POI-word转html

  1 package com.test.poiword;
  2 
  3 import android.app.Activity;
  4 import android.os.Bundle;
  5 import android.webkit.WebSettings;
  6 import android.webkit.WebView;
  7 
  8 import com.test.poiword.utils.FileUtils;
  9 
 10 import org.apache.poi.hwpf.HWPFDocument;
 11 import org.apache.poi.hwpf.converter.PicturesManager;
 12 import org.apache.poi.hwpf.converter.WordToHtmlConverter;
 13 import org.apache.poi.hwpf.usermodel.Picture;
 14 import org.apache.poi.hwpf.usermodel.PictureType;
 15 import org.w3c.dom.Document;
 16 
 17 import java.io.BufferedWriter;
 18 import java.io.ByteArrayOutputStream;
 19 import java.io.File;
 20 import java.io.FileInputStream;
 21 import java.io.FileNotFoundException;
 22 import java.io.FileOutputStream;
 23 import java.io.IOException;
 24 import java.io.OutputStreamWriter;
 25 import java.util.List;
 26 
 27 import javax.xml.parsers.DocumentBuilderFactory;
 28 import javax.xml.transform.OutputKeys;
 29 import javax.xml.transform.Transformer;
 30 import javax.xml.transform.TransformerFactory;
 31 import javax.xml.transform.dom.DOMSource;
 32 import javax.xml.transform.stream.StreamResult;
 33 
 34 /**
 35  * 
 36  */
 37 public class WordHtmlActivity extends Activity {
 38     //文件存储位置
 39     private String docPath = "/mnt/sdcard/doc/";
 40     //文件名称
 41     private String docName = "word3.doc";
 42     //html文件存储位置
 43     private String savePath = "/mnt/sdcard/doc/";
 44     @Override
 45     protected void onCreate(Bundle savedInstanceState) {
 46         super.onCreate(savedInstanceState);
 47         setContentView(R.layout.html);
 48         String name = docName.substring(0, docName.indexOf("."));
 49         try {
 50             convert2Html(docPath + docName, savePath + name + ".html");
 51         } catch (Exception e) {
 52             e.printStackTrace();
 53         }
 54         //WebView加载显示本地html文件
 55         WebView webView = (WebView)this.findViewById(R.id.office);
 56         WebSettings webSettings = webView.getSettings();
 57         webSettings.setLoadWithOverviewMode(true);
 58         webSettings.setSupportZoom(true);
 59         webSettings.setBuiltInZoomControls(true);
 60         webView.loadUrl("file:/"+savePath+name+".html");
 61     }
 62 
 63     /**
 64      * word文档转成html格式
 65      * */
 66     public void convert2Html(String fileName, String outPutFile) {
 67         HWPFDocument wordDocument = null;
 68         try {
 69             wordDocument = new HWPFDocument(new FileInputStream(fileName));
 70             WordToHtmlConverter wordToHtmlConverter = new WordToHtmlConverter(
 71                     DocumentBuilderFactory.newInstance().newDocumentBuilder().newDocument());
 72             //设置图片路径
 73             wordToHtmlConverter.setPicturesManager(new PicturesManager() {
 74                 public String savePicture(byte[] content,
 75                                           PictureType pictureType, String suggestedName,
 76                                           float widthInches, float heightInches) {
 77                     String name = docName.substring(0, docName.indexOf("."));
 78                     return name + "/" + suggestedName;
 79                 }
 80             });
 81             //保存图片
 82             List<Picture> pics=wordDocument.getPicturesTable().getAllPictures();
 83             if(pics!=null){
 84                 for(int i=0;i<pics.size();i++){
 85                     Picture pic = (Picture)pics.get(i);
 86                     System.out.println( pic.suggestFullFileName());
 87                     try {
 88                         String name = docName.substring(0,docName.indexOf("."));
 89                         String file = savePath+ name + "/"
 90                                 + pic.suggestFullFileName();
 91                         FileUtils.makeDirs(file);
 92                         pic.writeImageContent(new FileOutputStream(file));
 93                     } catch (FileNotFoundException e) {
 94                         e.printStackTrace();
 95                     }
 96                 }
 97             }
 98             wordToHtmlConverter.processDocument(wordDocument);
 99             Document htmlDocument = wordToHtmlConverter.getDocument();
100             ByteArrayOutputStream out = new ByteArrayOutputStream();
101             DOMSource domSource = new DOMSource(htmlDocument);
102             StreamResult streamResult = new StreamResult(out);
103 
104             TransformerFactory tf = TransformerFactory.newInstance();
105             Transformer serializer = tf.newTransformer();
106             serializer.setOutputProperty(OutputKeys.ENCODING, "utf-8");
107             serializer.setOutputProperty(OutputKeys.INDENT, "yes");
108             serializer.setOutputProperty(OutputKeys.METHOD, "html");
109             serializer.transform(domSource, streamResult);
110             out.close();
111             //保存html文件
112             writeFile(new String(out.toByteArray()), outPutFile);
113         } catch (Exception e) {
114             e.printStackTrace();
115         }
116     }
117 
118     /**
119      * 将html文件保存到sd卡
120      * */
121     public void writeFile(String content, String path) {
122         FileOutputStream fos = null;
123         BufferedWriter bw = null;
124         try {
125             File file = new File(path);
126             if(!file.exists()){
127                 file.createNewFile();
128             }
129             fos = new FileOutputStream(file);
130             bw = new BufferedWriter(new OutputStreamWriter(fos,"utf-8"));
131             bw.write(content);
132         } catch (FileNotFoundException fnfe) {
133             fnfe.printStackTrace();
134         } catch (IOException ioe) {
135             ioe.printStackTrace();
136         } finally {
137             try {
138                 if (bw != null)
139                     bw.close();
140                 if (fos != null)
141                     fos.close();
142             } catch (IOException ie) {
143             }
144         }
145     }
146 }

http://poi.apache.org/apidocs/overview-summary.html

原文地址:https://www.cnblogs.com/royi123/p/5202669.html