I have developed a script that outputs rough rules.xml and disambiguation.xml files (which need manual modifications) from the data within An Gramadoir. I’m considering putting this up on Github, but I have concerns over licensing.
Most of the rules in An Gramadoir are under the GPL, which is problematic with the LGPL, although the Irish data is also under the Perl “Artistic License”. One interpretation of this could be that the XML files can be under the GPL, whereas LanguageTool is the interpreter for the purposes of the license.
Only Irish and French (already converted?) have a reasonable number of rules, although the rules may be of use for starting a currently unsupported language.
Yes, the French rules were converted in the past. Until we have support for Irish, converting these makes little sense.
We already have some dictionary data that is not under LGPL (which is weird for dictionaries anyway, as GPL), so I guess we could also have some portions in rule files that are under some other license, provided that these are appropriately attributed.
This is awesome and it is not having the relevance that it should.
Maybe it will be easier if the authors are aware of this.
If you wish, I can compile and provide to you a list of relevant author e-mails from CVS Info for project gramadoir