@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("-------------------------------"); }