public String getParsedPage() {
List<String> alertHandler = new LinkedList<String>();
;
WebClient webClient = new WebClient(BrowserVersion.FIREFOX_24); // CHROME);
webClient.setAjaxController(new MyNicelyResynchronizingAjaxController());
webClient.getOptions().setJavaScriptEnabled(true);
webClient.getOptions().setTimeout(3500);
webClient.getOptions().setThrowExceptionOnScriptError(true);
webClient.getOptions().setCssEnabled(true);
webClient.getOptions().isRedirectEnabled();
webClient.setAlertHandler(
new CollectingAlertHandler(alertHandler)); // 将JavaScript中alert标签产生的数据保存在一个链表中
// webClient.getOptions().setThrowExceptionOnScriptError(false);
HtmlPage page = null;
JavaScriptEngine engine = new JavaScriptEngine(webClient);
webClient.setJavaScriptEngine(engine);
try {
page = webClient.getPage(data);
} catch (FailingHttpStatusCodeException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (MalformedURLException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
}
if (page != null) {
return page.asXml();
}
return null;
}
/**
* get the ajax url from the click button
*
* @param clickOfXpath:页面待点击按钮的xpath表达式
* @param index
* @return List<String>:链表的第一个信息是页面的title，以后的信息是所有的ajax的url
*/
public static List<String> getAjaxUrl(String targetUrl, String clickOfXpath, int index)
throws FailingHttpStatusCodeException, MalformedURLException, IOException {
// TARGET_URL =
// "http://app.flyme.cn/apps/public/detail?package_name=com.myzaker.zaker_phone_smartbar";
List<String> urls = new LinkedList<String>();
// 每次ajax请求时都会创建一个AjaxController对象，在该对象中可以查看ajax请求的地址
MyNicelyResynchronizingAjaxController ajaxController =
new MyNicelyResynchronizingAjaxController();
List alertHandler = new LinkedList();
// 模拟一个浏览器
WebClient webClient = new WebClient(BrowserVersion.FIREFOX_24);
// HtmlUnitDriver
// 设置webClient的相关参数
webClient.getOptions().setJavaScriptEnabled(true);
webClient.getOptions().setCssEnabled(false);
webClient.setAjaxController(ajaxController);
webClient.getOptions().setTimeout(35000);
webClient.getOptions().setThrowExceptionOnScriptError(false);
webClient.setAlertHandler(
new CollectingAlertHandler(alertHandler)); // 将JavaScript中alert标签产生的数据保存在一个链表中
// 模拟浏览器打开一个目标网址
HtmlPage rootPage = webClient.getPage(targetUrl);
urls.add(rootPage.getTitleText());
urls.add(ajaxController.getVisitUrl());
// System.out.println("url1:" + url);
HtmlElement elementA = (HtmlElement) rootPage.getByXPath(clickOfXpath).get(index);
Page page = elementA.click();
urls.add(ajaxController.getVisitUrl());
return urls;
}
/**
* @return
* @author douzh
* @time 2015-4-23下午4:36:05
*/
public static WebClient getClient() {
WebClient client = new WebClient(BrowserVersion.FIREFOX_24);
client.getOptions().setJavaScriptEnabled(true);
client.getOptions().setActiveXNative(false);
client.getOptions().setCssEnabled(false);
client.getOptions().setRedirectEnabled(true);
client.getOptions().setThrowExceptionOnScriptError(false);
client.getOptions().setThrowExceptionOnFailingStatusCode(false);
client.getOptions().setGeolocationEnabled(true);
// client.addWebWindowListener(new WebWindowListener() {
// public void webWindowOpened(WebWindowEvent event) {
// System.out.println("Web Window Openning");
// }
//
// public void webWindowContentChanged(WebWindowEvent event) {
// System.out.println("Web Content Changed");
// }
//
// public void webWindowClosed(WebWindowEvent event) {
// System.out.println("Web Window Closed");
// }
// });
client.setAjaxController(
new NicelyResynchronizingAjaxController() {
public boolean processSynchron(HtmlPage page, WebRequest settings, boolean async) {
System.out.println(settings.getUrl());
return super.processSynchron(page, settings, async);
}
});
return client;
}
public static void main(String[] args) throws Exception {
// WebClient webClient = new WebClient(BrowserVersion.FIREFOX_24, "54.186.230.121", 3128);
WebClient webClient = new WebClient(BrowserVersion.FIREFOX_24);
webClient.getOptions().setThrowExceptionOnScriptError(false);
webClient.setJavaScriptTimeout(10000);
webClient.getOptions().setJavaScriptEnabled(true);
webClient.setAjaxController(new NicelyResynchronizingAjaxController());
webClient.getOptions().setTimeout(10000);
// webClient.getOptions().setJavaScriptEnabled(false);
// webClient.getOptions().setAppletEnabled(false);
// webClient.getOptions().setCssEnabled(false);
// webClient.getOptions().setThrowExceptionOnScriptError(false);
// webClient.setJavaScriptTimeout(10000);
// webClient.getOptions().setJavaScriptEnabled(true);
// webClient.setAjaxController(new NicelyResynchronizingAjaxController());
// webClient.getOptions().setTimeout(10000);
// webClient.getOptions().setThrowExceptionOnFailingStatusCode(false);
// webClient.getOptions().setThrowExceptionOnScriptError(false);
HtmlPage currentPage = webClient.getPage("http://www.yandex.ru/");
// HtmlPage currentPage = webClient.getPage("http://www.google.ru");
// HtmlDivision div = currentPage.getHtmlElementById("del_competitors-1_42");
// HtmlElement clickable = (HtmlElement)
// currentPage.getHtmlElementById("del_competitors-1_42");
// currentPage = (HtmlPage) clickable.click();
HtmlAnchor advancedSearchAn = currentPage.getAnchorByText("Завести ящик");
currentPage = advancedSearchAn.click();
HtmlImage image = currentPage.<HtmlImage>getFirstByXPath("//img[@src='images/ash2008.jpg']");
currentPage = (HtmlPage) image.click();
System.out.println(currentPage.asXml());
// HtmlImage image =
// currentPage.<HtmlImage>getFirstByXPath("//img[@src='images/ash2008.jpg']");
// currentPage = (HtmlPage) image.click();
// HtmlImage imagetosave =
// currentPage.<HtmlImage>getFirstByXPath("//img[@src='//yastatic.net/www/1.977/yaru/i/logo.png']");
// HtmlImage image = currentPage.<HtmlImage>getHtmlElementById("add_competitors-1_3");
// currentPage = (HtmlPage) image.click();
// File imageFile = new File("test_new.jpg");
// image.saveAs(imageFile);
// System.out.println(currentPage.asXml());
System.out.println("It is done.");
webClient.closeAllWindows();
}
public static void homePage(String url) {
String str;
// 创建一个webclient
WebClient webClient = new WebClient();
// webClient.getWebConsole().setLogger(null);
// //htmlunit 对css和javascript的支持不好，所以请关闭之
webClient.getOptions().setCssEnabled(false);
// webClient.getOptions().setUseInsecureSSL(true);
webClient.getOptions().setThrowExceptionOnFailingStatusCode(false);
// webClient.getCookieManager().setCookiesEnabled(true);
webClient.setAjaxController(new NicelyResynchronizingAjaxController());
webClient.getOptions().setThrowExceptionOnScriptError(false);
webClient.waitForBackgroundJavaScript(1000);
webClient.waitForBackgroundJavaScriptStartingBefore(1000);
// 获取页面
HtmlPage page = null;
int times = 10;
try {
page = webClient.getPage(url);
for (int i = 0; i < times; i++) {
synchronized (page) {
page.wait(2000);
}
if (page.getByXPath("//object") != null) {
System.out.println("***************************yes**********************");
for (int j = 0; j < page.getByXPath("//object").size(); j++) {
System.out.println(page.getByXPath("//object").get(j));
}
break;
}
System.out.println("***************************no***************************");
}
} catch (Exception e) {
e.printStackTrace();
}
// 获取页面的TITLE
// str = page.getTitleText();
// System.out.println(str);
// //获取页面的XML代码
// System.out.println("***************************start***************************");
// System.out.println(page.asXml());
// System.out.println("***************************end***************************");
// //获取页面的文本
// str = page.asText();
// System.out.println(str);
}
/**
* Static method for <code>Login</code>. Uses dependencies <b>HttpUnit</b> in connecting to Keats.
*
* @see com.gargoylesoftware.htmlunit
* @param parent The parent window. instanceof<code>Scrape</code>, to set relative locations to
* @param link The url of which to retrieve information from.
* @param username The username to log in KEATS with.
* @param password The password to log in KEATS with.
* @return results.asText() Returns the content of the url if login successful. Returns null
* otherwise.
*/
public static String login(Scrape parent, String link, String username, String password) {
try {
WebClient client = new WebClient();
// Settings
client.getOptions().setThrowExceptionOnScriptError(false);
client.getOptions().setThrowExceptionOnScriptError(false);
client.getOptions().setThrowExceptionOnFailingStatusCode(false);
client.getOptions().setJavaScriptEnabled(false);
client.getOptions().setCssEnabled(false);
client.getOptions().setRedirectEnabled(true);
client.getOptions().setUseInsecureSSL(true);
client.getCookieManager().setCookiesEnabled(true);
HtmlPage page = client.getPage("https://login-keats.kcl.ac.uk/");
HtmlForm form =
page.getFirstByXPath("//form[@action='https://keats.kcl.ac.uk/login/index.php']");
HtmlInput usernameInput = form.getInputByName("username");
usernameInput.setValueAttribute(username);
HtmlInput passwordInput = form.getInputByName("password");
passwordInput.setValueAttribute(password);
page = form.getInputByValue("Log in").click();
HtmlPage results = client.getPage(link);
client.closeAllWindows();
return results.asText();
} catch (MalformedURLException e) {
JOptionPane.showMessageDialog(
parent,
"The URL you have provided is not recognised. Please double check your "
+ "input and try again.",
"MalformedURLException found.",
JOptionPane.ERROR_MESSAGE);
e.printStackTrace();
} catch (IOException e) {
JOptionPane.showMessageDialog(
parent,
"An error has occurred when attempting to read information from the "
+ "server. The input could be interrupted by external processes.",
"IOException found.",
JOptionPane.ERROR_MESSAGE);
e.printStackTrace();
}
return null;
}
/**
* 获取速度最快的浏览器
*
* @return
*/
public WebClient getFastWebClient() {
// TODO 这里可以启动webclient的ajax功能，读取技术指标的变化
webClient.getOptions().setCssEnabled(false); // if you don't need css
webClient.getOptions().setJavaScriptEnabled(false); // if you don't need js
webClient.getOptions().setActiveXNative(false);
webClient.getOptions().setAppletEnabled(false);
webClient.getOptions().setGeolocationEnabled(false);
// webClient.getOptions().setDoNotTrackEnabled(true); //不追踪隐私
webClient.getOptions().setThrowExceptionOnFailingStatusCode(false); // 页面有错，不抛出异常
webClient.getOptions().setPrintContentOnFailingStatusCode(false); // 页面有错，不打印
return webClient;
}
@Before
public void setUp() {
webUrl = System.getProperty("integration.url");
webClient = new WebClient();
webClient.getOptions().setJavaScriptEnabled(true);
webClient.setJavaScriptTimeout(60000);
}
@Before
public void setUp() {
ProxyConfig proxyConfig = new ProxyConfig("localhost", proxyPort);
proxyConfig.addHostsToProxyBypass("127.0.0.1");
webClient.getOptions().setProxyConfig(proxyConfig);
webClient.getOptions().setTimeout(2000);
}
@Test
@Ignore
public void test3() throws Exception {
System.out.println("-------------------------------");
WebClient webClient = new WebClient(BrowserVersion.CHROME);
webClient.getOptions().setCssEnabled(false);
webClient.getOptions().setJavaScriptEnabled(false);
HtmlPage page =
webClient.getPage("http://www.zjnu.edu.cn/news/common/article_show.aspx?article_id=19285");
System.out.println("---------------标题----------------");
HtmlSpan span1 = (HtmlSpan) page.getElementById("mytitle");
System.out.println(span1.asText());
System.out.println("-------------------------------");
System.out.println("---------------正文----------------");
HtmlSpan span2 = (HtmlSpan) page.getElementById("mycontent");
System.out.println(span2.asText());
System.out.println("-------------------------------");
System.out.println("---------------图片----------------");
DomNodeList<HtmlElement> elements = span2.getElementsByTagName("img");
for (HtmlElement element : elements) {
System.out.println(element.getAttribute("src"));
}
// log.debug("{}", elements);
System.out.println("-------------------------------");
webClient.close();
System.out.println("-------------------------------");
}
public static HtmlPage getPage(String url)
throws FailingHttpStatusCodeException, MalformedURLException, IOException {
WebClient wc = new WebClient(BrowserVersion.CHROME);
wc.getOptions().setCssEnabled(false);
HtmlPage page = wc.getPage(url);
return page;
}
public static ArrayList<String> trans(String url)
throws FailingHttpStatusCodeException, MalformedURLException, IOException {
ArrayList<String> hrefList = new ArrayList<String>();
WebClient webClient = new WebClient(BrowserVersion.CHROME);
webClient.getOptions().setJavaScriptEnabled(false);
webClient.getOptions().setCssEnabled(false);
try {
HtmlPage page = null;
try {
page = (HtmlPage) webClient.getPage(url);
} catch (ConnectException e) {
System.out.println("Connect fails here:" + e.getMessage());
}
InputStream temp = new ByteArrayInputStream(page.asText().getBytes());
InputStreamReader isr = new InputStreamReader(temp);
BufferedReader br = new BufferedReader(isr);
String str = null, rs = null;
while ((str = br.readLine()) != null) {
rs = str;
// System.out.println(rs);
if (rs != null) hrefList.add(rs);
}
System.out.println("从该网址查找的可能相关文本如下：");
for (int i = 0; i < hrefList.size(); i++) {
String string = hrefList.get(i);
string = getTextFromHtml(string);
if (string.length() >= 30) System.out.println("------" + i + ":" + string);
}
} catch (IOException e) {
}
return hrefList;
}
private HtmlPage getPage(String url) throws IOException {
final WebClient webClient =
new WebClient(
getBrowserVersionFromName(
searchEngine.getDefaultBrowser())); // BrowserVersion.FIREFOX_24);
webClient.getOptions().setJavaScriptEnabled(false);
return webClient.getPage(url);
}
@Test
@Ignore
public void test01() throws Exception {
System.out.println("-------------------------------");
WebClient webClient = new WebClient(BrowserVersion.CHROME);
try {
WebClientOptions options = webClient.getOptions();
options.setThrowExceptionOnFailingStatusCode(false);
options.setThrowExceptionOnScriptError(false);
options.setCssEnabled(false);
options.setJavaScriptEnabled(true);
options.setTimeout(50000);
// webClient.setAjaxController(new NicelyResynchronizingAjaxController());
HtmlPage pageOrgin = webClient.getPage("http://lvyou.baidu.com/jinhua/jingdian");
Thread.sleep(5000);
DomNodeList<DomNode> pageNodes = pageOrgin.querySelectorAll(".pagination");
HtmlDivision pageDiv = (HtmlDivision) pageNodes.get(0);
DomNodeList<HtmlElement> liElements = pageDiv.getElementsByTagName("li");
Integer pageSize = liElements.size() - 1;
for (int pageNow = 0; pageNow < pageSize; pageNow++) {
DomNodeList<HtmlElement> pageAnchors = liElements.get(pageNow).getElementsByTagName("a");
HtmlAnchor pageAnchor = (HtmlAnchor) pageAnchors.get(0);
HtmlPage page = pageAnchor.click();
Thread.sleep(10000);
DomElement jViewDom = page.getElementById("J-view-list-container");
DomNodeList<HtmlElement> lis = jViewDom.getElementsByTagName("li");
for (HtmlElement li : lis) {
DomNodeList<DomNode> titleNodes = li.querySelectorAll(".title");
HtmlAnchor titleAnchor = (HtmlAnchor) titleNodes.get(0);
System.out.println("---------------标题----------------");
log.debug("{}", titleAnchor.asText());
DomNodeList<DomNode> picNodes = li.querySelectorAll(".pic");
HtmlAnchor picAnchor = (HtmlAnchor) picNodes.get(0);
System.out.println("---------------详情URL----------------");
String detailUrl = "http://lvyou.baidu.com" + picAnchor.getAttribute("href");
log.debug("{}", detailUrl);
DomNodeList<HtmlElement> imgEelements = picAnchor.getElementsByTagName("img");
for (HtmlElement imgEelement : imgEelements) {
System.out.println("---------------图片----------------");
log.debug("{}", imgEelement.getAttribute("src"));
}
DomNodeList<DomNode> sumNodes = li.querySelectorAll(".view-userSays");
HtmlDivision sumDiv = (HtmlDivision) sumNodes.get(0);
DomNodeList<HtmlElement> sumElements = sumDiv.getElementsByTagName("p");
HtmlParagraph sumPara = (HtmlParagraph) sumElements.get(0);
System.out.println("---------------摘要----------------");
log.debug("{}", sumPara.asText());
}
}
} finally {
webClient.close();
}
System.out.println("-------------------------------");
}
@JsfTest(JsfVersion.JSF_2_2_5)
@Test
public void testSetForEach() throws Exception {
webClient.getOptions().setThrowExceptionOnFailingStatusCode(false);
HtmlPage page = webClient.getPage(webUrl + "faces/setForEach.xhtml");
assertEquals(200, page.getWebResponse().getStatusCode());
assertTrue(page.asXml().indexOf("foo") == -1);
assertTrue(page.asXml().indexOf("bar") != -1);
assertTrue(page.asXml().indexOf("baz") != -1);
}
public static void main(String[] args) throws Exception {
LogFactory.getFactory()
.setAttribute("org.apache.commons.logging.Log", "org.apache.commons.logging.impl.NoOpLog");
java.util.logging.Logger.getLogger("com.gargoylesoftware.htmlunit").setLevel(Level.OFF);
java.util.logging.Logger.getLogger("org.apache.commons.httpclient").setLevel(Level.OFF);
WebClient client = new WebClient();
client.getOptions().setCssEnabled(false);
client.getOptions().setJavaScriptEnabled(false);
client.getOptions().setThrowExceptionOnFailingStatusCode(false);
client.getOptions().setThrowExceptionOnScriptError(false);
// HtmlPage page = client.getPage(url);
// String source = page.getWebResponse().getContentAsString();
// p(source);
Request r = new Request();
r.getRequest(client);
while (true) ;
}
@Test
@Ignore
public void test() throws Exception {
System.out.println("-------------------------------");
WebClient webClient = new WebClient(BrowserVersion.CHROME);
webClient.getOptions().setCssEnabled(false);
webClient.getOptions().setJavaScriptEnabled(false);
HtmlPage page = webClient.getPage("http://news.163.com/domestic/");
// DomNodeList<HtmlElement> elements = page.getElementBy
System.out.println("---------------标题----------------");
DomNodeList<DomNode> domNodes = page.querySelectorAll(".item-top");
// log.debug("{}", domNodes);
for (DomNode domNode : domNodes) {
HtmlDivision htmlDivision = (HtmlDivision) domNode;
DomNodeList<HtmlElement> aElements = htmlDivision.getElementsByTagName("a");
HtmlAnchor htmlAnchor = (HtmlAnchor) aElements.get(0);
// HTMLHeadingElement htmlHeading2 = (HTMLHeadingElement)
// htmlDivision.getElementsByTagName("h2");
// HtmlAnchor htmlAnchor = (HtmlAnchor) htmlDivision.getElementsByTagName("a");
log.debug("{}", htmlAnchor.asText());
log.debug("{}", htmlAnchor.getAttribute("href"));
DomNodeList<HtmlElement> pElements = htmlDivision.getElementsByTagName("p");
HtmlParagraph htmlParagraph = (HtmlParagraph) pElements.get(0);
log.debug("{}", htmlParagraph.asText());
DomNodeList<HtmlElement> iEelements = htmlDivision.getElementsByTagName("img");
for (HtmlElement iEelement : iEelements) {
log.debug("{}", iEelement.getAttribute("src"));
}
String detailUrl = htmlAnchor.getAttribute("href");
if (detailUrl.equals("http://news.163.com/15/1215/17/BAT2L8RB00014JB6.html#f=dlist")) {
HtmlPage detailPage = webClient.getPage(detailUrl);
System.out.println("---------------正文----------------");
DomElement endTextElement = detailPage.getElementById("endText");
log.debug("{}", endTextElement.asText());
System.out.println("---------------图片----------------");
DomNodeList<DomNode> imgNodes = endTextElement.querySelectorAll(".f_center");
for (DomNode imgNode : imgNodes) {
HtmlParagraph imgpara = (HtmlParagraph) imgNode;
DomNodeList<HtmlElement> endImgs = imgpara.getElementsByTagName("img");
for (HtmlElement endImg : endImgs) {
log.debug("{}", endImg.getAttribute("src"));
}
}
}
}
webClient.close();
System.out.println("-------------------------------");
}
@Override
public void run() {
String str;
// 创建一个webclient
WebClient webClient = new WebClient();
// htmlunit 对css和javascript的支持不好，所以请关闭之
webClient.getOptions().setJavaScriptEnabled(false);
webClient.getOptions().setCssEnabled(false);
Cookie cookie =
new Cookie(
"www.qixin.com", "login_returnurl", "http%3A//www.qixin.com/search/prov/SH%3Fpage%3D2");
Cookie cookie1 =
new Cookie(
"www.qixin.com",
"userKey",
"QXBAdmin-Web2.0_5tUrhr/6EVtLT+GVfE+vU8k330y+oPICCM6jhUGEoLc%3D");
Cookie cookie2 =
new Cookie("www.qixin.com", "userValue", "4a68111b-0cfa-457f-91bd-b6fda97fa524");
Cookie cookie3 =
new Cookie(
"www.qixin.com",
"gr_session_id_955c17a7426f3e98",
"d25fe84e-fb1d-4ef8-8b4e-b530e5004b30");
Cookie cookie4 =
new Cookie("www.qixin.com", "_alicdn_sec", "5732cf53d99e48a838049be355d47a44000895ae");
CookieManager cookieManager = new CookieManager();
cookieManager.addCookie(cookie);
cookieManager.addCookie(cookie2);
cookieManager.addCookie(cookie3);
cookieManager.addCookie(cookie1);
cookieManager.addCookie(cookie4);
webClient.setCookieManager(cookieManager);
// 获取页面
HtmlPage page = null;
try {
page = webClient.getPage("http://www.qixin.com/search/prov/SH?page=20");
} catch (IOException e) {
e.printStackTrace();
}
// 获取页面的XML代码
List<HtmlAnchor> hbList = (List<HtmlAnchor>) page.getByXPath("//a");
Iterator iterator = hbList.iterator();
while (iterator.hasNext()) {
HtmlAnchor ha = (HtmlAnchor) iterator.next();
if ("search-result-title".equals(ha.getAttribute("class"))) {
System.out.println(ha.asText());
System.out.println("http://www.qixin.com" + ha.getAttribute("href"));
}
}
// 关闭webclient
webClient.close();
}
public static WebClient buildWebClient() {
WebClient webClient = new WebClient(BrowserVersion.FIREFOX_10);
webClient.setAjaxController(new NicelyResynchronizingAjaxController());
webClient.getOptions().setCssEnabled(true);
webClient.getOptions().setJavaScriptEnabled(true);
webClient.getOptions().setUseInsecureSSL(true);
webClient.setCssErrorHandler(
new ErrorHandler() {
@Override
public void warning(CSSParseException exception) throws CSSException {
// nothing to do here
}
@Override
public void error(CSSParseException exception) throws CSSException {
// todo: log or throw exception
}
@Override
public void fatalError(CSSParseException exception) throws CSSException {
// todo: log or throw exception
}
});
webClient.setIncorrectnessListener(
new IncorrectnessListener() {
@Override
public void notify(String message, Object origin) {
// todo: analyze and throw exception
}
});
webClient.waitForBackgroundJavaScript(100000);
webClient.getOptions().setThrowExceptionOnScriptError(false);
webClient.getOptions().setRedirectEnabled(true);
return webClient;
}
@Before
public void setUp() throws IOException {
if (server == null) {
server =
new Thread(
new Runnable() {
public void run() {
try {
Main.main(args);
} catch (IOException e) {
throw new RuntimeException(e);
}
}
});
server.start();
}
webClient.getOptions().setTimeout(30000);
}
private void ProcessTest(WebTest test) {
HtmlPage page = null;
// Orville.LOG().info(String.format("Executing test %s\n", event));
WebClient webClient = new WebClient();
webClient.setRefreshHandler(new ThreadedRefreshHandler());
webClient.getOptions().setJavaScriptEnabled(false);
try {
if (test.getMethod().toLowerCase().equals("post")) {
WebRequest request =
new WebRequest(UrlUtils.toUrlUnsafe(test.getUrlString()), HttpMethod.POST);
request.setRequestParameters(new ArrayList<NameValuePair>());
for (Map.Entry<String, String> entry : test.getPostData().entrySet()) {
request.getRequestParameters().add(new NameValuePair(entry.getKey(), entry.getValue()));
}
page = webClient.getPage(request);
} else {
page = webClient.getPage(test.getUrlString());
}
ProcessResult(test, page.getWebResponse());
} catch (FailingHttpStatusCodeException fsc) {
Orville.LOG()
.warning(
String.format(
"Failing HTTP Status code caught executing the test command: %s\n",
fsc.getMessage()));
} catch (IOException ioe) {
Orville.LOG()
.warning(
String.format(
"IO Exception caught executing the test command: %s\n", ioe.getMessage()));
}
webClient.closeAllWindows();
}
@JsfTest(JsfVersion.JSF_2_2_5)
@Test
public void testDropFlashCookie() throws Exception {
HtmlPage page = webClient.getPage(webUrl + "faces/flashDropCookie.xhtml");
webClient.getOptions().setRedirectEnabled(true);
HtmlTextInput textInput = (HtmlTextInput) page.getHtmlElementById("input");
textInput.setValueAttribute("test");
HtmlSubmitInput button = (HtmlSubmitInput) page.getHtmlElementById("submit");
int currentSize = webClient.getCookieManager().getCookies().size();
page = button.click();
HtmlElement element = page.getHtmlElementById("link");
page = element.click();
int newSize = webClient.getCookieManager().getCookies().size();
assertTrue(newSize < currentSize);
}
/**
* Set up method run before each test to create a new Web client and load the application's
* landing page.
*/
@Before
@RunAsClient
public void setUp() {
try {
browser = new WebClient();
browser.getOptions().setThrowExceptionOnScriptError(false);
landingPageResponse = browser.getPage(deploymentUrl.toString() + "index.xhtml");
Assert.assertEquals(
"Could not load the application landing page.",
"Cargo Tracker",
landingPageResponse.getTitleText());
} catch (IOException ex) {
Assert.fail(
"An IOException was thrown during the test setup for class \""
+ ViewDetailTest.class.getSimpleName()
+ "\" at method \""
+ testName.getMethodName()
+ "\" with message: "
+ ex.getMessage());
}
}
/**
* <span style="color:red">INTERNAL API - SUBJECT TO CHANGE AT ANY TIME - USE AT YOUR OWN
* RISK.</span><br>
*
* <p>Submits this form to the server. If <tt>submitElement</tt> is <tt>null</tt>, then the
* submission is treated as if it was triggered by JavaScript, and the <tt>onsubmit</tt> handler
* will not be executed.
*
* <p><b>IMPORTANT:</b> Using this method directly is not the preferred way of submitting forms.
* Most consumers should emulate the user's actions instead, probably by using something like
* {@link HtmlElement#click()} or {@link HtmlElement#dblClick()}.
*
* @param submitElement the element that caused the submit to occur
* @return a new page that reflects the results of this submission
*/
Page submit(final SubmittableElement submitElement) {
final HtmlPage htmlPage = (HtmlPage) getPage();
final WebClient webClient = htmlPage.getWebClient();
if (webClient.getOptions().isJavaScriptEnabled()) {
if (submitElement != null) {
isPreventDefault_ = false;
final ScriptResult scriptResult = fireEvent(Event.TYPE_SUBMIT);
if (isPreventDefault_) {
// null means 'nothing executed'
if (scriptResult == null) {
return htmlPage;
}
return scriptResult.getNewPage();
}
}
final String action = getActionAttribute().trim();
if (StringUtils.startsWithIgnoreCase(action, JavaScriptURLConnection.JAVASCRIPT_PREFIX)) {
return htmlPage
.executeJavaScriptIfPossible(action, "Form action", getStartLineNumber())
.getNewPage();
}
} else {
if (StringUtils.startsWithIgnoreCase(
getActionAttribute(), JavaScriptURLConnection.JAVASCRIPT_PREFIX)) {
// The action is JavaScript but JavaScript isn't enabled.
// Return the current page.
return htmlPage;
}
}
final WebRequest request = getWebRequest(submitElement);
final String target = htmlPage.getResolvedTarget(getTargetAttribute());
final WebWindow webWindow = htmlPage.getEnclosingWindow();
webClient.download(webWindow, target, request, false, "JS form.submit()");
return htmlPage;
}
public static void savePage(String path) {
// path = transformPath(path);
if (!Main.list.contains(path) && path.contains("https://www.darts-ip.com")) {
System.out.println("");
System.out.println("=================================================");
System.out.println("");
System.out.println(path);
System.out.println("");
Main.list.add(path);
String content = "";
String fileName = "";
HtmlPage page = null;
// ----------HTML-UNIT----------
try {
WebClient webClient = new WebClient();
webClient.getOptions().setThrowExceptionOnScriptError(false);
page = webClient.getPage(path);
WebResponse response = page.getWebResponse();
content = response.getContentAsString();
fileName = path + ".html";
} catch (Exception e) {
System.out.println("");
System.out.println("ERROR in Page.savePage #HTML-UNIT");
System.out.println(e);
System.out.println("");
}
// ----------SAVE-HTML-FILE----------
try {
PrintWriter writer = new PrintWriter(fileName, "UTF-8");
writer.print(content);
writer.close();
} catch (Exception e) {
System.out.println("");
System.out.println("ERROR in ERROR in Page.savePage #SAVE-HTML-FILE");
System.out.println(e);
System.out.println("");
}
// ----------ZIP-HTML-FILE----------
try {
FileInputStream fi = new FileInputStream(fileName);
BufferedInputStream buffi = new BufferedInputStream(fi, Main.BUFFER);
ZipEntry entry = new ZipEntry(fileName);
Main.out.putNextEntry(entry);
int count;
while ((count = buffi.read(Main.data, 0, Main.BUFFER)) != -1) {
Main.out.write(Main.data, 0, count);
}
Main.out.closeEntry();
buffi.close();
} catch (Exception e) {
System.out.println("");
System.out.println("ERROR in Page.savePage #ZIP-HTML-FILE");
System.out.println(e);
System.out.println("");
}
// ----------DELETE-TMP-FILE----------
try {
new File(fileName).delete();
} catch (Exception e) {
System.out.println("");
System.out.println("ERROR in Page.savePage #DELETE-TMP-FILE");
System.out.println(e);
System.out.println("");
}
System.out.println("");
System.out.println("=================================================");
// ----------EXPLORE-LINKS----------
exploreLinks(page);
}
}
/**
* Processes requests for both HTTP <code>GET</code> and <code>POST</code> methods.
*
* @param request servlet request
* @param response servlet response
* @throws ServletException if a servlet-specific error occurs
* @throws IOException if an I/O error occurs
*/
protected void processRequest(HttpServletRequest request, HttpServletResponse response)
throws ServletException, IOException {
response.setContentType("text/html;charset=UTF-8");
PrintWriter out = response.getWriter();
int no = 0;
try {
no = Integer.parseInt(request.getParameter("no"));
} catch (NumberFormatException exception) {
no = 0;
}
try {
/* TODO output your page here. You may use following sample code. */
out.println("<!DOCTYPE html>");
out.println("<html>");
out.println("<head>");
out.println("<title>Servlet NewServlet</title>");
out.println("</head>");
out.println("<body>");
final WebClient webClient = new WebClient();
webClient.getCookieManager().clearCookies();
webClient.getOptions().setUseInsecureSSL(true);
webClient.setJavaScriptEnabled(false);
final HtmlPage page = webClient.getPage("https://filestream.me/");
// out.println(page.getTitleText());
System.out.println("no of forms : " + page.getForms().size());
HtmlForm form = page.getForms().get(1);
final HtmlTextInput textField = form.getInputByName("login");
textField.setAttribute("value", "*****@*****.**");
final HtmlPasswordInput passField = form.getInputByName("password");
passField.setAttribute("value", "ramkrishnan18");
HtmlSubmitInput htmlSubmitInput = form.getInputByValue("login");
HtmlPage page1 = htmlSubmitInput.click();
HtmlTable htmlTable = page1.getHtmlElementById("fileCatTable");
List<HtmlTableRow> listOfHtmlTableRow = htmlTable.getBodies().get(0).getRows();
List<String> allLinks = new ArrayList<String>();
for (HtmlTableRow htmlTableRow : listOfHtmlTableRow) {
HtmlTableCell cell = htmlTableRow.getCells().get(htmlTableRow.getCells().size() - 1);
for (DomElement domElement : cell.getChildElements()) {
// out.println(domElement.getTagName());
if (domElement.getTagName().equals("div")) {
boolean flag = true;
for (DomElement celldomElement : domElement.getChildElements()) {
if (flag) {
// out.println(celldomElement.getTagName());
String title = celldomElement.getAttribute("title");
if ("Download".equals(title) || "Downloads".equals(title)) {
String link = celldomElement.getAttribute("onclick");
String http =
link.substring(
link.indexOf('\'') + 1, link.indexOf('\'', link.indexOf('\'') + 1));
// out.println(celldomElement.getAttribute("onclick") + "<br/>");
// out.println(http + "<br/>");
allLinks.add(http);
flag = false;
}
}
}
}
}
}
// InputStream is =anchorAttachment.click().getWebResponse().getContentAsStream();
out.println("************ Start *************");
// HtmlAnchor anchorElement=HTMLAnchorElement.;
// anchorElement.set
// anchorElement.setHref(allLinks.get(0));
// anchorElement.cl
final String u = allLinks.get(no);
Thread t =
new Thread(
new Runnable() {
@Override
public void run() {
try {
URL url = new URL(u);
System.out.println("Link : " + u);
HttpURLConnection con = (HttpURLConnection) url.openConnection();
// System.out.println("Response Code : " + con.getInputStream().available());
con.setRequestProperty("Accept-Encoding", "gzip,deflate");
con.setRequestProperty(
"User-Agent",
"Mozilla/5.0 (X11; Linux i686) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/28.0.1500.71 Safari/537.36");
Map<String, List<String>> map = con.getHeaderFields();
for (String object : map.keySet()) {
System.out.println(object + " --> " + map.get(object));
}
System.out.println("Response Code : " + con.getResponseCode());
System.out.println("length :" + con.getContentLength());
System.out.println("Message " + con.getResponseMessage());
InputStream inputStream = con.getInputStream();
String path = System.getenv("OPENSHIFT_JBOSSAS_DIR");
FileOutputStream fileOutputStream = new FileOutputStream(path + "test.zip");
byte[] buffer = new byte[1024];
System.out.println(inputStream.available());
int bytesRead = 10;
while (true) {
// System.out.print("--");
bytesRead = inputStream.read(buffer);
// System.out.println(bytesRead);
if (bytesRead == -1) {
break;
}
fileOutputStream.write(buffer, 0, bytesRead);
// System.out.print(">");
}
fileOutputStream.close();
System.out.println("File Completed");
} catch (IOException ex) {
Logger.getLogger(MakeData.class.getName()).log(Level.SEVERE, null, ex);
}
}
});
t.start();
// System.out.println(allLinks.get(0));
// System.out.println(webClient.getCookieManager().getCookies(new
// URL(allLinks.get(0))).size());
// // System.out.println(httpPage.asText());
// HtmlPage httpPage = webClient.getPage(allLinks.get(0));
//
// // final String pageAsXml = page1.asXml();
// // InputStream inputStream=httpPage.getWebResponse().getContentAsStream();
// // inputStream.available();
// System.out.println("Code : "+httpPage.getWebResponse().getStatusMessage());
// System.out.println("Code : "+httpPage.getWebResponse().getContentType());
// System.out.println("netCode :
// "+httpPage.getWebResponse().getContentCharset());
//
// out.println("<br/><br/>Code : "+httpPage.getWebResponse().getStatusCode());
// out.println("<br/><br/>Code : "+httpPage.getWebResponse().getContentType());
// out.println("<br/><br/>Code : "+httpPage.getWebResponse().getContentCharset());
// final String pageAsText = page1.asText();
// out.println(pageAsXml);
webClient.closeAllWindows();
out.println("<h1>Servlet NewServlet at " + request.getContextPath() + "</h1>");
out.println("</body>");
out.println("</html>");
} finally {
out.close();
}
}
public static void main(String[] args) throws IOException {
WebClient client = new WebClient(BrowserVersion.CHROME);
CookieManager cookie = new CookieManager();
client.setCookieManager(cookie);
client.getOptions().setJavaScriptEnabled(true);
client.getOptions().setActiveXNative(false);
client.getOptions().setCssEnabled(false);
client.getOptions().setThrowExceptionOnScriptError(false);
client.getOptions().setThrowExceptionOnFailingStatusCode(false);
client.getOptions().setDoNotTrackEnabled(true);
client.getOptions().setPrintContentOnFailingStatusCode(false);
client.setAjaxController(new NicelyResynchronizingAjaxController());
client.setJavaScriptTimeout(Long.MAX_VALUE);
List<Integer> TopCategory = new LinkedList<Integer>(TopCategoryUrl.keySet());
Collections.sort(TopCategory);
long today = new Date().getTime();
long ymd = today - 1000 * 60 * 60 * 24;
String date = sdf.format(ymd);
FileWriter fw = new FileWriter("ShuTaobaoTop/search-" + date, false);
BufferedWriter bw = new BufferedWriter(fw);
while (TopCategory.size() > 0) {
int cid = TopCategory.get(0);
// if(cid<50002766)
// {
// TopCategory.remove(0);
// continue;
// }
String cname = TopCategoryUrl.get(cid);
String pid = "0", pname = "";
String topUrl = "http://shu.taobao.com/top/" + cid + "/search";
boolean success = false;
ArrayList<String> result = new ArrayList<String>();
try {
// client.getOptions().setJavaScriptEnabled(true);
do {
System.out.println("剩余类目数=" + TopCategory.size());
// ProxyUnit.configProxy(client, ProxyUnit.CHECKURL);
} while (false == AccountLogin.loginTaoBao(client, AccountLogin.TAOBAOLOGINURL, true));
// client.getOptions().setJavaScriptEnabled(false);
HtmlPage queryTrade = client.getPage(topUrl);
Thread.sleep(1000L);
System.out.println(queryTrade.getTitleText());
/*
* 解析
*/
Document doc = Jsoup.parse(queryTrade.asXml());
Element time = doc.getElementsByAttributeValue("class", "time").get(0);
String startTime =
time.text().substring(6, 16).replace("-", ""); // 统计时间: 2014-05-17 - 2014-05-23
String endTime = time.text().substring(19).replace("-", "");
Elements elements =
doc.getElementsByAttributeValueStarting(
"class", "mod "); // <div class="mod odd mod-10 sm-ua"> <h3 class="title">
for (Element element : elements) {
String title = element.getElementsByTag("h3").get(0).text();
System.out.println(title);
Elements lis =
element.getElementsByTag("ol").get(0).getElementsByTag("li"); // <li class="up ">
for (Element li : lis) {
Elements spans = li.getElementsByTag("span");
String rank = spans.get(0).text();
String key = spans.get(1).text();
String rise = spans.get(2).text();
rise = rise.substring(0, rise.length() - 1);
URL href =
new URL(
URLDecoder.decode(
"http://shu.taobao.com"
+ spans.get(1).getElementsByTag("a").get(0).attr("href"),
"utf-8"));
String hrefQuery = href.getQuery();
if (hrefQuery.contains("cid=")) {
if ("0".equals(pid)) {
pid = String.valueOf(cid);
pname = cname;
}
cid = Integer.parseInt(hrefQuery.split("cid=")[1].split("&")[0]);
cname = title;
}
System.out.println(
startTime + "\001" + endTime + "\001" + cid + "\001" + cname + "\001" + pid + "\001"
+ pname + "\001" + rank + "\001" + key + "\001" + rise);
result.add(
startTime + "\001" + endTime + "\001" + cid + "\001" + cname + "\001" + pid + "\001"
+ pname + "\001" + rank + "\001" + key + "\001" + rise);
}
}
success = true;
} catch (Exception e) {
// TODO Auto-generated catch block
// e.printStackTrace();
}
if (success) {
TopCategory.remove(0);
System.out.println("剩余类目数=" + TopCategory.size());
for (String r : result) {
bw.write(r + "\n");
}
}
}
bw.close();
fw.close();
}
@Test
public void testNullInCompositionTemplate() throws Exception {
webClient.getOptions().setThrowExceptionOnFailingStatusCode(false);
HtmlPage page = webClient.getPage(webUrl + "faces/compositionWithNullTemplate.xhtml");
assertTrue(page.asText().contains("Invalid path"));
}
public InternetConnection() {
client.getOptions().setJavaScriptEnabled(false);
// Turns off annoying exceptions caused by broken javascript links,
// which still cause problems despite javascript being turned off.
client.getOptions().setThrowExceptionOnFailingStatusCode(false);
}
/** Setup before testing. */
@Before
public void setUp() {
webUrl = System.getProperty("integration.url");
webClient = new WebClient();
webClient.getOptions().setThrowExceptionOnFailingStatusCode(true);
}
