0-CreateProject.Rmd

# Create HBV model project
# ========================

Authors: Richard T. Gray, Neil Bretana

This Rmarkdown script is used to create a HBV modelling project specified 
by the user. The project specifications entered by the user are hardcoded 
in the "Project specifications" chunk.

```{r setup}
# Clear workspace
rm(list = ls()) 
library("readr")
library("dplyr")
library("ggplot2")

# Set working directory to source file directory first
basePath <- getwd()

```

## User inputs

The following chunk is where the user enters the project specifications

```{r Project specifications}
# Hard coded user specified project details

project_name <- "test_project"
project_directory <- file.path(basePath, "projects")

# Start of the first year
start_year <- 1951

#start of the final year represents end of year before 
end_year <- 2050

# Simulation timestep
timestep <- 0.1

# Populations modelled
full_population_names <- c("Infants aged 0 to 4", 
                           "Children aged 5 to 14",
                           "Adults aged 15 to 44", 
                           "Adults older tha 45")

population_names <- c("age0to4", 
                      "age5to14", 
                      "age15to44", 
                      "age45")
  
```

## Project creation

The following specifications are fixed for all models and are required 
for the model to run (i.e. they need to be changed if the model changes)

```{r Model specifications}
# Fixed compartmental names
full_state_names <- c("susceptible",  
                      "acute", 
                      "chronic", 
                      "cleared",
                      "immune")

# Short compartmental names
state_names <- c("s", "a", "ch", "cl", "i")

```

```{r Parameter specifications}

# Note parameters that are for specifying the initial population sizes
# so they can be stored separately
initialPopsNames <- c("init_pop", "pop_prop_", "sus_prop_", "acute_prop_",
                     "chronic_prop_", "clear_prop_", "immune_prop_")

# The following constant parameters are generated for each population 
constNamesPops <- c("sus_prop_", "acute_prop_", "chronic_prop_",
                    "clear_prop_", "immune_prop_", 
                    "pop_prop_", "prog_chron_", 
                    "clr_rate_", "ac_mort_rate_", 
                    "chr_mort_rate_", "vac_eff_",
                    "vac_prop_")

# The following constant parameters are the same for all populations or
# population agnostic.
constNames <- c("init_pop", "mig_series", "dis_rate", "ac_res_rate",
                "fol_mult", "vacc_tog", "chronic_beta")

# Note pop_prop and init_pop are only used for setting up the initial 
# population.

# Following parameters all vary over time
timeVaryNames <- c("years", "notifications", "births")

# TODO - make these population varying as well like the constants - 0 to 
# npops
timeVaryPops <- c("bgmort0to4", "bgmort5to14", "bgmort15to44", "bgmort45",
                  "mig0to4chronic", "mig0to4cleared", "mig0to4sus",
                  "mig5to14chronic", "mig5to14cleared", "mig5to14sus", 
                  "mig15to44chronic", "mig15to44cleared", "mig15to44sus", 
                  "mig45chronic", "mig45cleared", "mig45sus")

```

The following chunk creates the files used for inputing parameters based 
on the project specifications. Once these are created the user needs to
enter input values and ranges for the model to run

```{r Create directories and csv files}
# Create project directory
projectFolder <- file.path(project_directory, project_name)
dir.create(projectFolder, showWarnings = FALSE)

# Create other directories
dir.create(file.path(projectFolder, "results"), showWarnings = FALSE)
dir.create(file.path(projectFolder, "docs"), showWarnings = FALSE)
dir.create(file.path(projectFolder, "data"), showWarnings = FALSE)

# Create parameter files -------------------------------------------------
npops <- length(population_names)
nyears <- end_year - start_year + 1

numConstants <- npops * length(constNamesPops) + length(constNames)

# Create constants file
constants <- data_frame(parameter = character(numConstants),
                        value = 0,
                        lower = 1,
                        upper = 1)

for (param in 1:length(constNamesPops)) {
  popParams <- paste0(constNamesPops[param], as.character(0:(npops-1)))
  index <- (param - 1) * npops + 1
  constants$parameter[index:(index+npops-1)] <- popParams
}

constants$parameter[(index + npops):numConstants] <- constNames

# Separate initial population parameters from other parameters
fullInitialNames <- character()
for (param in 1:length(initialPopsNames)){
  if (initialPopsNames[param] == "init_pop") {
    fullInitialNames <- c(fullInitialNames, initialPopsNames[param])
  } else {
    initialParams <- paste0(initialPopsNames[param],
                            as.character(0:(npops-1)))
    fullInitialNames <- c(fullInitialNames, initialParams)
  }
}

initialPops <- constants %>%
  filter(parameter %in% fullInitialNames)
constants <- constants %>%
  filter(!(parameter %in% fullInitialNames))

# Write constants to file
write_csv(initialPops, file.path(projectFolder,
                                    "initial_populations.csv"))
write_csv(constants, file.path(projectFolder,
                                    "parameters_constants.csv"))

# Create time varying paramater file
# TODO: change when time varying population paramaters are included
timeVarying <- as.data.frame(matrix(0, nrow = nyears, 
  ncol = length(timeVaryNames) + length(timeVaryPops)))
colnames(timeVarying) <- c(timeVaryNames, timeVaryPops)

timeVarying$years <- start_year:end_year

write_csv(timeVarying, file.path(projectFolder,
                               "parameters_time_varying.csv"))

# Create time varying relative factors file - for each time varying 
# parameter create a start range and an upper range for the relative
# parameter values
timeFactors <- as.data.frame(matrix(1, nrow = ncol(timeVarying), 
                                    ncol = 5))
colnames(timeFactors) <- c("parameter", "start_lower", "start_upper", 
                           "end_lower", "end_upper")

timeFactors$parameter <- colnames(timeVarying)

write_csv(timeFactors, file.path(projectFolder,
                               "time_varying_ranges.csv"))

# Create associated transitions and waifw_matrix files -------------------
transitions <- as.data.frame(matrix(0, nrow = npops, ncol = npops + 1))
colnames(transitions) <- c("population", population_names)
transitions$population <- population_names

waifw_matrix <- transitions

write_csv(transitions, file.path(projectFolder,
                               "population_transitions.csv"))

write_csv(waifw_matrix, file.path(projectFolder,
                               "population_waifw_matrix.csv"))

```

The following chunk creates a project R data file for storing the project 
variables and inputs

```{r Create project specifications list}
# Create project specifications list
pg <- list()
pg$project_name <- project_name
pg$project_directory <- project_directory
pg$start_year <- start_year
pg$end_year <- end_year
pg$timestep <- timestep
pg$full_population_names <- full_population_names
pg$population_names <- population_names

# Additional useful specs - further specs will be appended when generating
# the parameter sets for a project
pg$npops <- npops
pg$years <- start_year:end_year
pg$nyears <- nyears

pg$full_state_names <- full_state_names
pg$state_names <- state_names
pg$nstates <- length(state_names)

# Create an inputs list
inputs <- list(initial_pops = initialPops, constants = constants,
               time_varying = timeVarying,
               time_varying_ranges = timeFactors,
               transitions = transitions, waifw_matrix = waifw_matrix)

# Create project .rda files
save(pg, inputs, file = file.path(projectFolder,
  paste0(project_name, ".rda")))

```