As the software is shipped to our customers, we cannot add the words manually. Searching the forum, I found a way of adding the ignore words (http://forum.languagetool.org/t/adding-additional-ignore-words/440). But I am not sure how this works server side, currently I have a map containing instances of all supported languages (i.e. AustrianGerman) and create a JLanguagetool instance per request. If I add the ignore words that way can I still cache the languages?
Is there a way to add words which should be used as suggestion during runtime?
EDIT: In the meantime I tried to add words to the dictionary at runtime. I was able do add them via hunspellDict.addWord. The word is recognized after the addition, but when I missspell it, it is not shown as a suggestion.
I searched the code for a solution and found that it would be possible to overwrite the getIngoreFileName and getSpellingFileName methods to hold the files outside the jar. Unfortunately the addIngoreWords method in the HunspellRule class uses the filename to load the contents from the jars resources directory.
Would it generally be possible to use external files for the spelling.txt and ingore.txt files? In my opinion this could be a good way to build up custom dictionaries.
Is there any better way to add words to the ignore and spelling lists during runtime? If coding would be required, I could try to supply a patch.