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