I'm trying to use the OpenNLP chunking API to chunk a portuguese sentence. So, first I tokenized a