI indexation in the package:

foi_index as used in the fit_seromodel function seems a little unsafe to me. I also wasn't sure what this means from the function documentation, "Integer vector specifying the age-groups for which force-of-infection values will be estimated. It can be specified by means of [get_foi_index]". To illustrate my confusion, I wasn't sure in your example in the above whether the first element of foi_index corresponded to the oldest-aged individuals or the youngest. How about adding structure to this input? e.g. a data frame with (year, foi_index) for time models and (age, foi_index) for age models?

This is related to #204.