My name is Kruno, and I am maintaining the Hunspell dictionary for Croatian spellchecking. The dictionary is now in a mature stage in terms of Hunspell, but needs refactoring to support post tags in LagugeTool so I would like to expand only spellchecker’s capabilities with LangugeTool to compensate Hunspell’s word-only checkings, not yet ready for full grammar/style checking.
Could we add support for the Croatian language to LanguageTool so that I can start developing XML rules directly through the LanguageTool platform? Of course, this process will take time, and I plan to work on it gradually, step by step. I hope that is okay.
Well, I did what I could in LibreOffice’s Lightproof. At the moment, I can’t get into the process of running LanguageTool independently because I don’t even know what ‘Maven’ is. But thanks for the link, I’ll look into it when I find some time to adapt the Hunspell dictionary to match LT and try to go through the entire process. Best regards
Take your time. You can start drafting Croatian rule patterns in simple XML locally or in the web editor while others in the community assist with the Maven project setup.
Maven can definitely be a bit of a wall when you’re just trying to get a dictionary or some basic rules integrated, especially if you’re coming from LibreOffice’s ecosystem rather than full Java development.
Don’t sweat not knowing it right away—building out the Hunspell dictionary and getting the word-lists solid is already 90% of the heavy lifting anyway. Take your time with it, and once you do get a chance to poke around the build tools, it gets a bit more intuitive. Really cool to see someone finally pushing to get proper Croatian support rolling on LT!
I managed to bring LibreOffice’s LightProof to a decent state. I’m satisfied. At first, I thought about simply adapting LightProof’s rules for LanguageTool using pure regex in XML, so that I could bring what I have there to other applications like web browsers, email clients, and the like. However, I realized that LightProof won’t work well either without dictionary tagging. I’ve done the bare minimum, but at least there is something now. So, for any serious spell/grammar checking, we need to have a tagged dictionary, which I don’t have. I will focus on that first. The only problem is that tagging and maintaining a dictionary, LightProof rules, and now trying to start with LanguageTool all at the same time is, unfortunately, more than one person can handle alone. I don’t want to take on more than I can handle. But I hope we will get there someday.