if we were to extract possible tags based on the content