Functionality for using nested variable and event lists

[pwinskill]

Adding capability to setup nested lists of variables and events. This can aid the user in making "grouped" variables or events and is helpful if the group size may vary. For example the variables list may contain a sub list of variables "vaccines" of length number of unique vaccines that can be used to track an individual's vaccine status and be flexible to changing numbers or available vaccines.

This goes some way to resolving #52

A couple of points that it would be good to hear your feedback on @giovannic :

1. This implementation does obfuscate the simulation loop code somewhat (it now calls a recursive function for events and variable updates). Not sure if there is a more clear way to do this?
2. Would appreciate your feedback on tests for this addition (I can implement some, but the testing in this package is very concise and neat).

If it helps here is some example code using the changes to model 4, independent diseases in a population, adapted from the SIR vignette:

# 4 Diseases each with unique R0, all other parameters are the same
n_disease <- 4
R0 <- c(2, 3, 4, 5)

N <- 1e3
I0 <- 5
S0 <- N - I0
dt <- 0.1
tmax <- 100
steps <- tmax/dt
gamma <- 1/10
beta <- R0 * gamma
health_states <- c("S","I","R")
health_states_t0 <- rep("S",N)
health_states_t0[sample.int(n = N,size = I0)] <- "I"

# Created nested list of variables where states is a sub list of the health state for each disease
variables <- list()
variables$states <- lapply(1:n_disease, function(x){
  CategoricalVariable$new(categories = health_states, initial_values = health_states_t0)
})

# Infection process
infection_process <- function(t){
  for(disease in 1:n_disease){
    I <- variables$states[[disease]]$get_size_of("I")
    foi <- beta[disease] * I/N
    S <- variables$states[[disease]]$get_index_of("S")
    S$sample(rate = pexp(q = foi * dt))
    variables$states[[disease]]$queue_update(value = "I",index = S)
  }
}

# Recovery events
events <- list()
events$recovery <- lapply(1:n_disease, function(x){
  TargetedEvent$new(population_size = N)
})

update_status <- function(variable, new_status){
  force(variable)
  force(new_status)
  function(timestep, target){
    variable$queue_update(new_status, target)
  }
}


# add listeners
for(disease in 1:n_disease){
  events$recovery[[disease]]$add_listener(
    update_status(variables$states[[disease]], "R"))
}
# Recovery process
recovery_process <- function(t){
  for(disease in 1:n_disease){
    I <- variables$states[[disease]]$get_index_of("I")
    already_scheduled <- events$recovery[[disease]]$get_scheduled()
    I$and(already_scheduled$not())
    rec_times <- rgeom(n = I$size(),prob = pexp(q = gamma * dt)) + 1
    events$recovery[[disease]]$schedule(target = I, delay = rec_times)
  }
}

# Render output
health_render <- Render$new(timesteps = steps)
render_states <- function(t){
  for(disease in 1:n_disease){
    for(i in seq_along(health_states)){
      size <- variables$states[[disease]]$get_size_of(health_states[i])
      health_render$render(paste0(health_states[i], disease), size, t)
    }
  }
}

# Run simulation
out <- simulation_loop(
  variables = variables,
  events = events,
  processes = list(infection_process,recovery_process,render_states),
  timesteps = steps
)

# Visualise output
states <- health_render$to_dataframe() %>%
  tidyr::pivot_longer(-timestep, names_to = "state", values_to = "n") %>%
  tidyr::separate(state, sep = 1,  c("State", "Disease"))

ggplot2::ggplot(states, ggplot2::aes(x = timestep, y = n, col = Disease, linetype = State)) +
  ggplot2::geom_line() +
  ggplot2::theme_bw()

[codecov]

Codecov Report

Merging #100 (21ddf40) into dev (c9b30ce) will decrease coverage by 0.24%.
The diff coverage is 75.00%.

@@            Coverage Diff             @@
##              dev     #100      +/-   ##
==========================================
- Coverage   92.50%   92.26%   -0.25%     
==========================================
  Files          26       26              
  Lines        1054     1060       +6     
==========================================
+ Hits          975      978       +3     
- Misses         79       82       +3     

Impacted Files	Coverage Δ
R/simulation.R	80.00% <75.00%> (-12.86%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 300a6fe...21ddf40. Read the comment docs.