Better to say the vast majority of them are later viral shedding or very minor, probably not contagious cases.

There's probably been a few reinfections, and some decent documentation of a couple. If you repeat an experiment millions of times on diverse individuals, you can expect some variation.