- html转txt
import org.apache.commons.lang3.StringEscapeUtils;- import org.jsoup.Jsoup;
- import org.jsoup.nodes.Document;
- import org.jsoup.nodes.Document.OutputSettings;
- import org.jsoup.safety.Whitelist;
- public class Html2Txt {
- public static String toPlainText(String html)
- {
- if (html == null)
- {
- return "";
- }
- final Document document =Jsoup.parse(html);
- final OutputSettings outputSettings = new Document.OutputSettings().prettyPrint(false);
- document.outputSettings(outputSettings);
- document.select("br").append("\\n");
- document.select("p").prepend("\\n");
- document.select("p").append("\\n");
- final String newHtml = document.html().replaceAll("\\\\n", "\n");
- final String plainText = Jsoup.clean(newHtml, "", Whitelist.none(), outputSettings);
- final String result = StringEscapeUtils.unescapeHtml4(plainText.trim());
- return result;
- }
- public static void main(String[] args){
- }
- }
-----------------
java操作xml
- import org.dom4j.Attribute;
- import org.dom4j.Element;
- public class ElmUtils {
- public static Element addNode(Element elm, String nodeName, String txt) {
- Element node = elm.addElement(nodeName);
- txt = txt == null ? "" : txt;
- node.addText(txt);
- return node;
- }
- public static String getElmTxt(Element root, String elmNM) {
- Element tmpElm = (Element) root.selectSingleNode("//" + elmNM);
- if (tmpElm == null) {
- return "";
- } else {
- return tmpElm.getText();
- }
- }
- public static String elementTxt(Element root, String elmNM) {
- Element tmpElm = (Element) root.element(elmNM);
- if (tmpElm == null) {
- return "";
- } else {
- return tmpElm.getText();
- }
- }
- public static String getCurElmTxt(Element root, String elmNM) {
- Element tmpElm = (Element) root.selectSingleNode(elmNM);
- if (tmpElm == null) {
- return "";
- } else {
- return tmpElm.getText();
- }
- }
- public static String getElmAttrVal(Element root, String attrNM) {
- Attribute temAttr = root.attribute(attrNM);
- if (temAttr == null) {
- return "";
- } else {
- return temAttr.getValue();
- }
- }
- public static Element setElmTxt(Element elm, String elmName, String elmTxt) {
- Element tempElm = (Element) elm.selectSingleNode("//" + elmName);
- if(tempElm!=null){
- tempElm.setText(elmTxt);
- }
- return elm;
- }
- public static Element setElmAttr(Element elm, String attrName, String attrVal){
- Attribute attr = elm.attribute(attrName);
- attr.setValue(attrVal);
- return elm;
- }
- }
以上就是使用java对Html操作的详细内容,更多关于使用java对Html操作的资料请关注九品源码其它相关文章!