Elasticsearch NLP Plugin

An Elasticsearch Plugin that Integrates NLP Capabilities into IR Scoring Algorithms

Pre-requisites

This plugin requires Java 8+

This plugin currently supports and requires Elasticsearch 7.3.0.

Due to fundamental changes in underlying lucene data structures between versions, it is currently unfeasible to support multiple versions of ES, although that may change in the future.

For the purposes of backreferencing, the supported Elasticsearch versions of each major release are listed below, in reverse chronological order, although features are not backported between versions

• v1.0.2 - Elasticsearch 7.3.0
• v1.0.1 - Elasticsearch 7.3.0
• v1.0.0 - Elasticsearch 6.6.0

Installation

1. Please direct yourself to the releases page and download the ES-NLP-PLUGIN.zip for your desired release.
2. Follow the installation steps for installing elasticsearch plugins from filesystem here, directing it at the downloaded zip file
3. Copy config.yml in the zip to <your_es_directory>/config/elasticsearch-nlp-plugin.yml and edit appropriately
4. Documents need to be indexed using the nlp tokenizer

For developers

Making customizations to this codebase is fairly straightforward - you will need JDK 8+, Apache Maven, and Apache Ant

1. First, clone this repository and make your desired changes.
2. Run mvn clean install to generate the plugin JAR
3. Run ant dist to generate the distribution zip that is consumed by Elasticsearch's plugin installer
4. Follow the installation steps for installing elasticsearch plugins from filesystem here, directing it at the downloaded zip file

Reference

A. Wen, Y. Wang, V. C. Kaggal, S. Liu, H. Liu and J. Fan, "Enhancing Clinical Information Retrieval through Context-Aware Queries and Indices," 2019 IEEE International Conference on Big Data (Big Data), Los Angeles, CA, USA, 2019, pp. 2800-2807, doi: 10.1109/BigData47090.2019.9006241.