Mining a comparable text corpus for a Vietnamese - French statistical machine translation system