An Gramadóir

Kevin P. Scannell

Summary

This is the home page for An Gramadóir, an open source grammar checking engine. It is intended as a platform for the development of sophisticated natural language processing tools for languages with limited computational resources. It is currently implemented for the Irish language (Gaeilge); this is, to the best of my knowledge, the first grammar checker developed for any minority language. Ports for Afrikaans, Akan, Cornish, Esperanto, French, Hiligaynon, Icelandic, Igbo, Languedocien, Scottish Gaelic, Tagalog, Walloon, and Welsh are currently underway; see the Developers' Guide for more information on porting.

News

(October 2007): Diarmaid Mac Mathúnra was interviewed about Ceart and GaelSpell on the Raidió na Gaeltachta program "Glór Aniar".
(January 2007): Article about Ceart in the Galway Advertiser.
(December 2006): Kevin Donnelly has launched an online version of Welsh Gramadóir called "Klebran".
(July 2006): "Ceart", a standalone GUI version of the Irish grammar checker, has been released by Cruinneog. It is available from Litriocht.com.

Features

An Gramadóir is:

Portable. As of version 0.5 the core engine is written entirely in Perl which means it will run on just about anything that plugs in. It has been tested on a variety of platforms and operating systems: Linux (x86, ppc, amd64), *BSD (x86), Sun Solaris, DEC alpha, Mac OSX, and Windows.
Modular. The Perl module provides separate interfaces for sentence segmentation, spell checking, part-of-speech tagging, and grammar checking. These components provide a platform for the development of applications for much more complex natural language processing tasks (e.g. parsing, machine translation).
Easy to use. There is a simple command line interface that reads text from standard input and writes errors to standard output (you can see some actual output on the Sampler page). Or, you can try the software now using the Web Interface. There are also interfaces that allow the grammar checker to be called from the text editors emacs, vim, and OpenOffice. Last but not least, if you have a Mac you can try the Java front-end "Ceart" developed by Cruinneog.
Corpus-based. Various components of the engine can be bootstrapped from corpora harvested by my web crawling software An Crúbadán. This is essential for languages with severely limited resources, allowing rapid development with a minimum of effort.
Easy to develop. I've tried to design the language developers' pack so that no programming experience is needed. All Perl code is generated automatically from a number of (hopefully simple) plain text input files. This is especially important for minority languages where in many cases there is a lack of trained linguists, software developers, or both.
Scalable. With as little as an hour or two of work (editing word lists output by my web crawler) a developer can have a decent spell checking package up and running. On the other hand, the engine is flexible enough to allow for full-scale grammar checking (as evidenced by the rococo Irish version).
Language independent. (More or less). Most open source language technology is designed with (Indo-)European languages in mind. To counter this trend I've included things like full Unicode support and better support for the rich morphological phenomena found in many non-European languages.
Free Software. It is released under the GNU General Public License which means (roughly speaking) that you are free to copy, modify, or even sell this software as long as redistributed versions preserve these same freedoms. Please consider contributing to help improve later releases.

An Gramadóir

Kevin P. Scannell

Summary

Features

Other Resources