Multiservice

ICS PAS Multiservice is a robust linguistic Web service for Polish, combining several mature offline linguistic tools in a common online platform. Packaged TEI P5-based annotation is used as representation format for the service. Apache Thrift is used for data exchange between underlying tools.

In contrast to most frequent approaches, the architecture supports asynchronous handling of requests to enable processing large amounts of text.

Documentation

• Users guide: http://glass.ipipan.waw.pl/redmine/projects/multiserwis/wiki/Usage

• Admins guide (Polish): Multiserwis-Administracja.pdf

• Developers guide (Polish): Multiserwis-Źródła.pdf

Reports and publications

• Ogrodniczuk M., Lenart M. (2013). A multi-purpose online toolset for NLP applications. In Elisabeth Métais, Farid Meziane, Mohamed Saraee, Vijay Sugumaran, and Sunil Vadera, editors, Proceedings of the 18th International Conference on Applications of Natural Language to Information Systems, volume 7934 of Lecture Notes in Computer Science, pages 392–395. Springer-Verlag, Berlin, Heidelberg.

• Ogrodniczuk M., Lenart M. (2012). Web Service integration platform for Polish linguistic resources. In Proceedings of the Eighth International Conference on Language Resources and Evaluation, LREC 2012, pages 1164–1168, Istanbul, Turkey. ELRA.

• Ogrodniczuk M., Lenart M. (2011). Multipurpose Linguistic Web Service for Polish. In Proceedings of the Language Technology for a Multilingual Europe workshop at the German Society for Computational Linguistics and Language Technology Conference (GSCL 2011), Hamburg, Germany.

• Ogrodniczuk M. (2011). The Packaged TEI P5-based Stand-off Annotation Format. Internal description of the Multiservice format

Please cite the LREC paper:

Demo

The Web demo of the service: http://glass.ipipan.waw.pl/multiservice/

Console client

Here you can download python console client script: wsclient.py.

Requires suds python library, you can install it using:

pip install suds

Requirements

• Python 2.7

• suds python package (https://fedorahosted.org/suds/)

Invocation

Console client can be invoked as following:

python wsclient.py <Service1> <Service2> ... < input.txt > output.xml

For example:

python wsclient.py Pantera Spejd DependencyParser < input.txt > output.xml

Output format

Output format is packaged TEI P5 representation (more information about the format is available here: http://glass.ipipan.waw.pl/redmine/projects/multiserwis/wiki/InOut.