一个更加简单粗暴的爬虫-必应今日美图

 1 import com.gargoylesoftware.htmlunit.FailingHttpStatusCodeException;
 2 import com.gargoylesoftware.htmlunit.WebClient;
 3 import com.gargoylesoftware.htmlunit.html.HtmlAnchor;
 4 import com.gargoylesoftware.htmlunit.html.HtmlPage;
 5 import org.apache.commons.io.FileUtils;
 6 
 7 import java.io.File;
 8 import java.io.IOException;
 9 import java.io.InputStream;
10 import java.net.MalformedURLException;
11 
12 /**
13  * <pre>
14  * class intro
15  * </pre>
16  * <p>
17  * Created by fujun on 2017/9/29.
18  */
19 public class BingPhoto {
20     public static void main(String[] args) {
21         WebClient webClient = new WebClient();
22         webClient.getOptions().setCssEnabled(false);
23         webClient.getOptions().setJavaScriptEnabled(false);
24         for (int i = 1; i < 491; i++) {
25             String url = "http://bing.plmeizi.com/show/" + i;
26             try {
27                 HtmlPage page = webClient.getPage(url);
28                 HtmlAnchor a = (HtmlAnchor) page.getElementById("picurl");
29                 String picUrl = a.getHrefAttribute();
30                 if (picUrl.contains("jpg")) {
31                     InputStream inputStream = webClient.getPage(picUrl).getWebResponse().getContentAsStream();
32                     FileUtils.copyToFile(inputStream, new File("E://bingpic//" + i + ".jpg"));
33                     System.out.println("SUCCEED:" + url);
34                 } else {
35                     System.out.println("FAILED:" + url);
36                 }
37 
38             } catch (FailingHttpStatusCodeException e) {
39                 // TODO Auto-generated catch block
40                 e.printStackTrace();
41             } catch (MalformedURLException e) {
42                 // TODO Auto-generated catch block
43                 e.printStackTrace();
44             } catch (IOException e) {
45                 // TODO Auto-generated catch block
46                 e.printStackTrace();
47             } finally {
48                 webClient.close(); // 关闭客户端，释放内存
49             }
50         }
51     }
52 }