launch_script.R

## This is my launch script for the comparison between the two models

## Cluster set up
## Prepare for cluster use
## see https://mrc-ide.github.io/hipercow/
hipercow::hipercow_init()
hipercow::hipercow_configure(driver = "windows")

## Provision packages required on the cluster
## see https://mrc-ide.github.io/hipercow/articles/packages.html
hipercow::hipercow_provision()
# hipercow::hipercow_configuration()
# hipercow::hipercow_hello()



################################
## Model equilibrium analyses ##
################################

## Local test
# test <- orderly2::orderly_run(
#   name = "equilibrium_comparison",
#   parameters = list(
#     investigation = "default",
#     het_group_outputs = F,
#     parameter_draws = F,
#     EIR_range = "sparse",
#     additional_outputs = F
#   )
# )


#'- Generate standard (default) @equilibrium solutions
#--- With heterogeneity group outputs (for mixed infection calculations)
#--- Default parameter assumptions only (no parameter draws)
#--- With detailed range of EIR (for complete results set)
#--- With no additional (age-structured) outputs

default_run <- hipercow::task_create_expr(
  orderly2::orderly_run(
    name = "equilibrium_comparison",
    parameters = list(
      investigation = "default",
      het_group_outputs = T,
      parameter_draws = F,
      EIR_range = "detailed",
      additional_outputs = F
    )
  )
)

hipercow::task_status(default_run)
# hipercow::task_log_show(default_run)
# hipercow::task_cancel(default_run)


#'- Generate *sensitivity analyses* @equilibrium solutions
#--- These are:
#'---- varying *schizont-targeting drug coverage*
#'---- varying *hypnozoite relapse rate*
#'---- varying *mosquito biting heterogeneity*
#'---- varying *mosquito bioparameters*: q0, death rate and biting rate
#--- With no heterogeneity group outputs (only for mixed infection calculations)
#--- Default assumptions only (no parameter draws)
#--- With detailed range of EIR (for complete results set)
#--- With no additional (age-structured) outputs

# Other investigations don't return het outputs
equlibrium_investigations <-
  c("vary_ft",
    "vary_relapse_rate",
    "vary_biting_het",
    "decrease_q0",
    "decrease_death_rate",
    "increase_biting"
  )

investigations_run_task_loop <- list()
for(investigation in equlibrium_investigations){
  investigations_run_task_loop[[investigation]] <- hipercow::task_create_expr(
    orderly2::orderly_run(
      name = "equilibrium_comparison",
      parameters = list(
        investigation = investigation,
        het_group_outputs = F,
        parameter_draws = F,
        EIR_range = "detailed",
        additional_outputs = F
      )
    )
  )
}

# table(unlist(lapply(investigations_run_task_loop, hipercow::task_status)))
# lapply(investigations_run_task_loop, hipercow::task_cancel)
# hipercow::task_log_show(investigations_run_task_loop[[1]])


#'- Generate standard (default) @equilibrium solutions [over parameter draws]
#--- With heterogeneity group outputs (for mixed infection calculations)
#'-- [Over 50 parameter draws]
#--- With detailed range of EIR (for complete results set)
#--- With no additional (age-structured) outputs

default_draws_run_task <- hipercow::task_create_expr(
  orderly2::orderly_run(
    name = "equilibrium_comparison",
    parameters = list(
      investigation = "default",
      het_group_outputs = T,
      parameter_draws = 50,
      EIR_range = "detailed",
      additional_outputs = F
    )
  ),
  parallel = hipercow::hipercow_parallel("parallel"),
  # You could tailor cores if number to calibrate < 32
  resources = hipercow::hipercow_resources(cores = 32)
)

# hipercow::task_status(default_draws_run_task)
# hipercow::task_log_show("9b05438cc6750357b236b99d1e5b0bc9")
# lapply(default_draws_run_task, hipercow::task_cancel)

#'- Generate *sensitivity analyses* @equilibrium solutions [over parameter draws]
#--- These are:
#'---- varying *schizont-targeting drug coverage*
#'---- varying *hypnozoite relapse rate*
#'---- varying *mosquito biting heterogeneity*
#'---- varying *mosquito bioparameters*: q0, death rate and biting rate
#--- With no heterogeneity group outputs (only for mixed infection calculations)
#'-- [Over 50 parameter draws]
#--- With detailed range of EIR (for complete results set)
#--- With no additional (age-structured) outputs

investigations_draws_run_task_loop <- list()
for(investigation in equlibrium_investigations){
  investigations_draws_run_task_loop[[investigation]] <- hipercow::task_create_expr(
    orderly2::orderly_run(
      name = "equilibrium_comparison",
      parameters = list(
        investigation = investigation,
        het_group_outputs = F,
        parameter_draws = 50,
        EIR_range = "detailed",
        additional_outputs = F
      )
    ),
    parallel = hipercow::hipercow_parallel("parallel"),
    # You could tailor cores if number to calibrate < 32
    resources = hipercow::hipercow_resources(cores = 32)
  )
}

# table(unlist(lapply(investigations_draws_run_task_loop, hipercow::task_status)))
# lapply(investigations_draws_run_task_loop, hipercow::task_log_show)
# lapply(investigations_draws_run_task_loop, hipercow::task_cancel)


#'- Generate standard (default) [age-structured] @equilibrium solutions
#--- With no heterogeneity group outputs (for mixed infection calculations)
#--- Default parameter assumptions only (no parameter draws)
#--- With detailed range of EIR (for complete results set)
#'-- With [age-structured] outputs

age_outputs_run_task <- hipercow::task_create_expr(
  orderly2::orderly_run(
    name = "equilibrium_comparison",
    parameters = list(
      investigation = "default",
      het_group_outputs = F,
      parameter_draws = F,
      EIR_range = "detailed",
      additional_outputs = T
    )
  ),
  parallel = hipercow::hipercow_parallel("parallel"),
  # You could tailor cores if number to calibrate < 32
  resources = hipercow::hipercow_resources(cores = 32)
)

# hipercow::task_status(age_outputs_run_task)
# hipercow::task_cancel(age_outputs_run_task)


#'- [Plot equilibrium results]
orderly2::orderly_run("equilibrium_comparison_plotting")




##################################################
## Malaria atlas project co-prevalence analyses ##
##################################################

# Set analysis admin level
admin_level <- 1

#'- [Fetch and compile site file prevalence, population and genetic estimates]
orderly2::orderly_run("download_country_site_files",
                      parameters = list(admin_level = admin_level))

#'- [Generate Pf/Pv endemic regions for use in analysis (with key results and plots)]
orderly2::orderly_run("assess_site_prevalences",
                      parameters = list(admin_level = admin_level))

#'- [Perform optimisation of kmeans clustering]
orderly2::orderly_run("clustering_algorithm_trajectories",
                      parameters = list(admin_level = admin_level))

#'- [Plot MAP co-prevalence results and key analysis outputs]
#'- *This includes all:*
#'--- *clustered trajectories (including by country and geographically)*
#'--- *clustered trajectories compared with equilibrium results (including sensitivity analyses)*
#'- Number of clusters is set to @7 in line with results from clustering_algorithm_trajectories
orderly2::orderly_run("clustering_plotting_trajectories",
                      parameters = list(k_groups = 7,
                                        admin_level = admin_level))





###########################
## Intervention simulations
###########################

#'- [Create scenarios: ITN and treatment simulations]
#'- *treatment at 90% for 10 years*
#'- *a single ITN distribution at 90% coverage*

scenarios <- c("control", "treatment", "itns")
for(scenario in scenarios){
  orderly2::orderly_run(
    name = "sim_create_scenarios",
    parameters = list(
      scenario = scenario
    )
  )
}

#'- [Test run functions]
# orderly2::orderly_run(
#   name = "sim_run_simulations",
#   parameters = list(
#     scenario = "control",
#     total_M_to_match = 5,
#     species_to_run = 'falciparum',
#     simulation_years = 5,
#     repetitions = 1,
#     parallel = FALSE
#   )
# )


#'- [Run scenarios: over a range of EIR values]
#'- *EIRs correspond with mosquito density*

total_Ms <- c(0.2, 0.3, 0.35, 0.4, 0.45, 0.47, 0.49, 0.5, 1, 5, 10)
species_v <- c("falciparum", "vivax")

run_scenarios_task <- list()
for(scenario in scenarios){
  for(total_M in total_Ms[2:4]){
    for(species in species_v){
      run_scenarios_task[[paste0(c(scenario, total_M, species), collapse = "-")]] <- hipercow::task_create_expr(
        orderly2::orderly_run(
          name = "sim_run_simulations",
          parameters = list(
            scenario = scenario,
            total_M_to_match = total_M,
            species_to_run = species,
            simulation_years = 20,
            repetitions = 20,
            parallel = TRUE
          )
        ),
        parallel = hipercow::hipercow_parallel("parallel"),
        # You could tailor cores if number to calibrate < 32
        resources = hipercow::hipercow_resources(cores = 32)
      )
    }
  }
}

table(unlist(lapply(run_scenarios_task, hipercow::task_status)))
# lapply(run_scenarios_task, hipercow::task_cancel)
# hipercow::task_log_show(run_scenarios_task$`itns-5-falciparum`)

#'- [Plot scenario results]
orderly2::orderly_run(name = "sim_simulation_plots",
                      parameters = list(scenarios = paste0(scenarios, collapse = "--"),
                                        total_Ms = paste0(total_Ms[c(2:4, 5, 9, 11)], collapse = "--")))