Galaaz

Galaaz 2.0 drives GNU R from JRuby over a process bridge. You write Ruby (R.c(...), dplyr-style chains, ggplot2) and talk to the same R you use for CRAN and Bioconductor. It is not the older GraalVM / TruffleRuby / FastR stack.

Install

You need JRuby (tested with 10.1.1.0), Java 21, GNU R with Rscript, a C++ toolchain, and Rcpp (install.packages("Rcpp")). MRI Ruby is not supported.

jruby -S gem install galaaz
gem_dir="$(jruby -e "puts Gem::Specification.find_by_name('galaaz').full_gem_path")"
make -C "${gem_dir}/ext/new_bridge" all

Full steps, WSL/Docker notes, and the contributor clone path are in the manual and the README. The gem is on RubyGems (changelog).

Try (Docker)

A public try image has JRuby, GNU R, and Galaaz already installed and starts gstudio (IRB). Docker only; nothing to gem-install:

docker run --rm -it ghcr.io/rbotafogo/galaaz-try:gstudio

Then vec = R.c(1, 2, 3); puts vec. From a checkout you can also build it locally with ./docker/try-gstudio/run.sh. This is a demo shell, not the cold-install proof.

Maintainers can prove a RubyGems install in a throwaway Ubuntu container (no git repo inside):

./docker/cold-install/run.sh published-specs

Manual

• Galaaz manual (HTML)
• Galaaz manual (PDF)
• Manual source (Markdown)

Blogs

• Oh my… Galaaz · PDF
• gKnit · PDF
• ggplot with Galaaz · PDF
• NSE and dplyr · PDF
• Ruby plot · PDF

Repository

• Source on GitHub (branch galaaz2_0)
• galaaz on RubyGems
• Examples (source)
• Specs
• Example rendered PDFs on this site

Source and runnable examples ship in the gem; knitted HTML/PDF live here. After install, use gstudio for an interactive shell.