Contemporaneous sampling

[ArtPoon]

Should be just a matter of having sim.dynamics run to simTime endpoint and then taking a random sample of infected hosts at that time.

[ArtPoon]

• we should support either a fixed probability that members of a compartment are sampled (with probability p), or a fixed number of individuals to sample
• first step in sim.outer.tree is to check whether there are sufficient numbers in each compartment being sampled at end of dynamics simulation
• append events to event log that correspond to sampling from compartments at end time
• then the rest of sim.outer.tree should proceed as usual

[ArtPoon]

We might need to create a fake sampled compartment because the model does not specify one under contemp sampling