Android抓取CSDN首页极客头条内容 Android抓取CSDN首页极客头条内容完整实例

HttpClient httpClient = new DefaultHttpClient(); //创建一个HttpClient 
 
HttpGet httpGet = new HttpGet(“http://www.csdn.net/”); //创建一个GET请求 
 
HttpResponse response = httpClient.execute(httpGet); //发送GET请求，并响应内容 

<li><a title="宇宙员在太空中如何洗澡、睡觉、上厕所？" href="http://geek.csdn.net/news/detail/1478" rel="external nofollow" target="_blank" onclick="LogClickCount(this,363);">宇宙员在太空中如何洗澡、睡觉、上厕所？</a></li> 

<li><a title="(.*?)" href="(.*?)" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" target="_blank" onclick="LogClickCount(this,363);">\1</a></li> 

title="(.*?)" href="(.*?)" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" target="_blank" onclick="LogClickCount(this,363) 

title="(.*?)" href="(.*?)" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" .*?363 

"title=\"(.*?)\" href=\"(.*?)\".*?363" 

Pattern p = Pattern.compile("title=\"(.*?)\" href=\"(.*?)\".*?363"); 
Matcher m = p.matcher(csdnString); //csdn首页的源代码字符串 
while (m.find()) { //循环查找匹配字串 
  MatchResult mr=m.toMatchResult(); 
  Map<String, Object> map = new HashMap<String, Object>(); 
  map.put("title", mr.group(1));//找到后group(1)是表达式第一个括号的内容 
  map.put("url", mr.group(2));//group(2)是表达式第二个括号的内容 
  result.add(map); 
} 

public class MainActivity extends ListActivity {
	ListView listview;
	Handler handler;
	List<Map<String, Object>> data;
	final String CSDNURL = "http://www.csdn.net/";
	@Override 
	  protected void onCreate(Bundle savedInstanceState) {
		super.onCreate(savedInstanceState);
		handler = getHandler();
		ThreadStart();
	}
	/** 
   * 新开辟线程处理联网操作 
   */
	private void ThreadStart() {
		new Thread() {
			public void run() {
				Message msg = new Message();
				try {
					data = getCsdnNetDate();
					msg.what = data.size();
				}
				catch (Exception e) {
					e.printStackTrace();
					msg.what = -1;
				}
				handler.sendMessage(msg);
			}
		}
		.start();
	}
	/** 
   * 联网获得数据 
   * @return 数据 
   */
	private List<Map<String, Object>> getCsdnNetDate() {
		List<Map<String, Object>> result = new ArrayList<Map<String, Object>>();
		String csdnString = http_get(CSDNURL);
		//<li><a title="(.*?)" href="(.*?)" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" target="_blank" onclick="LogClickCountthis,363;">\1</a></li> 
		//title="(.*?)" href="(.*?)" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" rel="external nofollow" .*?,363\) 
		Pattern p = Pattern.compile("title=\"(.*?)\" href=\"(.*?)\".*?363");
		Matcher m = p.matcher(csdnString);
		while (m.find()) {
			MatchResult mr=m.toMatchResult();
			Map<String, Object> map = new HashMap<String, Object>();
			map.put("title", mr.group(1));
			map.put("url", mr.group(2));
			result.add(map);
		}
		return result;
	}
	/** 
   * 处理联网结果，显示在listview 
   * @return 
   */
	private Handler getHandler() {
		return new Handler(){
			public void handleMessage(Message msg) {
				if (msg.what < 0) {
					Toast.makeText(MainActivity.this, "数据获取失败", Toast.LENGTH_sHORT).show();
				} else {
					initListview();
				}
			}
		}
		;
	}
	/** 
   * 在listview里显示数据 
   * @author Lai Huan 
   * @created 2013-6-20 
   */
	private void initListview() {
		listview = getListView();
		SimpleAdapter adapter = new SimpleAdapter(this, data, 
		        android.R.layout.simple_list_item_1, new String[] { "title"}, 
		        new int[] { android.R.id.text1 });
		listview.setAdapter(adapter);
		listview.setOnItemClickListener(new OnItemClickListener() {
			@Override 
			      public void onItemClick(AdapterView<?> arg0, View arg1, int arg2, 
			          long arg3) {
				Map<String, Object> map = data.get(arg2);
				String url = (String)(map.get("url"));
				Intent intent = new Intent(Intent.ACTION_VIEW);
				intent .setData(Uri.parse(url));
				startActivity(intent);
			}
		}
		);
	}
	/** 
   * get请求URL，失败时尝试三次 
   * @param url 请求网址 
   * @return 网页内容的字符串 
   */
	private String http_get(String url) {
		final int RETRY_TIME = 3;
		HttpClient httpClient = null;
		HttpGet httpGet = null;
		String responseBody = "";
		int time = 0;
		do {
			try {
				httpClient = getHttpClient();
				httpGet = new HttpGet(url);
				HttpResponse response = httpClient.execute(httpGet);
				if (response.getStatusLine().getStatusCode() == 200) {
					//用utf-8编码转化为字符串 
					byte[] bResult = EntityUtils.toByteArray(response.getEntity());
					if (bResult != null) {
						responseBody = new String(bResult,"utf-8");
					}
				}
				break;
			}
			catch (IOException e) {
				time++;
				if (time < RETRY_TIME) {
					try {
						Thread.sleep(1000);
					}
					catch (InterruptedException e1) {
					}
					continue;
				}
				e.printStackTrace();
			}
			finally {
				httpClient = null;
			}
		}
		while (time < RETRY_TIME);
		return responseBody;
	}
	private HttpClient getHttpClient() {
		HttpParams httpParams = new BasicHttpParams();
		//设定连接超时和读取超时时间 
		HttpConnectionParams.setConnectionTimeout(httpParams, 6000);
		HttpConnectionParams.setSoTimeout(httpParams, 30000);
		return new DefaultHttpClient(httpParams);
	}
}