新手问题 nokogiri 如何解析这样的片段

suupic · October 23, 2012 · Last by hexawing replied at May 05, 2014 · 8080 hits

想爬一点微博的数据,发现在未登录状态,虽然可以浏览,但微博正文数据是放在 script 标签里的,不知怎样转换才能得到正确的中文?

<script>STK && STK.pageletM && STK.pageletM.view({"pid":"pl_content_weiboDetail","js":[],"css":["style\/css\/module\/global\/person_info_big.css?version=f05544dd408986d7","style\/css\/module\/list\/feed.css?version=f05544dd408986d7","style\/css\/module\/forms\/feed_repeat.css?version=f05544dd408986d7","style\/css\/module\/list\/comment_list.css?version=f05544dd408986d7","style\/css\/module\/global\/pages.css?version=f05544dd408986d7","style\/css\/module\/tab\/tab_c.css?version=f05544dd408986d7","style\/css\/module\/layer\/layer_faces.css?version=f05544dd408986d7","style\/css\/module\/layer\/layer_addfavor_tags.css?version=f05544dd408986d7","style\/css\/module\/layer\/layer_forward.css?version=f05544dd408986d7","style\/css\/module\/layer\/layer_menu_list.css?version=f05544dd408986d7","style\/css\/module\/tab\/tab_b.css?version=f05544dd408986d7","style\/css\/module\/layer\/layer_send_pic.css?version=f05544dd408986d7"],"html":"<div node-type=\"weibo_info\" class=\"feed_lists W_linka W_texta\" action-data=\"&ispower=1\">\n\t<dl class=\"feed_list clearfix feed_list_hover W_no_border\" mid=\"3502493067895526\" >\n\t<dd class=\"content\">\n\t<p><em nick-name=\"\u5415\u6674\u6625-\">\u6709\u505a\u5efa\u7b51\u8bbe\u8ba1\u65b9\u9762\u7684\u670b\u53cb\u4e48\uff1f\u8bf7\u5404\u4f4d\u4eb2\u4eec\u5e2e\u624b\u4ecb\u7ecd\u4ecb\u7ecd\uff0c<img src=\"http:\/\/img.t.sinajs.cn\/t35\/style\/images\/common\/face\/ext\/normal\/c3\/zy_org.gif\" title=\"[\u6324\u773c]\" alt=\"[\u6324\u773c]\" type=\"face\" \/><\/em><\/p>\n<!--pic-->\n\t<!--\/pic-->\n<!--retweeted-->\n\t<!--\/retweeted-->\n\t<div class=\"wTablist W_linkb W_textb\" node-type =\"feed_list_tagList\" style=\"display:none;\">\u6807\u7b7e\uff1a\n\t  \t  \t  \n\t  <\/div>\n\t  \t\t<p class=\"info W_linkb W_textb\">\n\t\t<span>\t\n\t\t\t  <a href=\"javascript:void(0);\" action-type=\"login\">\u8f6c\u53d1(10)<\/a><i class=\"W_vline\">|<\/i>\n\t\t\t<em class=\"hover\">\n\t\t\t<a href=\"javascript:void(0);\" action-type=\"login\">\u6536\u85cf<\/a><i class=\"W_vline\">|<\/i>\n\t<\/em>\t\n\t<a href=\"javascript:void(0);\" action-type=\"login\">\u8bc4\u8bba(5)<\/a>\t\t\n\t\t<\/span>\n\t\t\u4eca\u5929 14:48 \u6765\u81ea<a target=\"_blank\" href=\"http:\/\/se.360.cn\/?fromweibo\" rel=\"nofollow\">360\u5b89\u5168\u6d4f\u89c8\u5668<\/a>\n\t\t\t<em class=\"hover\"><i class=\"W_vline\">|<\/i><a href=\"javascript:void(0);\" action-type=\"login\">\u4e3e\u62a5<\/a><\/em>\n\t\t<\/p>\n\t<\/dd>\n\t<\/dl>\t\n<\/div>\n<div class=\"unlogin_vip_info\">\n<span class=\"icon_warn\"><\/span>\n<a target=\"_blank\" href=\"http:\/\/weibo.com\/signup\/signup.php?inviteCode=1832604244&entry=weiyonghu\" >\u5feb\u901f\u5f00\u901a\u5fae\u535a<\/a>\u4f60\u53ef\u4ee5\u67e5\u770b\u66f4\u591a\u5185\u5bb9\uff0c\u8fd8\u53ef\u4ee5\u8bc4\u8bba\u3001\u8f6c\u53d1\u5fae\u535a\u3002\n<\/div>\n"})</script>

需要提取的部分在 content 中,如下:

<dd class=\"content\">\n\t<p><em nick-name=\"\u5415\u6674\u6625-\">\u6709\u505a\u5efa\u7b51\u8bbe\u8ba1\u65b9\u9762\u7684\u670b\u53cb\u4e48\uff1f\u8bf7\u5404\u4f4d\u4eb2\u4eec\u5e2e\u624b\u4ecb\u7ecd\u4ecb\u7ecd\uff0c<img src=\"http:\/\/img.t.sinajs.cn\/t35\/style\/images\/common\/face\/ext\/normal\/c3\/zy_org.gif\" title=\"[\u6324\u773c]\" alt=\"[\u6324\u773c]\" type=\"face\" \/><\/em>

这个 Nokogiri 搞不了,用正则

@suupic 要爬的链接发来看看

#1 楼 @ywencn 用 Nokogiri 很简单啊

dom.at('script dd.content').content ????

#2 楼 @zfjoy520 就是从每条微博左下角发表时间的链接点进去,比如http://weibo.com/1832604244/z12R4x7Zc

#5 楼 @suupic 你可以换个思路。抓微博手机版的,没那么复杂。如:http://weibo.cn/1832604244/z12R4x7Zc 把 weibo.com 换成 weibo.cn 其他不变。

#5 楼 @suupic 我这边尝试了下 weibo.cn 抓出来很容易的。

#7 楼 @zfjoy520 哈,多谢指点,这样简单多了

#8 楼 @suupic 一直在玩 weibo.cn。你仔细研究,你会发现微博手机版带给我们带来了很多惊喜。你懂的。

source = Nokogiri::HTML(open(#{uri})) source.css('class=\"content\"').each do |class_content| puts class_content.content.lstrip end

这样可以吗?

#6 楼 @zfjoy520 奇怪,我点你的链接就可以看到内容,点我自己的围脖换成.cn 就要我登录……

#11 楼 @hexawing 。。。这是挖墓啊。

呃,求知嘛,挖一下不算错吧……

You need to Sign in before reply, if you don't have an account, please Sign up first.