Great module! Seems the .txt is not stripped whereas .csv is when generating vocabulary name based on file name.