文档库--解析docx

jero-boot/jero-boot-modules/src/main/java/com/jero/modules/document/service/impl/BussDocumentLibraryEOServiceImpl.java

@@ -1729,7 +1729,7 @@ public class BussDocumentLibraryEOServiceImpl extends ServiceImpl<BussDocumentLi
 
                         String wordContent = null;
                         try {
-                            wordContent = ReadWordUtil.readWord(ossFile.getUrl());
+                            wordContent = ReadWordUtil.readWord(ossFile.getUrl(),uploadpath);
                         } catch (Exception e) {
                             log.error("文档库: 读取word内容失败");
                         }
@@ -1872,7 +1872,7 @@ public class BussDocumentLibraryEOServiceImpl extends ServiceImpl<BussDocumentLi
                     || ossFile.getFileName().endsWith(".DOCX")) {
                 String wordContent = null;
                 try {
-                    wordContent = ReadWordUtil.readWord(ossFile.getUrl());
+                    wordContent = ReadWordUtil.readWord(ossFile.getUrl(),uploadpath);
                 } catch (Exception e) {
                     log.error("文档库:读取word内容失败");
                 }
@@ -2665,7 +2665,7 @@ public class BussDocumentLibraryEOServiceImpl extends ServiceImpl<BussDocumentLi
                     || ossFile.getFileName().endsWith(".DOCX")) {
                 String wordContent = null;
                 try {
-                    wordContent = ReadWordUtil.readWord(ossFile.getUrl());
+                    wordContent = ReadWordUtil.readWord(ossFile.getUrl(),uploadpath);
                 } catch (Exception e) {
                     log.error("文档库:读取word文件内容失败");
                 }

jero-boot/jero-boot-modules/src/main/java/com/jero/modules/document/utils/ReadWordUtil.java

@@ -1,6 +1,7 @@
 package com.jero.modules.document.utils;
 
 import com.jero.common.exception.JeroBootException;
+import com.jero.common.util.oss.CosBootUtil;
 import org.apache.poi.POIXMLDocument;
 import org.apache.poi.POIXMLTextExtractor;
 import org.apache.poi.hwpf.extractor.WordExtractor;
@@ -8,32 +9,50 @@ import org.apache.poi.openxml4j.opc.OPCPackage;
 import org.apache.poi.xwpf.extractor.XWPFWordExtractor;
 
 import java.io.File;
-import java.io.FileInputStream;
 import java.io.InputStream;
 
+import static com.jero.modules.document.service.impl.BussDocumentLibraryEOServiceImpl.copyFile;
+
 /**
  * @description
  * @date 2022/3/21 17:04
  * @auth zhn
  */
 public class ReadWordUtil {
-
+    public static String readWord(String path,String uploadpath) {
-    public static String readWord(String path) {
 //        String path = "E:\\DeskTop\\1111.doc";
 //        path = "E:\\DeskTop\\22222.docx";
 //        String path = "E:\\DeskTop\\标准所ISO-用户手册.doc";
         String str = "";
         try {
             if (path.endsWith(".doc")) {
-                InputStream is = new FileInputStream(new File(path));
+                if(CosBootUtil.doesObjectExist(path)){
-                WordExtractor ex = new WordExtractor(is);
+                    InputStream in = CosBootUtil.download(path);
-                str = ex.getText();
+                    WordExtractor ex = new WordExtractor(in);
-                ex.close();
+                    str = ex.getText();
+                    ex.close();
+                }
             } else if (path.endsWith("docx")) {
-                OPCPackage opcPackage = POIXMLDocument.openPackage(path);
+                //先把文件存到本地, 用完后在删除
-                POIXMLTextExtractor extractor = new XWPFWordExtractor(opcPackage);
+                if(CosBootUtil.doesObjectExist(path)){
-                str = extractor.getText();
+                    InputStream in = CosBootUtil.download(path);
-                extractor.close();
+                    String fileNowPath = uploadpath + "/tempZip/" ;
+                    File file = new File(fileNowPath);
+                    if (file.exists()) {
+                        file.delete();
+                    }
+                    file.mkdirs();
+                    File fileOld = new File(path);
+                    long currentTime = System.currentTimeMillis();
+                    copyFile(in, fileNowPath + File.separator + currentTime + fileOld.getName());
+                    File fileTemp = new File(fileNowPath + File.separator + currentTime + fileOld.getName());
+                    String pathTemp = fileTemp.getPath();
+                    OPCPackage opcPackage = POIXMLDocument.openPackage(pathTemp);
+                    POIXMLTextExtractor extractor = new XWPFWordExtractor(opcPackage);
+                    str = extractor.getText().trim();
+                    extractor.close();
+                    fileTemp.delete();
+                }
             } else {
                 throw new JeroBootException("此文件不是word文件");
             }
@@ -44,4 +63,30 @@ public class ReadWordUtil {
             throw new JeroBootException("文件内容读取失败");
         }
     }
+//    public static String readWord(String path) {
+////        String path = "E:\\DeskTop\\1111.doc";
+////        path = "E:\\DeskTop\\22222.docx";
+////        String path = "E:\\DeskTop\\标准所ISO-用户手册.doc";
+//        String str = "";
+//        try {
+//            if (path.endsWith(".doc")) {
+//                InputStream is = new FileInputStream(new File(path));
+//                WordExtractor ex = new WordExtractor(is);
+//                str = ex.getText();
+//                ex.close();
+//            } else if (path.endsWith("docx")) {
+//                OPCPackage opcPackage = POIXMLDocument.openPackage(path);
+//                POIXMLTextExtractor extractor = new XWPFWordExtractor(opcPackage);
+//                str = extractor.getText();
+//                extractor.close();
+//            } else {
+//                throw new JeroBootException("此文件不是word文件");
+//            }
+//            return str;
+//        } catch (Exception e) {
+//            throw new JeroBootException("文件内容读取失败");
+//        }catch (NoSuchFieldError e){
+//            throw new JeroBootException("文件内容读取失败");
+//        }
+//    }
 }