This Java code gets called by the scripts in scripts-gnbr to handle the dependency parsing and formatting of dependency paths. The source is available on GitHub and can be compiled into a runnable JAR using Maven.

Regenerating GNBR requires a series of manipulations of the original PubTator data that (1) convert it to dependency paths and (2) assign theme scores to those paths. This package contains the collection of Python scripts that handle these manipulations, as well as a README that tells you the order in which to run things.