Stanford CoreNLP addon for Apache Tika's NER Parser

This project supplies necessary resources to Apache Tika's NamedEntityParser and demonstrates how to activate the NER implementation based on Stanford CoreNLP's CRF classifiers.

The usage of this addon has been documented at http://wiki.apache.org/tika/TikaAndNER#Using_Stanford_CoreNLP_NER

Requirements

• newer maven - tested on maven 3.3
• newer JDK - Tested on JDK 1.8
• (Other dependencies will be fetched from Maven repositories, requires internet)

How to Build :

• to get jar for dropping into tika's classpath
  mvn clean compile assembly:single -PtikaAddon

• To test : mvn exec:java -Dexec.args=README.md NOTE: README.md is a CLI argument

How to configure tika

1. Get the jar obtained in previous step (target/tika-ner-corenlp-addon-1.0-SNAPSHOT-jar-with-dependencies.jar) and add it to tika's classpath (requires Tika 1.12).

Alternatively, it is simple if your are using maven.

• Build and install this project to local maven repo by running mvn install on this project
• Add this dependency to your project xml <dependency> <groupId>edu.usc.ir.tika</groupId> <artifactId>tika-ner-corenlp</artifactId> <version>1.0-SNAPSHOT</version> </dependency>

2. Set system property ner.impl.class to org.apache.tika.parser.ner.corenlp.CoreNLPNERecogniser. An example usage is shown in test case NamedEntityParserTest.java

3. Activate org.apache.tika.parser.ner.NamedEntityParser. An example configuration is at src/main/resources/tika-config.xml

Contributors :

• Thamme Gowda N - tgowdan at gmail.com