MicroMoB-connector

This runs MicroMoB using the Beverton-Holt model of aquatic mosquito dynamics and the Ross-Macdonald model of adult mosquito dynamics for use in the Model Runner.

The sample parameters were made with:

p <- 2
tmax <- 365

surv_a <- 0.9

adult_stochastic <- TRUE
adult_f <- 0.3
adult_q <- 0.9
adult_eip <- rep(10, tmax)

input <- replicate(n = 3, expr = {
    mu <- log(1 / surv_a)
    adult_p <- t(replicate(n = p, expr = {pexp((sin((1:tmax)/365*2*pi) + 2)/2 * mu, lower.tail = FALSE)}))
    adult_psi <- diag(p)
    adult_nu <- 25
    adult_M <- rpois(n = p, lambda = 100)
    adult_Y <- rep(0, times = p)
    adult_Z <- rep(0, times = p)
    
    molt <- 0.1
    surv <- 0.9
    
    aqua_stochastic <- TRUE
    aqua_molt <- matrix(data = molt, nrow = p, ncol = tmax)
    aqua_surv <- matrix(data = surv, nrow = p, ncol = tmax)
    
    lambda <- adult_M*(1-surv_a)
    eggs <- adult_nu * adult_f * adult_M
    
    aqua_L <- lambda * ((1/molt) - 1) + eggs
    aqua_K <- - (lambda * aqua_L) / (lambda - aqua_L*molt*surv)
    aqua_K <- t(sapply(X = aqua_K, FUN = function(x) {rep(x, tmax)}))
    
    list(
        p = p,
        tmax = tmax,
        aqua_stochastic = aqua_stochastic,
        aqua_molt = aqua_molt,
        aqua_surv = aqua_surv,
        aqua_K = aqua_K,
        aqua_L = aqua_L,
        adult_stochastic = adult_stochastic,
        adult_f = adult_f,
        adult_q = adult_q,
        adult_eip = adult_eip,
        adult_p = adult_p,
        adult_psi = adult_psi,
        adult_nu = adult_nu,
        adult_M = adult_M,
        adult_Y = adult_Y,
        adult_Z = adult_Z
    )
}, simplify = FALSE)

jsonlite::write_json(x = input, path = "test/input.json",digits=17,pretty=TRUE)
jsonvalidate::json_validate(json = "test/input.json", schema = "schema/input.json")

You can read in and plot the output in R to check everything is working correctly with:

library(data.table)
library(ggplot2)

output <- jsonlite::read_json(path = "output/data.json", simplifyVector = TRUE)

MYZ <- NULL
for (i in 1:length(output$output$MYZ)) {
  
  MYZ_i <- as.data.table(output$output$MYZ[[i]])
  data.table::setnames(x = MYZ_i, old = c("V1", "V2", "V3"), new = c("Patch", "Day", "Stage"))
  
  MYZ_i[,
      Stage := fcase(
        Stage == 1, "M",
        Stage == 2, "Y",
        Stage == 3, "Z"
      )]
  MYZ_i[, 'run' := as.integer(i)]
  MYZ <- rbind(MYZ, MYZ_i)
}

ggplot(data = MYZ) +
  geom_line(aes(x = Day, y = value, group = run), alpha = 0.75) +
  facet_wrap(Patch ~ Stage, scales = "free") +
  theme_bw()

You should see output which resembles the following plot, although if you used stochastic model(s) there will be sampling variation.