Some valuable BioNLP algorithms are implemented in JVM languages (e.g. SETH, Opsin). In order to make these accessible to python, we use Py4J to call them from a Kazu Step instance.

In order to make such algorithms available to Py4J, we need to compile a fatjar, which can then be placed on the classpath of the JVM process started by Py4J.

We use the gradle wrapper concept https://docs.gradle.org/current/userguide/gradle_wrapper.html

./gradlew shadowJar

Tests can be run as follows:

./gradlew test