Support for Welsh language spell / grammar checking

Hi there,

I am looking for ways to support the Welsh language within an implementation of LanguageTools.

Can you help by providing advice on:

Many thanks

Darren Collins

If you’re only interested in spell checking, you could try something like this (example for Norwegian) in the properties file that is specified using --config:

lang-no=Norwegian (Bokmal)
lang-no-dictPath=/home/languagetool/dynamic-languages/no/nb_NO.dict

.dict is a Morfologik-based dictionary or a hunspell .dic file.

If you intend to use the browser add-on, the problem will be that Welsh will not be in the list of languages, though. So this is only helpful if you intend to use the API.

Hi Daniel,

Many thanks for the suggestion. I will pass this on to our development teams for feedback.

Regards

Darren

Ideally you need to find a Welsh dictionary with grammatical tags i.e tag words to indicate if they are nouns, adjectives, whether they are singular, plural, tag inflected words such as conjugations of verbs but also conjugation of preposition (which happens in Welsh).

Several years ago, I added Breton to LanguageTool. Breton and Welsh have similarities (e.g both have mutations, conjugated prepositions…). Both are in the same family of languages i.e. celtic → Insular celtic → Brittonic. For Breton, I found grammatical tags in the files of the Apertium French <-> Breton translator. For Welsh, I just did a quick Google search and I see a Welsh dictionary with grammatical tags at http://www.eurfa.org.uk/ which may be suitable as input of a script to import in LanguageTool

I started working on rules for Welsh in my own branch some 6 years ago, but hit a bit of a barrier with which word dictionary to use, and converting Constraint Grammar into languagetool rules.

I did write a converter for An Gramadoir rules for LanguageTool, mainly for converting the Welsh language rules, but due to concerns with which dictionary should be used, I put this on hold whilst awaiting output from the CorCenCC project.

It would be awesome if this could be picked back up!