RMixtCompIO is the minimal R interface of the MixtComp C++ library. It is recommended to use RMixtComp which is more versatile and convenient to use. The version number of RMixtCompIO follows that of MixtComp C++ library.

You can download it from CRAN in an R console:

install.packages("RMixtCompIO", repos = "https://cran.rstudio.com")

The C++ library CppNumericalSolvers is required in the src folder. It is available via a git submodule. Run in a terminal:

git submodule init
git submodule update

R must be installed, along with the following packages from the CRAN: RcppEigen, BH, doParallel, foreach. Several packages are used for testing but are not required for installing RMixtComp: testthat is used to perform unit testing in R, xml2 for xml report, RInside is used for testing procedure in C++ using R command, roxygen2 is required for generating the documentation and devtools (with roxygen2) is required to load all functions of RMixtCompIO on the CI server.

1. Install R from terminal or download it from R-project:

   sudo apt install r-base
   

2. Install required R packages

   Packages are required from Ubuntu repositories:

   sudo apt install libssl-dev libcurl4-openssl-dev libxml2-dev
   

   Install required R packages from terminal:

   Rscript -e "install.packages(c(\"BH\", \"RcppEigen\", \"doParallel\", \"foreach\"), repos = \"https://cran.rstudio.com\")"
   # OPTIONAL: for testing purpose
   Rscript -e "install.packages(c(\"testthat\", \"RInside\", \"xml2\", \"devtools\", \"roxygen2\", \"Rmixmod\", \"blockcluster\"), repos = \"https://cran.rstudio.com\")"
   

   from R:

   install.packages(c("BH", "RcppEigen", "doParallel", "foreach"), repos = "https://cran.rstudio.com")
   # OPTIONAL: for testing purpose
   install.packages(c("testthat", "RInside", "xml2", "devtools", "roxygen2", "Rmixmod", "blockcluster"), repos = "https://cran.rstudio.com")
   

Compiling R packages requires Rtools.

Run in the current location (./RMixtCompIO/ folder):

make updateLib
make all

The target updateLib copies the MixtComp library C++ folder from MixtComp/src/lib folder to src/lib folder.

R CMD build RMixtCompIO

It generates a tar.gz file

R CMD check --as-cran RMixtCompIO

or you can build the package before

R CMD check --as-cran RMixtCompIO_*.tar.gz

If you do not have windows, build the source and upload the tar.gz at https://win-builder.r-project.org/upload.aspx

To check on (almost) the same linux configuration as the CRAN, you can use the following procedure: https://r-hub.github.io/containers/local.html

1. Install Docker
2. Create a check folder and put the tar.gz package inside
3. Run

docker run -v `pwd`/check:/check ghcr.io/r-hub/containers/gcc13:latest r-check

See https://r-hub.github.io/containers/containers.html for available containers (clang16, clang-asan, ...)

Tests are implemented in the tests/testthat folder of RMixtCompIO. They require the testthat package.

You can run tests via R:

library(RMixtCompIO)
library(testthat)
library(devtools)
load_all()
test_dir("tests/testthat/")

or

library(methods); 
library(devtools); 
options(testthat.output_file = "RMCIOtest.xml"); 
test(".", reporter = JunitReporter)

but it requires the xml2 package and generates a xml report file named RMCIOtest.xml.

Via shell:

Rscript -e 'library(methods); library(devtools); options(testthat.output_file = "RMCIOtest.xml"); test(".", reporter = JunitReporter)'

In Rstudio, if packages devtools and testthat are installed, you can run the testing procedure by pressing CRTL + SHIFT + T or tests are run during a R CMD check CRTL + SHIFT + E.

For your own testing procedure, you can generate your own data using dataGeneratorNewIO function (from files R/GENDATA_dataGeneratorNewIO.R) for example.

See functions in files R/TEST_*.R for creating a parameter object for each type of data. Have a look at tests/testthat/test.run.R for writing a test with testthat.

It requires RInside package. Tests are in the src/test folder. You can compile and run it from Eclipse.