HWPFDocument读取doc,wps文档（含图片读取）

使用HWPFDocument对象读取03版doc文件或wps文件

导包

 代码：

1、图片工具类

package com.poi.test;

import java.util.ArrayList;
import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Set;

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.Picture;
import org.apache.poi.hwpf.usermodel.Range;

/**
 * Provides access to the pictures both by offset, iteration over the
 * un-claimed, and peeking forward
 */
public class PicturesSource {//这个类是poi官网找的
    private PicturesTable picturesTable;
    private Set<Picture> output = new HashSet<Picture>();
    private Map<Integer, Picture> lookup;
    private List<Picture> nonU1based;
    private List<Picture> all;
    private int pn = 0;

    public PicturesSource(HWPFDocument doc) {
        picturesTable = doc.getPicturesTable();
        all = picturesTable.getAllPictures();

        // Build the Offset-Picture lookup map
        lookup = new HashMap<Integer, Picture>();
        for (Picture p : all) {
            lookup.put(p.getStartOffset(), p);
        }

        // Work out which Pictures aren't referenced by
        //  a \u0001 in the main text
        // These are \u0008 escher floating ones, ones
        //  found outside the normal text, and who
        //  knows what else...
        nonU1based = new ArrayList<Picture>();
        nonU1based.addAll(all);
        Range r = doc.getRange();
        for (int i = 0; i < r.numCharacterRuns(); i++) {
            CharacterRun cr = r.getCharacterRun(i);
            if (picturesTable.hasPicture(cr)) {
                Picture p = getFor(cr);
                int at = nonU1based.indexOf(p);
                nonU1based.set(at, null);
            }
        }
    }

    private boolean hasPicture(CharacterRun cr) {
        return picturesTable.hasPicture(cr);
    }

    private void recordOutput(Picture picture) {
        output.add(picture);
    }

    private boolean hasOutput(Picture picture) {
        return output.contains(picture);
    }

    private int pictureNumber(Picture picture) {
        return all.indexOf(picture) + 1;
    }

    public Picture getFor(CharacterRun cr) {
        return lookup.get(cr.getPicOffset());
    }

    /**
     * Return the next unclaimed one, used towards the end
     */
    private Picture nextUnclaimed() {
        Picture p = null;
        while (pn < nonU1based.size()) {
            p = nonU1based.get(pn);
            pn++;
            if (p != null)
                return p;
        }
        return null;
    }
}

2、处理图片和段落文字

package com.poi.test;

import java.io.ByteArrayOutputStream;
import java.io.File;
import java.io.FileInputStream;

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.Paragraph;
import org.apache.poi.hwpf.usermodel.Picture;
import org.apache.poi.hwpf.usermodel.Range;

public class PoiForWord {
    /**
     * 使用HWPFDocument解析word文档
     * wps按doc处理即可
     */
    public void parseDocByHWPFDocument(){
        try(FileInputStream is = new FileInputStream(new File("c:\\a.wps"));HWPFDocument document = new HWPFDocument(is);){
            ByteArrayOutputStream baos = new ByteArrayOutputStream();//字节流，用来存储图片
            PicturesSource pictures = new PicturesSource(document);
            PicturesTable pictureTable = document.getPicturesTable();
            
            Range r = document.getRange();//区间
            for(int i=0;i<r.numParagraphs();i++){
                Paragraph p = r.getParagraph(i);//段落
                int fontSize = p.getCharacterRun(0).getFontSize();//字号，字号和是否加粗可用来当做标题或者某一关键标识的判断
                   boolean isBold = p.getCharacterRun(0).isBold();//是否加粗
                String paragraphText = p.text();//段落文本
                
                //以下代码解析图片,这样获取的图片是在文档流中的，是和文本按顺序解析的，可以很好的解决图片定位问题
                for(int j=0;j<p.numCharacterRuns();j++){
                    CharacterRun cr = p.getCharacterRun(j);//字符
                    if(pictureTable.hasPicture(cr)){
                        Picture picture = pictures.getFor(cr);
                        //如果是在页面显示图片，可转换为base64编码的图片
                        picture.writeImageContent(baos);//将图片写入字节流
//                        String base64Image = "<img src='data:image/png;base64,"+new BASE64Encoder().encode(baos.toByteArray())+"'/>";
                    }
                }
            }
        }catch(Exception e){
            e.printStackTrace();
        }
    }
    
}

3、处理表格

 

 

/**
     * 使用HWPFDocument解析word文档
     * wps按doc处理即可
     */
    @Test
    public void parseDocTableByHWPFDocument(){
        try(FileInputStream is = new FileInputStream(new File("d:\\b.doc"));HWPFDocument document = new HWPFDocument(is);){
            Range r = document.getRange();//区间
            for(int i=0;i<r.numParagraphs();i++){
                Paragraph p = r.getParagraph(i);//段落
                String text = p.text();
                
                if(text.indexOf("序号")!=-1){//解析表格需要从表格第一个单元格获取表格，另一种表格的方式是直接获取所有表格，但是无法判断表格在文档中的位置
                    Table table = r.getTable(p);
                    
                    int numRows = table.numRows();//获取行数
                    
                    for(int j=0;j<numRows;j++){
                        TableRow row = table.getRow(j);
                        int numCells = row.numCells();//当前行列数
                        for(int k=0;k<numCells;k++){
                            TableCell cell = row.getCell(k);
                            System.out.print(cell.text()+" @ ");
                        }
                        System.out.println();
                    }
                }
            }
        }catch(Exception e){
            e.printStackTrace();
        }
    }

 

 字符"?"可通过字符串替换或截取来解决

 

另一种解析的方式，只支持解析文本内容，且无法获取字号和加粗等字体格式

WordExtractor extor = new WordExtractor(is);
            String[] paragraphText = extor.getParagraphText();