當前位置: 首頁>>代碼示例>>Java>>正文

Java HtmlPage.asXml方法代碼示例

本文整理匯總了Java中com.gargoylesoftware.htmlunit.html.HtmlPage.asXml方法的典型用法代碼示例。如果您正苦於以下問題：Java HtmlPage.asXml方法的具體用法？Java HtmlPage.asXml怎麽用？Java HtmlPage.asXml使用的例子？那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在類com.gargoylesoftware.htmlunit.html.HtmlPage的用法示例。


在下文中一共展示了HtmlPage.asXml方法的11個代碼示例，這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚，您的評價將有助於係統推薦出更棒的Java代碼示例。

示例1: mockLogin

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
public String mockLogin(URL url, String username, String password) {
	try {
		HtmlPage page1 = webClient.getPage(url.getURLValue());  
		// find the login form
        HtmlForm form = page1.getForms().get(0);  
        
        // fill in the input
        HtmlInput hi = form.getInputByName("commit");
        HtmlTextInput textField = form.getInputByName("login");  
        HtmlPasswordInput pass = form.getInputByName("password");  
        
        textField.click();
        textField.setValueAttribute(username);  
        pass.click();
        pass.setValueAttribute(password);  
      
        // push the button
        HtmlPage page2 = hi.click();  
        return page2.asXml();

	} catch(IOException ioe)  {
		ioe.printStackTrace();
		return null;
	}
}
開發者ID:knshen，項目名稱:JSearcher，代碼行數:26，代碼來源:GithubLoginSimulator.java

示例2: main

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
public static void main(String[] args) throws IOException {

        // 瀏覽器
        WebClient webClient = new WebClient(BrowserVersion.CHROME);
        webClient.getOptions().setUseInsecureSSL(true);//支持https
        webClient.getOptions().setJavaScriptEnabled(true); // 啟用JS解釋器，默認為true
        webClient.getOptions().setCssEnabled(false); // 禁用css支持
        webClient.getOptions().setThrowExceptionOnScriptError(false); // js運行錯誤時，是否拋出異常
        webClient.getOptions().setTimeout(10000); // 設置連接超時時間 ，這裏是10S。如果為0，則無限期等待
        webClient.getOptions().setDoNotTrackEnabled(false);
        webClient.setJavaScriptTimeout(8000);//設置js運行超時時間
        webClient.waitForBackgroundJavaScript(500);//設置頁麵等待js響應時間,

        // proxy
        //webClient.getOptions().setProxyConfig(new ProxyConfig("IP", 80));

        HtmlPage page = webClient.getPage("http://---");
        String pageXml = page.asXml(); //以xml的形式獲取響應文本
        System.out.println(pageXml);

    }
開發者ID:xuxueli，項目名稱:xxl-incubator，代碼行數:22，代碼來源:Demo.java

示例3: allowAccess

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
private HtmlPage allowAccess(WebClient webClient, HtmlPage allowAccessPage) throws IOException {
	HtmlButton allowAccessButton = (HtmlButton) allowAccessPage.getElementById("submit_approve_access");
	if (allowAccessButton == null) {
		throw new RuntimeException("Cannot find allow access button in html page :\n" + allowAccessPage.asXml());
	}
	webClient.waitForBackgroundJavaScriptStartingBefore(WAIT_DELAY_MS);
	// allowAccessButton.click() does not work because
	// allowAccessButton.isVisible() is false
	// for some reason (click() was working with htmlunit 2.23)
	HtmlPage tokenPage = clickButtonIgnoringVisibility(allowAccessButton);
	return tokenPage;
}
開發者ID:cchabanois，項目名稱:mesfavoris，代碼行數:13，代碼來源:HtmlUnitAuthorizationCodeInstalledApp.java

示例4: getHtmlPageHtmlUnit

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
/**
 * HtmlUnit을 이용한 HTML 코드 파싱.
 * @param eachArchiveAddress 실제 만화가 담긴 아카이브 주소
 * @return 성공 시 html 코드를 리턴
 */
private String getHtmlPageHtmlUnit(String eachArchiveAddress) throws Exception {
	/* 필수! 로그 메세지 출력 안함 -> HtmlUnit 이용시 Verbose한 로그들이 너무 많아서 다 끔 */
	java.util.logging.Logger.getLogger("com.gargoylesoftware").setLevel(Level.OFF); 
	System.setProperty("org.apache.commons.logging.Log", "org.apache.commons.logging.impl.NoOpLog");
	
	System.out.print("일반 연결 시도중 ... ");
	
	WebClient webClient = new WebClient();
	webClient.getOptions().setRedirectEnabled(true);
	
	WebRequest req = new WebRequest(new URL(eachArchiveAddress));
	req.setHttpMethod(HttpMethod.POST);
	req.setAdditionalHeader("User-Agent", UserAgent.getUserAgent());
	req.setAdditionalHeader("Accept-Encoding", "gzip"); //20171126 gzip 추가
	req.getRequestParameters().add(new NameValuePair("pass", PASSWORD)); //비밀번호 post 방식 전송
	
	HtmlPage page = webClient.getPage(req);
	
	//Html코드를 포함한 페이지 소스코드가 담길 스트링
	String pageSource = page.asXml();
	
	/** 여기도 페이지 파싱 실패 시 검증하는 코드 들어가야 됨 **/
	
	webClient.close();
	System.out.println("성공");
	return pageSource;
}
開發者ID:occidere，項目名稱:MMDownloader，代碼行數:33，代碼來源:Downloader.java

示例5: getSuapPage

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
public String getSuapPage(String url) throws IOException {
    HtmlPage pagina = conn.getPage(url);

    if (pagina.getTitleText().equals(SuapConnection.SUAP_LOGIN_PAGE_TITLE)) {
        return login(url).asXml();
    } else {
        return pagina.asXml();
    }
}
開發者ID:marcocspc，項目名稱:SUAPPasswordResetter，代碼行數:10，代碼來源:SuapConnection.java

示例6: main

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
public static void main(String[] args) throws FailingHttpStatusCodeException, MalformedURLException, IOException {
    
    // 屏蔽HtmlUnit等係統 log
    LogFactory.getFactory().setAttribute("org.apache.commons.logging.Log","org.apache.commons.logging.impl.NoOpLog");
    java.util.logging.Logger.getLogger("com.gargoylesoftware").setLevel(Level.OFF);
    java.util.logging.Logger.getLogger("org.apache.http.client").setLevel(Level.OFF);
    
    String url = "https://www.douyin.com/share/video/6496703951436516621/?mid=6484356820260686606";
    System.out.println("Loading page now-----------------------------------------------: "+url);
    
    /* HtmlUnit 模擬瀏覽器 */
    WebClient webClient = new WebClient(BrowserVersion.CHROME);
    webClient.getOptions().setJavaScriptEnabled(true);              // 啟用JS解釋器，默認為true  
    webClient.getOptions().setCssEnabled(false);                    // 禁用css支持  
    webClient.getOptions().setThrowExceptionOnScriptError(false);   // js運行錯誤時，是否拋出異常
    webClient.getOptions().setThrowExceptionOnFailingStatusCode(false);
    webClient.getOptions().setTimeout(10 * 1000);                   // 設置連接超時時間
    HtmlPage page = webClient.getPage(url);
    webClient.waitForBackgroundJavaScript(30 * 1000);               // 等待js後台執行30秒

    String pageAsXml = page.asXml();
    
    /* Jsoup解析處理 */
    // Document doc = Jsoup.parse(pageAsXml, "https://bluetata.com/");
    Document doc = Jsoup.parse(pageAsXml);  
    Elements pngs = doc.select("img[src$=.png]");                   // 獲取所有圖片元素集
    // 其他操作
    System.out.println(doc.toString());
}
開發者ID:bluetata，項目名稱:crawler-jsoup-maven，代碼行數:30，代碼來源:htmlunitTest.java

示例7: download

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
/**
 * return the html of url
 * @param url : form data
 * @param inputs : input name -> input value
 * @return
 */
public String download(URL url, Map<String, String> inputs) {
	HtmlSubmitInput button = null;
	HtmlPage nextPage = null;
	
	try {
		HtmlPage page = webClient.getPage(url.getURLValue()); 
		for(Map.Entry<String, String> input : inputs.entrySet()) {
			String form_name = input.getKey().split("\\.")[0];
			String input_name = input.getKey().split("\\.")[1];
			
			HtmlForm form = page.getFormByName(form_name);
			if(input_name.equals("button")) {
				button = form.getInputByValue(input.getValue());
			}
			else {
				HtmlTextInput text_input = form.getInputByName(input_name);
				text_input.setValueAttribute(input.getValue());
			}
		}
		nextPage = button.click();
	} catch(IOException ioe) {
		ioe.printStackTrace();
	}
	  
	return nextPage.asXml();
}
開發者ID:knshen，項目名稱:JSearcher，代碼行數:33，代碼來源:PostDownloader.java

示例8: getDocumentFromGateway

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
public static DownloadableDocument getDocumentFromGateway(HtmlPage gatewayPage, String courseName, Term term) {
    String pageContent = gatewayPage.asXml();
    String name = fetchFileName(gatewayPage);
    String downloadLink = fetchDownloadLink(pageContent);
    String extension = fetchFileExtension(downloadLink);
    return new DownloadableDocument(name, downloadLink, courseName, extension, term);
}
開發者ID:theovier，項目名稱:lernplattform-crawler，代碼行數:8，代碼來源:DocumentCrawler.java

示例9: login

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
public static boolean login(WebClient client, String userName,
		String passWrod) throws Exception {
	boolean flag = false;
	try {
		String s = "https://passport.sina.cn/signin/signin?entry=wapsso&vt=4&r=http%3A%2F%2Fmy.sina.cn%2F%3Fpos%3D108%26vt%3D4%26m%3D78fc51068140045a973a3aeab4db2381&amp;revalid=1";
		HtmlPage page = client.getPage(s);
		// <input type="text" placeholder="微博帳號/手機號/郵箱" autocorrect="off"
		// autocapitalize="off" id="loginName">

		HtmlInput htmlInput = page
				.getFirstByXPath("//input[@id='loginName']");
		htmlInput.click();
		htmlInput.setAttribute("value", userName);

		// <input type="password" placeholder="密碼" id="loginPassword">
		htmlInput = page.getFirstByXPath("//input[@id='loginPassword']");
		htmlInput.click();
		htmlInput.setAttribute("value", passWrod);

		// <a id="loginAction" class="btn_login" href="javascript:;">登錄</a>

		HtmlElement a = page.getFirstByXPath("//a[@id='loginAction']");
		page = a.click();

		client.setJavaScriptTimeout(5000);
		String str = page.asXml();
		if (!str.contains("QQ帳號登錄")) {
			flag = true;
			logger.error("登錄成功");
		}
	} catch (Exception e) {
		logger.error(e);
		flag = false;
	}
	return flag;
}
開發者ID:xiaomin0322，項目名稱:alimama，代碼行數:37，代碼來源:HtmlUnitUtil.java

示例10: selectEmailRecoveryAsSignInChallenge

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
private HtmlPage selectEmailRecoveryAsSignInChallenge(WebClient webClient, HtmlPage signInChallengePage)
		throws IOException {
	List<HtmlForm> forms = signInChallengePage.getForms();
	Optional<HtmlForm> kpeForm = forms.stream()
			.filter(form -> "/signin/challenge/kpe/2".equals(form.getActionAttribute())).findFirst();
	if (!kpeForm.isPresent()) {
		throw new RuntimeException(
				"Cannot find recovery by email form in html page :\n" + signInChallengePage.asXml());
	}
	HtmlButton button = (HtmlButton) kpeForm.get().getElementsByTagName("button").get(0);
	HtmlPage htmlPage = button.click();
	webClient.waitForBackgroundJavaScriptStartingBefore(WAIT_DELAY_MS);
	return htmlPage;
}
開發者ID:cchabanois，項目名稱:mesfavoris，代碼行數:15，代碼來源:HtmlUnitAuthorizationCodeInstalledApp.java

示例11: download

import com.gargoylesoftware.htmlunit.html.HtmlPage; //導入方法依賴的package包/類
public HttpResponse download(HttpRequest request, int timeout) throws DownloadException {
	try {
		URL url = new URL(request.getUrl());
		WebRequest webRequest = new WebRequest(url);
		webRequest.setHttpMethod(HttpMethod.GET);
		if(request instanceof HttpPostRequest) {//post
			HttpPostRequest post = (HttpPostRequest)request;
			webRequest.setHttpMethod(HttpMethod.POST);
			List<NameValuePair> requestParameters = new ArrayList<NameValuePair>();
			for(Map.Entry<String, Object> entry : post.getFields().entrySet()) {
				NameValuePair nvp = new NameValuePair(entry.getKey(), entry.getValue().toString());
				requestParameters.add(nvp);
			}
			webRequest.setRequestParameters(requestParameters);	
		}
		//header
		boolean isMobile = SpiderThreadLocal.get().getEngine().isMobile();
		webRequest.setAdditionalHeader("User-Agent", UserAgent.getUserAgent(isMobile));
		webRequest.setAdditionalHeaders(request.getHeaders());
		//proxy
		HttpHost proxy = Proxys.getProxy();
		if(proxy != null) {
			webRequest.setProxyHost(proxy.getHostName());
			webRequest.setProxyPort(proxy.getPort());
		}
		//timeout
		this.webClient.getOptions().setTimeout(timeout);
		//request,response
		webClient.getPage(webRequest);
		HtmlPage page = webClient.getPage(request.getUrl());
		HttpResponse resp = new HttpResponse();
		WebResponse webResponse = page.getWebResponse();
		int status = webResponse.getStatusCode();
		resp.setStatus(status);
		if(status == 302 || status == 301) {
			String redirectUrl = webResponse.getResponseHeaderValue("Location");
			resp.setContent(UrlUtils.relative2Absolute(request.getUrl(), redirectUrl));
		} else if(status == 200) {
			String content = page.asXml();
			resp.setContent(content);
			resp.setRaw(webResponse.getContentAsStream());
			String contentType = webResponse.getContentType();
			resp.setContentType(contentType);
			String charset = getCharset(request.getCharset(), contentType);
			resp.setCharset(charset);
		} else {
			throw new DownloadException("ERROR : " + status);
		}
		return resp;
	} catch(Exception ex) {
		throw new DownloadException(ex);
	}
}
開發者ID:xtuhcy，項目名稱:gecco-htmlunit，代碼行數:54，代碼來源:HtmlUnitDownloder.java


注：本文中的com.gargoylesoftware.htmlunit.html.HtmlPage.asXml方法示例由純淨天空整理自Github/MSDocs等開源代碼及文檔管理平台，相關代碼片段篩選自各路編程大神貢獻的開源項目，源碼版權歸原作者所有，傳播和使用請參考對應項目的License；未經允許，請勿轉載。