Could you lead me to the java class in which is responsible for Khmer tokenizer? I have the same issues. LanguageTool detects handful errors on correct text. At the moment, the string tokenizer for Khmer confused incomplete sentences or a cluster of vocabularies with missing characters as a token.
