Welcome to Apache OpenNLP Add-ons!

Apache OpenNLP Add-ons is a set of optional components built on the Apache OpenNLP API.

Each add-on is published as its own Maven artifact, so an application can take a single integration without taking the dependencies of the others. Shared contracts stay in OpenNLP core; an implementation belongs here when it needs a distinct model format, external data, or an optional third-party library.

Useful Links

Overview

Currently, the repository has the following modules:

geoentitylinker : Links place names found in text to entries of a gazetteer index.
japanese : Feature generators and a name finder factory for Japanese text.
liblinear : LIBLINEAR based machine learning implementation.
modelbuilder : Utilities for semi-supervised generation of name finder models.
opennlp-addons-docs : The DocBook sources of the add-ons manual.

Getting Started

Depend on the module you need, not on the parent opennlp-addons POM:

Maven

<dependency>
    <groupId>org.apache.opennlp.addons</groupId>
    <artifactId>japanese</artifactId>
    <version>${opennlp-addons.version}</version>
</dependency>

Gradle

compile group: "org.apache.opennlp.addons", name: "japanese", version: "${opennlp-addons.version}"

[!NOTE] No 3.x add-ons release is published yet. Until one is, build this repository locally to use its artifacts.

Building OpenNLP Add-ons

At least JDK 21 and Maven 3.9.6 are required to build the add-ons.

After cloning the repository go into the destination directory and run:

mvn install

A verify build additionally runs Checkstyle, forbidden API checks, the RAT license header check, and a dependency license report. Use -Pjacoco for coverage.

The manual is written in DocBook XML under opennlp-addons-docs/src/docbkx and builds with the rest of the reactor.

Contributing

Every contribution is welcome, from a documentation typo fix to a new add-on. To get involved, please follow the instructions here