xtas

• easy access to numerous text processing and analysis tools
• full support for Dutch and English
• can use Elasticsearch for document storage
• can be run as a service

xtas is a collection of natural language processing and text mining tools, brought together in a single software package with built-in distributed computing and support for the Elasticsearch document store.

xtas functionality consists partly of wrappers for existing packages, with automatic installation of software and data; and partly of custom-built modules coming out of research. Currently offered are various parsers for Dutch and English (Alpino, CoreNLP, Frog, Semafor), named entity recognizers (Frog, Stanford and custom-built ones), a temporal expression tagger (Heideltime) and a sentiment tagger based on SentiWords.

A basic installation of xtas works like a Python module. Built-in package management and a simple, uniform interface take away the hassle of installing, configuring and using many existing NLP tools.

xtas’s open architecture makes it possible to include custom code, run this in a distributed fashion and have it communicate with Elasticsearch to provide document storage and retrieval.