html转Word与PDFjar包与代码

html转Word与PDFjar包与代码，转Word使用poi，PDF使用pd4ml，包含demo，包含字体fonts的jar包

package com.solid.util;
 
import java.io.BufferedReader;
import java.io.ByteArrayInputStream;
import java.io.File;
import java.io.FileOutputStream;
import java.io.FileReader;
import java.io.IOException;
 
import org.apache.poi.poifs.filesystem.DirectoryEntry;
import org.apache.poi.poifs.filesystem.DocumentEntry;
import org.apache.poi.poifs.filesystem.POIFSFileSystem;
/**
 * 将html文档转为doc
 * @author soildwang
 *
 */
public class HtmlToDoc {
       /**
        * 读取html文件到word
        * @param filepath html文件的路径
        * @return
        * @throws Exception
        */
       public boolean writeWordFile（String filepath） throws Exception {
              boolean flag = false;
              ByteArrayInputStream bais = null;
              FileOutputStream fos = null;
              String path = “C:/“;  //根据实际情况写路径
              try {
                     if （!““.equals（path）） {
                            File fileDir = new File（path）;
                            if （fileDir.exists（）） {
                                   String content = “lns:v=“+“‘urn:schemas-microsoft-com:vml‘“+ 
                                		   “xmlns:o=“+“‘urn:schemas-microsoft-com:office:office‘“+ 
                                		   “xmlns:w=“+“‘urn:schemas-microsoft-com:office:word‘“+ 
                                		   “xmlns:m=“+“‘http://schemas.microsoft.com/office/2004/12/omml‘“+ 
                                		   “xmlns=“+“‘http://www.w3.org/TR/REC-html40‘>“+readFile（filepath）+““;
                                   byte b[] = content.getBytes（）;
                                   bais = new ByteArrayInputStream（b）;
                                   POIFSFileSystem poifs = new POIFSFileSystem（）;
                                   DirectoryEntry directory = poifs.getRoot（）;
                                   DocumentEntry documentEntry = directory.createDocument（“WordDocument“ bais）;
                                   fos = new FileOutputStream（path + “temp.doc“）;
                                   poifs.writeFilesystem（fos）;
                                   bais.close（）;
                                   fos.close（）;
                            }
                     }
 
              } catch （IOException e） {
                     e.printStackTrace（）;
              } finally {
                     if（fos != null） fos.close（）;
                     if（bais != null） bais.close（）;
              }
              return flag;
       }
 
       /**
        * 读取html文件到字符串
        * @param filename
        * @return
        * @throws Exception
        */
       public String readFile（String filename） throws Exception {
              StringBuffer buffer = new StringBuffer（““）;
              BufferedReader br = null;
              try {
                     br = new BufferedReader（new FileReader（filename））;
                     buffer = new StringBuffer（）;
                     while （br.re

 属性            大小     日期    时间   名称
----------- ---------  ---------- -----  ----

     文件       3826  2014-08-08 10:09  html转Word与PDFjar包与代码\代码\Test.java

     文件       3492  2014-08-07 12:36  html转Word与PDFjar包与代码\代码\HtmlToDoc.java

     文件     335536  2014-08-08 09:46  html转Word与PDFjar包与代码\转wordjar包\pd4ml.jar

     文件     123328  2014-08-08 09:46  html转Word与PDFjar包与代码\转wordjar包\ss_css2.jar

     文件   26762300  2014-08-08 09:46  html转Word与PDFjar包与代码\转wordjar包\fonts.jar

     文件    1675036  2014-08-08 10:06  html转Word与PDFjar包与代码\转PDFjar包\poi-3.7-20101029.jar

     目录          0  2014-08-08 10:08  html转Word与PDFjar包与代码\代码

     目录          0  2014-08-08 12:47  html转Word与PDFjar包与代码\转wordjar包

     目录          0  2014-08-08 12:48  html转Word与PDFjar包与代码\转PDFjar包

     目录          0  2014-08-08 09:53  html转Word与PDFjar包与代码

----------- ---------  ---------- -----  ----

             28903518                    10