Including distribution of Latency & Infection period about covid19_scenarios

[nnoll]

This is a great idea. In reality, we aren't simulating individual people so we don't have a clean way to assign actual recovery/latency times. In its current form, the model could only capture such effects by sampling the deterministic rates we currently use by a distribution that's not Poissonian.

The code as written can easily accomodate stochastic sampling. The evolve function accepts a sampling function that you can program to do just this. If you (or anyone reading this) want to give a go at an input widget, I'll hook it up to the model in the branch and we can play around.

However, I personally think we should hold off on such a feature until we have a more detailed model that can directly plug into such an input.

[rneher]

We have implemented non-exponential distributions in #182
Custom distributions would require agent based models.