survival in bloodmeal sampling

[slwu89]

Let the current time step be t0, and the next time step begins at t1 = t0 + dt. When we are sampling S->E events in mosquitoes (https://github.com/dd-harp/mini-mash/blob/main/src/disaggregated/pop/bloodmeal.cpp#L41) the issue is that not all of these bites will go to mosquitoes that are alive at t1 of the next time step. If we assume all bites sampled during the bloodmeal module during [t0, t0+dt) go to mosquitoes alive at time t1, we will overestimate the number of S->E events, because some bites go to mosquitoes that die before reaching t1. We currently fix this by sampling the survival probability for each bite within the mosquito module when the bites are fed in: https://github.com/dd-harp/mini-mash/blob/main/src/disaggregated/pop/mosquito.cpp#L107

Additionally when calculating the intensity, we must subtract the total amount of bites sampled so far; H2M_lambda = a * c * (IH/NH) * (SV - (double)H2M_tot) * delta;. This is because the risk set of mosquitoes gets smaller as bites happen, because those mosquitoes are no longer susceptible.