用java语言通过POI实现word文档的按标题提取

  1 import java.io.FileInputStream;
  2 import java.io.FileOutputStream;
  3 import java.io.IOException;
  4 import java.io.InputStream;
  5 import java.io.OutputStream;
  6 import java.math.BigInteger;
  7 import java.util.HashMap;
  8 import java.util.List;
  9 import java.util.Map;
 10 
 11 import org.apache.poi.xwpf.usermodel.XWPFDocument;
 12 import org.apache.poi.xwpf.usermodel.XWPFParagraph;
 13 import org.apache.poi.xwpf.usermodel.XWPFRun;
 14 import org.apache.poi.xwpf.usermodel.XWPFStyle;
 15 import org.apache.poi.xwpf.usermodel.XWPFStyles;
 16 import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTDecimalNumber;
 17 import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTOnOff;
 18 import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTPPr;
 19 import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTString;
 20 import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTStyle;
 21 import org.openxmlformats.schemas.wordprocessingml.x2006.main.STStyleType;
 22 
 23 
 24 public class test {
 25 private static Map<String,Map<String,Object>> orderMap =new HashMap<String, Map<String,Object>>();
 26 
 27     public void init(String targetPath,String sourcePath){
 28         InputStream is = null;
 29         XWPFDocument doc=null;
 30         OutputStream out=null;
 31         try {
 32             XWPFDocument createDoc = new XWPFDocument();
 33 
 34             is = new FileInputStream(sourcePath);
 35             doc = new XWPFDocument(is);
 36             //获取段落
 37             List<XWPFParagraph> paras=doc.getParagraphs();
 38 
 39             for (XWPFParagraph para : paras){
 40  //             System.out.println(para.getCTP());//得到xml格式
 41               System.out.println(para.getStyleID());//段落级别
 42               System.out.println(para.getParagraphText());//段落内容
 43 
 44                 String titleLvl = getTitleLvl(doc,para);//获取段落级别
 45                 if("a5".equals(titleLvl)||"HTML".equals(titleLvl)||"".equals(titleLvl)||null==titleLvl){
 46                     titleLvl = "8";
 47                 }
 48               System.out.println(titleLvl+"-----");//0,1,2
 49                 if(!"8".equals(titleLvl)){
 50                     System.out.println(titleLvl+"===="+para.getParagraphText());
 51                 }
 52 
 53 
 54                 XWPFParagraph ctPara = createDoc.createParagraph();
 55                 //一个XWPFRun代表具有相同属性的一个区域。
 56                 XWPFRun ctRun = ctPara.createRun();
 57                 String ctText = para.getParagraphText();
 58                 ctRun.setFontFamily("宋体");//字体
 59                 ctRun.setFontSize(12);
 60 
 61                 if(null!=titleLvl&&!"".equals(titleLvl)&&!"8".equals(titleLvl)){
 62                     addCustomHeadingStyle(createDoc,titleLvl,Integer.parseInt(titleLvl));
 63                     String orderCode = getOrderCode(titleLvl);//获取编号
 64                     ctText = orderCode+" "+ctText;
 65                     ctRun.setBold(true);//标题加粗
 66                     ctRun.setFontSize(14);
 67 
 68                     ctPara.setStyle(titleLvl);
 69 
 70                 }else{//正文
 71                     ctPara.setIndentationFirstLine(567);//首行缩进：567==1厘米
 72 //                  ctRun.setTextPosition(6);//设置行间距
 73                 }
 74 
 75                 ctRun.setText(ctText);//内容
 76             }
 77             out=new FileOutputStream(targetPath);
 78             createDoc.write(out);
 79         } catch (Exception e) {
 80             e.printStackTrace();
 81         } finally{
 82             try {
 83                 if(null!=out){
 84                     out.close();
 85                 } 
 86                 if(null!=is){
 87                     is.close();
 88                 }
 89             }catch (IOException e) {
 90                 e.printStackTrace();
 91             }
 92         }
 93     }
 94 
 95     /**
 96      * Word中的大纲级别，可以通过getPPr().getOutlineLvl()直接提取，但需要注意，Word中段落级别，通过如下三种方式定义： 
 97      *  1、直接对段落进行定义； 
 98      *  2、对段落的样式进行定义； 
 99      *  3、对段落样式的基础样式进行定义。 
100      *  因此，在通过“getPPr().getOutlineLvl()”提取时，需要依次在如上三处读取。
101      * @param doc
102      * @param para
103      * @return
104      */
105     private static String getTitleLvl(XWPFDocument doc, XWPFParagraph para) {
106         String titleLvl = "";
107         try {
108             //判断该段落是否设置了大纲级别
109             if (para.getCTP().getPPr().getOutlineLvl() != null) {
110                 // System.out.println("getCTP()");
111 //              System.out.println(para.getParagraphText());
112 //              System.out.println(para.getCTP().getPPr().getOutlineLvl().getVal());
113 
114                 return String.valueOf(para.getCTP().getPPr().getOutlineLvl().getVal());
115             }
116         } catch (Exception e) {
117 
118         }
119 
120         try {
121             //判断该段落的样式是否设置了大纲级别
122             if (doc.getStyles().getStyle(para.getStyle()).getCTStyle().getPPr().getOutlineLvl() != null) {
123 
124                 // System.out.println("getStyle");
125 //              System.out.println(para.getParagraphText());
126 //              System.out.println(doc.getStyles().getStyle(para.getStyle()).getCTStyle().getPPr().getOutlineLvl().getVal());
127 
128                 return String.valueOf(doc.getStyles().getStyle(para.getStyle()).getCTStyle().getPPr().getOutlineLvl().getVal());
129             }
130         } catch (Exception e) {
131 
132         }
133 
134         try {
135             //判断该段落的样式的基础样式是否设置了大纲级别
136             if (doc.getStyles().getStyle(doc.getStyles().getStyle(para.getStyle()).getCTStyle().getBasedOn().getVal())
137                     .getCTStyle().getPPr().getOutlineLvl() != null) {
138                 // System.out.println("getBasedOn");
139 //              System.out.println(para.getParagraphText());
140                 String styleName = doc.getStyles().getStyle(para.getStyle()).getCTStyle().getBasedOn().getVal();
141 //              System.out.println(doc.getStyles().getStyle(styleName).getCTStyle().getPPr().getOutlineLvl().getVal());
142 
143                 return String.valueOf(doc.getStyles().getStyle(styleName).getCTStyle().getPPr().getOutlineLvl().getVal());
144             }
145         } catch (Exception e) {
146 
147         }
148 
149         try {
150             if(para.getStyleID()!=null){
151                 return para.getStyleID();
152             }
153         } catch (Exception e) {
154 
155         }
156 
157         return titleLvl;
158     }
159 
160     /**
161      * 增加自定义标题样式。这里用的是stackoverflow的源码
162      * 
163      * @param docxDocument 目标文档
164      * @param strStyleId 样式名称
165      * @param headingLevel 样式级别
166      */
167     private static void addCustomHeadingStyle(XWPFDocument docxDocument, String strStyleId, int headingLevel) {
168 
169         strStyleId = String.valueOf(Integer.parseInt(strStyleId)+1);
170         CTStyle ctStyle = CTStyle.Factory.newInstance();
171         ctStyle.setStyleId(strStyleId);
172 
173         CTString styleName = CTString.Factory.newInstance();
174         styleName.setVal(strStyleId);
175         ctStyle.setName(styleName);
176 
177         CTDecimalNumber indentNumber = CTDecimalNumber.Factory.newInstance();
178         indentNumber.setVal(BigInteger.valueOf(headingLevel));
179 
180         // lower number > style is more prominent in the formats bar
181         ctStyle.setUiPriority(indentNumber);
182 
183         CTOnOff onoffnull = CTOnOff.Factory.newInstance();
184         ctStyle.setUnhideWhenUsed(onoffnull);
185 
186         // style shows up in the formats bar
187         ctStyle.setQFormat(onoffnull);
188 
189         // style defines a heading of the given level
190         CTPPr ppr = CTPPr.Factory.newInstance();
191         ppr.setOutlineLvl(indentNumber);
192         ctStyle.setPPr(ppr);
193 
194         XWPFStyle style = new XWPFStyle(ctStyle);
195 
196         // is a null op if already defined
197         XWPFStyles styles = docxDocument.createStyles();
198 
199         style.setType(STStyleType.PARAGRAPH);
200         styles.addStyle(style);
201 
202     }
203     /**
204      * 获取标题编号
205      * @param titleLvl
206      * @return
207      */
208     private static String getOrderCode(String titleLvl) {
209         String order = "";
210 
211         if("0".equals(titleLvl)||Integer.parseInt(titleLvl)==8){//文档标题||正文
212             return "";
213         }else if(Integer.parseInt(titleLvl)>0&&Integer.parseInt(titleLvl)<8){//段落标题
214 
215             //设置最高级别标题
216             Map<String,Object> maxTitleMap = orderMap.get("maxTitleLvlMap");
217             if(null==maxTitleMap){//没有，表示第一次进来
218                 //最高级别标题赋值
219                 maxTitleMap = new HashMap<String, Object>();
220                 maxTitleMap.put("lvl", titleLvl);
221                 orderMap.put("maxTitleLvlMap", maxTitleMap);
222             }else{
223                 String maxTitleLvl = maxTitleMap.get("lvl")+"";//最上层标题级别(0,1,2,3)
224                 if(Integer.parseInt(titleLvl)<Integer.parseInt(maxTitleLvl)){//当前标题级别更高
225                     maxTitleMap.put("lvl", titleLvl);//设置最高级别标题
226                     orderMap.put("maxTitleLvlMap", maxTitleMap);
227                 }
228             }
229 
230             //查父节点标题
231             int parentTitleLvl = Integer.parseInt(titleLvl)-1;//父节点标题级别
232             Map<String,Object> cMap = orderMap.get(titleLvl);//当前节点信息
233             Map<String,Object> pMap = orderMap.get(parentTitleLvl+"");//父节点信息
234 
235             if(0==parentTitleLvl){//父节点为文档标题，表明当前节点为1级标题
236                 int count= 0;
237                 //最上层标题，没有父节点信息
238                 if(null==cMap){//没有当前节点信息
239                     cMap = new HashMap<String, Object>();
240                 }else{
241                     count = Integer.parseInt(String.valueOf(cMap.get("cCount")));//当前序个数
242                 }
243                 count++;
244                 order = count+"";
245                 cMap.put("cOrder", order);//当前序
246                 cMap.put("cCount", count);//当前序个数
247                 orderMap.put(titleLvl, cMap);
248 
249             }else{//父节点为非文档标题
250                 int count= 0;
251                 //如果没有相邻的父节点信息，当前标题级别自动升级
252                 if(null==pMap){
253                     return getOrderCode(String.valueOf(parentTitleLvl));
254                 }else{
255                     String pOrder = String.valueOf(pMap.get("cOrder"));//父节点序
256                     if(null==cMap){//没有当前节点信息
257                         cMap = new HashMap<String, Object>();
258                     }else{
259                         count = Integer.parseInt(String.valueOf(cMap.get("cCount")));//当前序个数
260                     }
261                     count++;
262                     order = pOrder+"."+count;//当前序编号
263                     cMap.put("cOrder", order);//当前序
264                     cMap.put("cCount", count);//当前序个数
265                     orderMap.put(titleLvl, cMap);
266                 }
267             }
268 
269             //字节点标题计数清零
270             int childTitleLvl = Integer.parseInt(titleLvl)+1;//子节点标题级别
271             Map<String,Object> cdMap = orderMap.get(childTitleLvl+"");//
272             if(null!=cdMap){
273                 cdMap.put("cCount", 0);//子节点序个数
274                 orderMap.get(childTitleLvl+"").put("cCount", 0);
275             }
276         }
277         return order;
278     }
279 
280     public static void main(String[] args) {
281         InputStream is = null;
282         XWPFDocument doc=null;
283         OutputStream out=null;
284         String[] title= new String [276];
285         String[] concent= new String [276];
286         String[] type= new String [276];
287         int i=0;
288         try {
289             XWPFDocument createDoc = new XWPFDocument();
290 
291             is = new FileInputStream("E:/doc/a.docx");
292             doc = new XWPFDocument(is);
293             //获取段落
294             List<XWPFParagraph> paras=doc.getParagraphs();
295 
296             for (XWPFParagraph para : paras){
297               //System.out.println(para.getCTP());//得到xml格式
298               //System.out.println(para.getStyleID());//段落级别
299              // System.out.println(para.getParagraphText());//段落内容
300               String titleLvl = getTitleLvl(doc,para);//获取段落级别
301               if("a5".equals(titleLvl)||"HTML".equals(titleLvl)||"".equals(titleLvl)||null==titleLvl){
302                   titleLvl = "8";
303               }
304             //System.out.println(titleLvl+"-----");//0,1,2
305              if(!"8".equals(titleLvl)){
306                 //System.out.println(titleLvl+"===="+para.getParagraphText());
307                 
308                 if("3".equals(titleLvl)) {
309                     
310                     if(concent[i]!=null)
311                     concent[i]=concent[i]+para.getParagraphText();
312                     else
313                     concent[i]=para.getParagraphText();
314                     //System.out.println(concent[i]);
315                 }
316                 if("2".equals(titleLvl)) {
317                     i++;
318                     title[i]=para.getParagraphText();
319                     type[i]=type[i-1];
320                     //System.out.println(title[i]);
321                 }
322                 if("1".equals(titleLvl)) {
323                     i++;
324                     type[i]=para.getParagraphText();     
325                     //System.out.println(title[i]);
326                 }
327             }
328             
329 
330             }
331             for(int j=2;j<title.length;j++) {
332                 if(title[j]!=null) {
333                     String sql = "INSERT INTO shuju (title,concent,type)VALUES (‘"+title[j]+"‘,‘"+concent[j]+"‘,‘"+type[j]+"‘)";
334                     DBUtil jdbc=new DBUtil();
335                     int result=jdbc.executeUpdate(sql);
336                     jdbc.close();
337                     //System.out.println(type[j]);
338                 //System.out.println(title[j]);
339                 //System.out.println(concent[j]);
340                 }
341             }
342         } catch (Exception e) {
343             e.printStackTrace();
344         } finally{
345             try {
346                 if(null!=out){
347                     out.close();
348                 } 
349                 if(null!=is){
350                     is.close();
351                 }
352             }catch (IOException e) {
353                 e.printStackTrace();
354             }
355         }
356     }
357 }

  1 package com.poi.test;
  2  
  3 import java.io.File;
  4 import java.io.FileInputStream;
  5 import java.io.InputStream;
  6  
  7 import org.apache.poi.POIXMLDocument;
  8 import org.apache.poi.POIXMLTextExtractor;
  9 import org.apache.poi.hwpf.extractor.WordExtractor;
 10 import org.apache.poi.openxml4j.opc.OPCPackage;
 11 import org.apache.poi.xwpf.extractor.XWPFWordExtractor;
 12  
 13 public class testPoi {
 14     /**
 15      * 读取word文件内容
 16      * 
 17      * @param path
 18      * @return buffer
 19      */
 20  
 21     public String readWord(String path) {
 22         String buffer = "";
 23         try {
 24             if (path.endsWith(".doc")) {
 25                 InputStream is = new FileInputStream(new File(path));
 26                 WordExtractor ex = new WordExtractor(is);
 27                 buffer = ex.getText();
 28                 ex.close();
 29             } else if (path.endsWith("docx")) {
 30                 OPCPackage opcPackage = POIXMLDocument.openPackage(path);
 31                 POIXMLTextExtractor extractor = new XWPFWordExtractor(opcPackage);
 32                 buffer = extractor.getText();
 33                 extractor.close();
 34             } else {
 35                 System.out.println("此文件不是word文件！");
 36             }
 37  
 38         } catch (Exception e) {
 39             e.printStackTrace();
 40         }
 41  
 42         return buffer;
 43     }
 44  
 45     public static void main(String[] args) {
 46         // TODO Auto-generated method stub
 47         testPoi tp = new testPoi();
 48         String content = tp.readWord("自己的路径.doc");
 49         //String arr[]=content.split("\\d+");
 50         //String arr[]=content.split("第"+"\\w"+"章");
 51         String arr[]=content.split("\\r\\n");
 52         /*String[] a=arr[13].split("\\d+");
 53         String[] b=a[1].split("\\s+");
 54         System.out.println(b[1]);*/
 55         String[] reci = new String [276];;
 56         for(int i=12,j=0;i<290;i++,j++) {
 57             arr[i]=arr[i]+"1";
 58             if(!arr[i].equals("1")) {
 59             
 60             if(i<27) {//判断页面数是否为单数
 61                 String[] a=arr[i].split("\\d+|\\s+");
 62             if(arr[i]!="\\s+") {//判断该元素是否为连续空格
 63                 if(a.length==2) {//判断该元素是否为标题即分割成2个段
 64                     reci[j]=a[1];
 65                 System.out.println(a[1]);
 66                 }
 67                 else if(a.length==1) {
 68                     reci[j]=a[1];
 69                 System.out.println(arr[i]);
 70                 }
 71                 else//否则该元素是平常元素可以分割成3个段
 72                     {
 73                     reci[j]=a[2];
 74                     System.out.println(a[2]);
 75                     }
 76                 }
 77         }
 78             else {
 79                 String[] a=arr[i].split("\\d{2,3}|\\s+|\\t");
 80                     if(arr[i]!="\\s+") {//判断该元素是否为连续空格
 81                         if(a.length==2) {
 82                             reci[j]=a[1];
 83                             System.out.println(a[1]);
 84                         }
 85                         else if(a.length==1) {
 86                             reci[j]=a[1];
 87                             System.out.println(arr[i]+i);
 88                         }
 89                         else if(a.length==4) {
 90                             reci[j]=a[1];
 91                             System.out.println(a[1]);
 92                         }
 93                         else
 94                             {reci[j]=a[2];
 95                             System.out.println(a[2]);
 96                             }
 97                     }
 98             }
 99         }
100     }
101         String fengefu=reci[0];
102         for(int i=1;i<276;i++) {
103             if(reci[i]!=null)
104         fengefu=fengefu+"|"+reci[i];
105 
106         }
107         System.out.println(reci[275]);
108         System.out.println(fengefu);
109         String arr2[]=content.split(fengefu);
110         for(int i=0;i<200;i++)
111             System.out.println(arr2[i]);
112     }
113 }

 1 package com.xxx.util;
 2  
 3 import java.io.File;
 4 import java.io.FileInputStream;
 5 import java.io.IOException;
 6  
 7 import org.apache.poi.hwpf.extractor.WordExtractor;
 8  
 9 public class DocUtil {
10     /**
11      * 读取doc文件内容
12      * 
13      * @param file
14      *            想要读取的文件对象
15      * @return 返回文件内容
16      * @throws IOException
17      */
18     public static String doc2String(FileInputStream fs) throws IOException {
19         StringBuilder result = new StringBuilder();
20         WordExtractor re = new WordExtractor(fs);
21         result.append(re.getText());
22         re.close();
23         return result.toString();
24     }
25  
26     public static String doc2String(File file) throws IOException {
27         return doc2String(new FileInputStream(file));
28     }
29  
30     public static void main(String[] args) {
31         File file = new File("自己的路径.doc");
32         try {
33             System.out.println(doc2String(file));
34         } catch (IOException e) {
35             e.printStackTrace();
36         }
37     }
38 }