// 深度控制的getLinks方法
public static void getLinks(MyUrl Url, int ConDepth) throws IOException {
Document doc;
// String codeUrl = "http://"
// + URLEncoder.encode(Url.getUrl().substring(7), "UTF-8");
HttpClient httpclient = new DefaultHttpClient();
System.out.print(Url);
int Depth = Url.getDepth();
String title = "k";
String time = "k";
String content = "k";
if (Url.getDepth() > ConDepth) {
} else {
try {
HttpGet httpget = new HttpGet(Url.getUrl());
HttpResponse response = httpclient.execute(httpget);
int resStatus = response.getStatusLine().getStatusCode();
if (resStatus == HttpStatus.SC_OK) {
doc = Jsoup.connect(Url.getUrl()).get();
Elements links = doc.select("a[href]");
Element Title = doc.select("title").first();
Elements context = doc.select("p");
Element Time = doc.select("div.time").first();
// List<String> Url2 = new LinkedList<String>();
content = context.text();
System.out.println(Title + content);
if (Title.text() != null && Time != null) {
title = Title.text();
time = Time.text();
}
System.out.println(Url.getUrl());
System.out.println(title + time + content + Url.getUrl());
// 组装Beans
try {
WebnewBean news = new WebnewBean(title, time, content, Url.getUrl());
boolean result = DBOperation.MyInsert(news);
} catch (Exception e) {
System.err.println(e);
}
try {
for (Element link : links) {
String linkHref = link.attr("abs:href");
MyUrl Assmble = new MyUrl();
Assmble.setUrl(linkHref);
Assmble.setDepth(Depth + 1);
// Url2.add(linkHref);
// System.out.println(linkHref);
if (Assmble.getUrl().length() > 22
&& Assmble.getUrl().substring(0, 22).equals("http://green.sohu.com/")) {
MyLinkDB.addUnVisitedUrl(Assmble);
}
}
} catch (Exception e) {
System.err.println(e);
}
}
} catch (Exception e) {
System.err.println(e);
}
}
}
