求poi将word转换为html文件的例子

搞什么哦 2010-03-31 12:58:59

...全文

460 3 打赏收藏转发到动态举报

写回复

用AI写文章

3 条回复

切换为时间正序

请发表友善的回复…

发表回复

F1226759 2012-10-27

打赏
举报

package com.test;
import java.io.File;
import java.io.FileInputStream;
import java.io.ObjectInputStream.GetField;

import org.apache.poi.hpbf.model.MainContents;
import org.apache.poi.hwpf.HWPFDocument;
import org.apache.poi.hwpf.model.PicturesTable;
import org.apache.poi.hwpf.usermodel.CharacterRun;
import org.apache.poi.hwpf.usermodel.Range;
/**
*
* @author 张廷下午
*
*/
public class WordToHtml {
/**
* 回车符ASCII码
*/
private static final short ENTER_ASCII = 13;
/**
* 空格符ASCII码
*/
private static final short SPACE_ASCII = 32;
/**
* 水平制表符ASCII码
*/
private static final short TABULATION_ASCII = 9;
private String htmlText = "";
/**
* 读取每个文字样式
*
* @param fileName
* @throws Exception
*/
public void getWordAndStyle(String fileName) throws Exception {
FileInputStream in = new FileInputStream(new File(fileName));
HWPFDocument doc = new HWPFDocument(in);
// 取得文档中字符的总数
int length = doc.characterLength();
// 创建图片容器
PicturesTable pTable = doc.getPicturesTable();
htmlText = "<html><head><title>" + doc.getSummaryInformation().getTitle() + "</title></head><body>";
// 创建临时字符串,好加以判断一串字符是否存在相同格式
String tempString = "";
for (int i = 0; i < length - 1; i++) {
// 整篇文章的字符通过一个个字符的来判断,range为得到文档的范围
Range range = new Range(i, i + 1, doc);
CharacterRun cr = range.getCharacterRun(0);
if (pTable.hasPicture(cr)) {
// 读写图片
//暂时不读取
this.readPicture(pTable, cr);
this.readPicture(pTable, cr);
} else {
Range range2 = new Range(i + 1, i + 2, doc);
// 第二个字符
CharacterRun cr2 = range2.getCharacterRun(0);
// 当前字符
char currentChar = cr.text().charAt(0);
// 判断是否为回车符
if (currentChar == ENTER_ASCII)
tempString += "<br/>";
// 判断是否为空格符
else if (currentChar == SPACE_ASCII)
tempString += " ";
// 判断是否为水平制表符
else if (currentChar == TABULATION_ASCII)
tempString += " ";
// 比较前后2个字符是否具有相同的格式
boolean flag = compareCharStyle(cr, cr2);
String fontStyle = "<span style='font-family:" + cr.getFontName() + ";font-size:" + cr.getFontSize() / 2 + "pt;";
if (cr.isBold())
fontStyle += "font-weight:bold;";
if (cr.isItalic())
fontStyle += "font-style:italic;";
if (flag && i != length - 2)
tempString += currentChar;
else if (!flag) {
htmlText += fontStyle + "'>" + tempString + currentChar + "</span>";
tempString = "";
} else
htmlText += fontStyle + "'>" + tempString + currentChar + "</span>";
}
htmlText += "</body></html>";
this.writeFile(htmlText);
}
}

}

maggiehexu 2011-07-15