Centralise validation of mvepiestim inputs

[sangeetabhatia03]

This PR address issue #189

• created new helpers to validate inputs to mv-epiestim functions;
• added an argument to control the validation, set to FALSE for draw_R and draw_epsilon which are called via estimate_advantage
• created a helper for normalising SI that is called at the beginning of estimate_advantage; our current handling of this was inconsistent as we normalised if the user did not specify t_min and did not normalise if t_min was provided. Further, the code for normalising was also incorrect (fixed now)

[sangeetabhatia03]

hi @annecori @steffilazerte just flagging that the lint errors on this PR have been addressed in PR#244.

[steffilazerte]

Hi @sangeetabhatia03, I reviewed as best I can, considering my unfamiliarity. Looks good to me and I think has a lot of added value in the DRY principle!

I asked a couple of questions, but mostly just for points of thought, not things that must change (for example, should you add .call = FALSE to the warnings and stops that you move? I'm not sure they'll be delinted otherwise?)

See what you think. Merge when you're ready!

[steffilazerte]

Just in case you wanted a review from me: looks good 😁