//*[@class="body"]//text()[name(..)!="style" and name(..)!="script"] 提取文本时不包含script标签
substring-after(//*[@class="twid"]/text()[3],"来源:") 提取/text()[3] 中来源后面的文字,substring-before用法相同
substring-before(substring-after(.//*[@class="twid"]/text(),"/"),"/") 可以嵌套使用
concat(.//*[@class="article"]/span[2]//text()," ",.//*[@class="From"]/span[3]//text()) 提取2个文本用空格拼接在一起