View Javadoc
1   package us.codecraft.webmagic.samples;
2   
3   import us.codecraft.webmagic.Page;
4   import us.codecraft.webmagic.Site;
5   import us.codecraft.webmagic.Spider;
6   import us.codecraft.webmagic.processor.PageProcessor;
7   
8   /**
9    * @author code4crafter@gmail.com <br>
10   * Date: 13-5-20
11   * Time: 下午5:31
12   */
13  public class KaichibaProcessor implements PageProcessor {
14      @Override
15      public void process(Page page) {
16          //http://progressdaily.diandian.com/post/2013-01-24/40046867275
17          int i = Integer.valueOf(page.getUrl().regex("shop/(\\d+)").toString()) + 1;
18          page.addTargetRequest("http://kaichiba.com/shop/" + i);
19          page.putField("title",page.getHtml().xpath("//Title"));
20          page.putField("items", page.getHtml().xpath("//li[@class=\"foodTitle\"]").replace("^\\s+", "").replace("\\s+$", "").replace("<span>.*?</span>", ""));
21      }
22  
23      @Override
24      public Site getSite() {
25          return Site.me().setDomain("kaichiba.com").setCharset("utf-8").
26                  setUserAgent("Mozilla/5.0 (Macintosh; Intel Mac OS X 10_7_2) AppleWebKit/537.31 (KHTML, like Gecko) Chrome/26.0.1410.65 Safari/537.31");
27      }
28  
29      public static void main(String[] args) {
30          Spider.create(new KaichibaProcessor()).addUrl("http://kaichiba.com/shop/41725781").run();
31      }
32  }