import java.io.FileOutputStream;
import java.io.InputStream;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import org.apache.commons.io.IOUtils;
import com.gargoylesoftware.htmlunit.Page;
import com.gargoylesoftware.htmlunit.WebClient;
public class DownloadFile {
public static void main(String[] args) throws Exception {
String baseUrl = "<a href="http://hanyu.iciba.com/hanzi/1.shtml";" target="_blank">http://hanyu.iciba.com/hanzi/1.shtml";</a>
String bihuaRegex = "class=\"guanggao\"[^<]*<[^<]*<param\\s*name=\"movie\"\\s*value=\"([^\"]*)";
String aSoundRegex = "class=\"js12\">ā.*?name=\"FlashVars\"\\s*value=\"f=([^\"]*)";
String eSoundRegex = "class=\"js12\">ē.*?name=\"FlashVars\"\\s*value=\"f=([^\"]*)";
WebClient client = new WebClient();
client.getOptions().setCssEnabled(false);
client.getOptions().setJavaScriptEnabled(false);
client.getOptions().setThrowExceptionOnFailingStatusCode(false);
client.getOptions().setThrowExceptionOnScriptError(false);
Page page = client.getPage(baseUrl);
String source = page.getWebResponse().getContentAsString();
Matcher mBihuan = Regex(source, bihuaRegex);
Matcher mA = Regex(source, aSoundRegex);
Matcher mE = Regex(source, eSoundRegex);
while(mBihuan.find()) {
String url = "<a href="http://hanyu.iciba.com/" + mBihuan.group" target="_blank">http://hanyu.iciba.com/" + mBihuan.group</a>(1);
page = client.getPage(url);
saveFile(page, "d:/testDownload/bihua.swf");
}
while(mA.find()) {
String url = mA.group(1);
page = client.getPage(url);
saveFile(page, "d:/testDownload/a.mp3");
}
while(mE.find()) {
String url = mE.group(1);
page = client.getPage(url);
saveFile(page, "d:/testDownload/e.mp3");
}
}
public static Matcher Regex(String source, String regex) {
Pattern p = Pattern.compile(regex, Pattern.DOTALL);
return p.matcher(source);
}
public static void saveFile(Page page, String file) throws Exception {
InputStream is = page.getWebResponse().getContentAsStream();
FileOutputStream output = new FileOutputStream(file);
IOUtils.copy(is, output);
output.close();
}
}
利用htmlunit下载网页上的文件(未测试)
最新推荐文章于 2024-04-08 14:34:05 发布