SimpleITKRInstaller

A remotes based installer for SimpleITK in R.

Default configuration, minimal compute resources but takes a long time, single core compilation:

remotes::install_github("SimpleITK/SimpleITKRInstaller")

Turn on mutlicore compilation, recommended configuration, six cores in this example:

Sys.setenv(MAKEJ=6)
remotes::install_github("SimpleITK/SimpleITKRInstaller")

Use multicore compilation and build additional modules not included in the default build setup such as SimpleElastix (registration) and DCMTK (additional DICOM IO option beyond the default GDCM).

Sys.setenv(MAKEJ=6)
Sys.setenv(ADDITIONAL_SITK_MODULES="-DSimpleITK_USE_ELASTIX=ON -DModule_ITKIODCMTK:BOOL=ON")
remotes::install_github("SimpleITK/SimpleITKRInstaller")

Note: On Linux and Mac requires CMake and git in the path.

On Windows requires rtools installation and setting the RTOOLS_HOME environment variable. Install the rtools version that matches the targeted R version. For example, using the R4.5.3 and rtools4.5 combination, set the enviornment variable to:

Sys.setenv(RTOOLS_HOME = "C:/rtools45")

Building a specific SimpleITK version

If you want to create a custom build for a specific SimpleITK release or a development version you need to:

1. Set the SITK_TARGET in the DESCRIPTION file to the appropriate committish (hash, tag or branch).
2. Run the sitk_r_version_date.sh script which will update the Version and Date entries in the DESCRIPTION file to match the expected format. For Version this is MAJOR.MINOR.PATCH.[9000] where the last entry is an "R universe" convention denoting "in development".

How to Cite

If you find the R version of SimpleITK useful in your research, support our efforts by citing it as:

R. Beare, B. C. Lowekamp, Z. Yaniv, "Image Segmentation, Registration and Characterization in R with SimpleITK", J Stat Softw, 86(8), https://doi.org/10.18637/jss.v086.i08, 2018.