Multistate Population Projection Model for Demographic and HIV Estimation • leapfrog

Installation

Please install from our r-universe:

install.packages(
  "leapfrog",
  repos = c("https://hivtools.r-universe.dev", "https://cloud.r-project.org"))

You can install the latest release from GitHub with:

# install.packages("remotes")
remotes::install_github("hivtools/leapfrog", subdir = "leapfrogr", ref = "r-release")

To install a development version on a branch checkout the repo to the appropriate branch and take the following steps:

(One time only) setup the LEAPFROG_INCLUDE env var using
```
./scripts/setup_dev_env.R
```
Generate the C++
```
../scripts/generate
```
Install the package
```
Rscript -e  "devtools::install()"
```

Simulation model

The “leapfrog” R package is an R interface around the “leapfrog” simulation model. The model is written in C++ and has interfaces in R, Python and C. Note the R package is in a subdirectory leapfrogr but the installed package is called leapfrog.

The simulation model is implemented in a header-only C++ library located in leapfrog-core/include/. The headers are bundled into inst/include in the installed package which allows the C++ code to be imported in other R packages via specifying LinkingTo: leapfrog in the DESCRIPTION file.

[!IMPORTANT] We use C++20 for this package. Please make sure you have a compiler that is compatible.

The simulation model is callable in R via a wrapper function run_model() created with Rcpp.

You can control how the simulation model is run by setting the configuration argument, see

leapfrog::list_model_configurations()

for the list of available models. At the time of writing these are

DemographicProjection - runs on the demographic projections
HivFullAgeStratification - runs the demographic projection and HIV adult projection with single-year age groups
HivCoarseAgeStratification - runs the demographic projection and HIV adult projection with 5-year age groups
ChildModel - runs the demographic projection, single-year age stratified HIV adult projection and HIV child projection

Example

The file inst/pjnz/bwa_aim-adult-art-no-special-elig_v6.13_2022-04-18.PJNZ contains an example Spectrum file constructed from default country data for Botswana with Spectrum (April 2022).

Prepare model inputs.

library(leapfrog)

pjnz <- system.file("pjnz/bwa_aim-adult-art-no-special-elig_v6.13_2022-04-18.PJNZ",
                    package = "leapfrog", mustWork = TRUE)

parameters <- process_pjnz(pjnz)
#> ℹ Some tags were not found in file:
#> • Tag not found in DP for `adult_non_aids_excess_mort`, returning `NULL`
#> • Tag not found in DP for `incidence_input`, returning `NULL`
#> • Tag not found in DP for `adult_art_adj_factor`, returning `NULL`
#> • Tag not found in DP for `adult_art_adj_factor_flag`, returning `NULL`
#> • Tag not found in DP for `adult_pats_alloc_to_from_other_region`, returning
#>   `NULL`
#> • Tag not found in DP for `nosocomial_infections`, returning `NULL`
#> This message is displayed once per session.

Simulate adult ‘full’ age group (single-year age) model from 1970 to 2030 with 10 HIV time steps per year.

lsimF <- run_model(parameters, "HivFullAgeStratification", 1970:2030)

You can also simulate a model with ‘coarse’ age group (5-year age groups). You need to first prepare the coarse age group parameters

params_coarse <- process_pjnz(pjnz, use_coarse_age_groups = TRUE)
lsimC <- run_model(params_coarse, "HivCoarseAgeStratification", 1970:2030)

Compare the HIV prevalence age 15-49 years and AIDS deaths 50+ years. Deaths 50+ years are to show some noticeable divergence between the "full" and "coarse" age group simulations.

prevF <- colSums(lsimF$p_hivpop[16:50,,],,2) / colSums(lsimF$p_totpop[16:50,,],,2)
prevC <- colSums(lsimC$p_hivpop[16:50,,],,2) / colSums(lsimC$p_totpop[16:50,,],,2)

deathsF <- colSums(lsimF$p_hiv_deaths[51:81,,],,2)
deathsC <- colSums(lsimC$p_hiv_deaths[51:81,,],,2)

plot(1970:2030, prevF, type = "l", main = "Prevalence 15-49")
lines(1970:2030, prevC, col = 2)


plot(1970:2030, deathsF, type = "l", main = "AIDS Deaths 50+ years")
lines(1970:2030, deathsC, col = 2)

Benchmarking

Install the package and then run the benchmarking script ../scripts/benchmark

lint

Lint R code with lintr

lintr::lint_package()

Lint C++ code with cpplint

cpplint leapfrog-core/include/*

Code design

R functions

The file src/leapfrog.cpp contains R wrapper functions for the model simulation via Rcpp.

Development notes

Package loading

The C++ header library lives in leapfrog-core/include/ (outside the R package), with generated headers produced by a codegen step. How this is handled depends on how you are building/installing the package:

Method	`configure` runs in	Headers used	Manual steps	Notes
`devtools::load_all()`	`leapfrogr/` (source)	`LEAPFROG_INCLUDE` or relative path to `../leapfrog-core/include`	Run codegen `../scripts/generate` manually on first run or when schemas change.	When headers in `leapfrog-core/include/` change, C++ recompiles automatically via `Config/build/extra-sources`.
`devtools::install()` (default)	Temp dir (tarball)	`LEAPFROG_INCLUDE` env var	`LEAPFROG_INCLUDE` must be set and run codegen manually	`configure` runs from temp dir so `../leapfrog-core` is absent.
`R CMD INSTALL leapfrogr/`	`leapfrogr/` (source)	`../leapfrog-core/include`	Nothing (in monorepo)	Will need `LEAPFROG_INCLUDE` env var if using R CMD build
`remotes::install_github(..., ref = "r-release")`	Temp dir (subdir only)	`inst/include` (bundled)	Nothing	Self-contained. `r-release` branch has headers bundled. Other branches will fail — use `build=FALSE` from a clone instead.
R-universe	Build server	`inst/include` (bundled)	Nothing	Preferred for end users. `r-release` branch is used per `packages.json`. Updated on every merge to `main`.

To set up LEAPFROG_INCLUDE for local development, run the helper script once:

./scripts/setup_dev_env.R

Note that LEAPFROG_INCLUDE is not needed for remotes::install_github(..., ref = "r-release") or r-universe installs — those use the r-release branch which has the C++ headers bundled into inst/include/. The r-release branch is updated automatically on every merge to main.

For CRAN submission (when we add it), build the tarball from the r-release branch — the headers are already present so it will pass R CMD check.

Testing

There is some pre-prepared test data available to make tests run faster. This is generated and saved by ../scripts/create_test_data.R as hdf5 format files.

This data is used in leapfrog-py for pythont tests and cpp_interface for testing standalone C++ model.

leapfrog R package