packagecom.szy.project.utils;importjava.io.BufferedInputStream;importjava.io.BufferedOutputStream;importjava.io.BufferedReader;importjava.io.DataOutputStream;importjava.io.File;importjava.io.FileInputStream;importjava.io.FileOutputStream;importjava.io.IOException;importjava.io.InputStream;importjava.io.InputStreamReader;importjava.io.OutputStreamWriter;importjava.io.Reader;importjava.io.Writer;importjava.util.Enumeration;importjavax.mail.MessagingException;importjavax.mail.Multipart;importjavax.mail.Session;importjavax.mail.internet.MimeBodyPart;importjavax.mail.internet.MimeMessage;importjavax.mail.internet.MimeMultipart;importorg.jsoup.Jsoup;importorg.jsoup.nodes.Document;/*** 转换工具 ---------- 需要引入第三方依赖 javaMail转换格式 和 jsoup解析HTML
* jsoup 文档地址 :http://www.open-open.com/jsoup/parse-document-from-string.htm* 将mht 转化成 HTML
*@author隔壁老王
**/
public classMht2HtmlUtil {public static void main(String[] args) throwsIOException {/*** 转换*/
//mht2html("f:\\job_111.mht", "f:\\test.htm");
/*** 获取姓名和性别*/String nameAndSex= Mht2HtmlUtil.findResultValue("f:\\test.htm", "li", "info_name");
String tmpString= nameAndSex.replaceAll("(?i)[^a-zA-Z0-9\u4E00-\u9FA5]", "");//去掉所有中英文符号
char[] carr =tmpString.toCharArray();for (int i = 0; i < tmpString.length(); i++) {if (carr[i] < 0xFF) {
carr[i]= ' ';//过滤掉非汉字内容
}
}
System.out.println(tmpString.substring(0, tmpString.length()-1)); //姓名
System.out.println(tmpString.substring(tmpString.length()-1)); //性别
/*** 获取教育经历*/File htmlf=new File("f:\\test.htm");
Document doc=Jsoup.parse(htmlf, "UTF-8");
String ss=doc.body().toString();//class等于masthead的li标签
Object[] aa= doc.select("div.detaile_box").toArray();for (int i = 0; i < aa.length; i++) {if(i==3){
String strtext=aa[i].toString();
Document docs=Jsoup.parse(strtext);
Object[] bb= docs.select("b.edu_main_sch").toArray();for (int j = 0; j < bb.length; j++) {
String tt=bb[j].toString();
Document doct=Jsoup.parse(tt);
String result= doct.select("b.edu_main_sch").text();
String a=