hello-world-aggregated-pear

An Apache UIMA Hello World Annotator aggregating two annotators bundled as a PEAR package.

A PEAR package is a way of packaging text analysis components and their resources for use with the Apache UIMA framework. Every PEAR can contain additional Java libraries and UIMA knows how to keep PEARs separate from each other such that there are no problems if two different PEARs use incompatible versions of some library.

This project provides a minimal example how to build an aggregated analysis engine and how to bundle it as a self contained UIMA PEAR package.

Build prerequisites

• JDK 1.8 or later
• Maven

Build instructions

mvn clean install

The Maven build process performs of the following steps:

• Compile the two UIMA components SimpleSentenceAnnotator and SimpleTokenAnnotator` included in this project.
• Generate a UIMA aggregate analysis engine descriptor (pipeline) using the GenerateAggregateAnalysisDescriptor class included in this project. This pipeline becomes the top-level analysis component of the PEAR - i.e. what is executed when the PEAR is invoked by UIMA.
• Uses the Apache UIMA PearPackagingMavenPlugin to generate a PEAR file consisting of the two UIMA components included in this project, all required dependencies (in particular the Averbis type system which is used by the components) and the UIMA aggregate analysis engine descriptor generated in the previous step. The file is written to the target folder.
• Executes the PearPackageIntegrationTest which uses Apache UIMA which temporarily installs the PEAR file, uses it to process a sample document, and verifies that the expected Token and Sentence annotations have been created.