A Julia package implementing methods for planning and evaluation of exact single-arm two-stage designs with binary endpoint. Documentation is available here.
Several key features depend on Integer Linear Programs and a suitable solver for the JuMP-package must be installed prior to using these. The package is tested against the commercial solver Gurobi and its Julia interface Gurobi.jl. Free academic licenses for Gurobi can be obtained here. Make sure Gurobi/Gurobi.jl/JuMP are working correctly before using any of the functionality involving optimization.
The package is not yet registered and the current development verions can be installed via:
Julia> Pkg.clone("https://github.com/imbi-heidelberg/BinaryTwoStageDesigns")
Note that these versions need not be stable (minor fixes and documentation changes). A specific release, which is just a tagged commit, can be checked out from the cloned git repository. Switch to ~/.julia/v0.6/BinaryTwoStageDesigns in your user folder and checkout the tag directly (git must be installed), e.g., by
git checkout v0.1.0
Note that this will cause a warning about a 'detached HEAD' state. This is fine for usage of the specific version and only has implications for developers. Alternatively, simply download the release as .zip folder from https://github.com/imbi-heidelberg/BinaryTwoStageDesigns/releases and unpack to your local .julia/v0.6 directory.
The documentation of the current development version is available here and previous versions can be found locally under /docs/build/index.html of the respective release.