fix 文档拆分--公式

laws-modules/src/main/java/com/jero/modules/split/service/impl/FileSpiltService.java

@@ -198,17 +198,17 @@ public class FileSpiltService {
                         //} else
                         if (child instanceof CTOMath) {
                             CTOMath ctoMath = (CTOMath) child;
-                            stringBuilder.append(ctoMath.xmlText());
+                            //stringBuilder.append(ctoMath.xmlText());
                             // 处理word中存在的公式成mathML格式
-                            //handleParagraphOMath(stringBuilder, ctoMath);
+                            String imageName = OmmlUtils.convertOmathToPng(ctoMath.xmlText());
                             if (messageList.size() > 0) {
-                                addItermsConditionsMath(messageList.get(messageList.size() - 1), itemValList, stringBuilder.toString());
+                                addItermsMathImage(messageList.get(messageList.size() - 1), itemValList, imageName);
                             }
                         } else if (child instanceof CTOMathPara) {
                             CTOMathPara ctoMathPara = (CTOMathPara) child;
-                            stringBuilder.append(ctoMathPara.xmlText());
+                            String imageName = OmmlUtils.convertOmathToPng(ctoMathPara.xmlText());
                             if (messageList.size() > 0) {
-                                addItermsConditionsMath(messageList.get(messageList.size() - 1), itemValList, stringBuilder.toString());
+                                addItermsMathImage(messageList.get(messageList.size() - 1), itemValList, imageName);
                             }
                         }
                     }
@@ -276,22 +276,6 @@ public class FileSpiltService {
                             continue;
                         }
                     }
-                    //if (StringUtils.isNotEmpty(p.getNumLevelText()) && p.getNumLevelText().indexOf(".") > 0) {
-                    //    String numberParent = p.getNumLevelText().substring(0, p.getNumLevelText().lastIndexOf(".")).replaceAll("%","");
-                    //    if (numberMap.containsKey(numberParent)) {
-                    //        Integer numbernow = (Integer) numberMap.get(numberParent) + 1;
-                    //        if (messageList != null && !messageList.isEmpty() && messageList.get(messageList.size() - 1).getItemNum().equals(numberParent + "." + String.valueOf(numbernow))) {
-                    //            numberMap.put(numberParent, numbernow + 1);
-                    //            paragraphString = numberParent + "." + String.valueOf(numbernow + 1) + paragraphString;
-                    //        } else {
-                    //            numberMap.put(numberParent, numbernow);
-                    //            paragraphString = numberParent + "." + String.valueOf(numbernow) + paragraphString;
-                    //        }
-                    //    } else {
-                    //        numberMap.put(numberParent, 1);
-                    //        paragraphString = numberParent + ".1" + paragraphString;
-                    //    }
-                    //}
                     if (messageNumber == 0) {
                         //ptest = Pattern.compile("^1\\s{0,50}范围");
                         Matcher matcher = RANGE_PATTERN_1.matcher(paragraphString);
@@ -477,32 +461,6 @@ public class FileSpiltService {
         return "";
     }
 
-    /**
-     * 处理公式
-     *
-     * @param content
-     * @param child
-     */
-    private static void handleParagraphOMath(StringBuilder content, CTOMath child) {
-        log.info(">> 开始解析段落中的CTOMath公式");
-        //将word中的omml格式转换成mathML
-        String mathMLFromNode = MathmlUtils.convertOMML2MML(child.xmlText());
-        content.append(mathMLFromNode);
-    }
-
-    /**
-     * 处理公式(将word中的omML格式的公式转换成mathML格式的工单)
-     *
-     * @param content
-     * @param child
-     */
-    private static void handleParagraphOMath(StringBuilder content, CTOMathPara child) {
-        log.info(">> 开始解析段落中的CTOMathPara公式");
-        //将word中的omml格式转换成mathML
-        String mathMLFromNode = MathmlUtils.convertOMML2MML(child.xmlText());
-        content.append(mathMLFromNode);
-    }
-
     /**
      * 通过传递过来的编号，得出下次可能出现的编号
      *
@@ -1158,6 +1116,15 @@ public class FileSpiltService {
         return tableToHtmlStr.toString();
     }
 
+    // 像每个条款中依次插入每一段的内容
+    private void addItermsMathImage(SarFileSplitItemsEO message, List<SarFileSplitItemsValEO> itemsValList, String imageName) {
+        String path = splitUrl + imageName;
+        String imgConVal = "<img class=\"wordImg\" style=\"width:100%;height:100%\" src=\"" + path + "\">";
+        message.getItemsCondi().add(imgConVal);
+        message.setItemContent(message.getItemContent() + imgConVal);
+        itemsValList.add(getSplitItemsValObject(message.getId(), SplitFilePragraTypeEnum.IMG.getValue(), path, message.getItemsCondi().size(), null));
+    }
+
     // 像每个条款中依次插入每一段的内容
     private void addItermsConditionsImage(SarFileSplitItemsEO message, List<SarFileSplitItemsValEO> itemsValList, XWPFParagraph p, List<String> imageBundleList) {
         for (String pictureId : imageBundleList) {

laws-modules/src/main/java/com/jero/modules/split/service/impl/SarFileSplitMenuEOServiceImpl.java

@@ -126,7 +126,7 @@ public class SarFileSplitMenuEOServiceImpl extends ServiceImpl<SarFileSplitMenuE
     public void deleteMenu(String id){
         SarFileSplitMenuEO sarFileSplitMenuEO = dao.selectByPrimaryKey(id);
         // 水平越权
-        documentSplitService.isHorizontalOverstep(sarFileSplitMenuEO.getInfoId());
+        //documentSplitService.isHorizontalOverstep(sarFileSplitMenuEO.getInfoId());
         List<String> menuIds = new ArrayList<>();
         SarFileSplitMenuEO sarMenuEO = new SarFileSplitMenuEO();
         sarMenuEO.setId(id);

laws-modules/src/main/java/com/jero/modules/split/util/ImageParse.java

@@ -0,0 +1,54 @@
+package com.jero.modules.split.util;
+
+import com.jero.common.util.MinioUtil;
+import com.jero.common.util.UUIDUtils;
+import org.apache.commons.io.IOUtils;
+import org.springframework.beans.factory.annotation.Value;
+
+import java.io.ByteArrayInputStream;
+import java.io.FileOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+
+public class ImageParse {
+    @Value("${jero.path.uploadCos}")
+    private String filePathCos;
+    @Value("${jero.splitUrl}")
+    private String splitUrl;
+
+    private int number = 0;
+//    private String targetDir;
+
+    public ImageParse() {
+        super();
+//        this.targetDir = targetDir;
+    }
+
+    public String parse(byte[] data, String extName) {
+        String imageName = UUIDUtils.randomUUID10() + extName;
+
+        String filepathandname = filePathCos + "/" + imageName;
+        MinioUtil.uploadByte(data, filepathandname);
+        return imageName;
+    }
+
+    public String parse(InputStream in, String extName) {
+        if (extName.lastIndexOf(".") > -1) {
+            extName = extName.substring(extName.lastIndexOf(".") + 1);
+        }
+        String filename = "image_" + (number++) + "." + extName;
+//        File target = new File(targetDir);
+//        if (!target.exists()) {
+//            target.mkdirs();
+//        }
+
+        try {
+            IOUtils.copy(in, new FileOutputStream(filename));
+        } catch (IOException e) {
+            e.printStackTrace();
+        }
+        return filename;
+    }
+
+
+}

laws-modules/src/main/java/com/jero/modules/split/util/MathmlUtils.java

@@ -1,113 +0,0 @@
-package com.jero.modules.split.util;
-
-import javax.xml.transform.*;
-import javax.xml.transform.stream.StreamResult;
-import javax.xml.transform.stream.StreamSource;
-import java.io.InputStream;
-import java.io.StringReader;
-import java.io.StringWriter;
-
-/**
- * @author liJiaRao
- * @date 2024-01-03 15:36
- */
-public class MathmlUtils {
-    /**
-     * Description: xsl转换器
-     */
-    public static String xslConvert(String s, String xslpath, URIResolver uriResolver) {
-        TransformerFactory tFac = TransformerFactory.newInstance();
-        if (uriResolver != null)
-            tFac.setURIResolver(uriResolver);
-        StreamSource xslSource = new StreamSource(MathmlUtils.class.getResourceAsStream(xslpath));
-        StringWriter writer = new StringWriter();
-        try {
-            Transformer t = tFac.newTransformer(xslSource);
-            Source source = new StreamSource(new StringReader(s));
-            Result result = new StreamResult(writer);
-            t.transform(source, result);
-        } catch (TransformerException e) {
-            e.printStackTrace();
-        }
-        return writer.getBuffer().toString();
-    }
-
-    /**
-     * Description: 将mathml转为latx
-     * @param mml
-     * @return
-     */
-    public static String convertMML2Latex(String mml) {
-        mml = mml.substring(mml.indexOf("?>") + 2, mml.length()); //去掉xml的头节点
-        URIResolver r = new URIResolver() { //设置xls依赖文件的路径
-            @Override
-            public Source resolve(String href, String base) throws TransformerException {
-                InputStream inputStream = MathmlUtils.class.getResourceAsStream("/conventer/mml2tex/" + href);
-                return new StreamSource(inputStream);
-            }
-        };
-        String latex = xslConvert(mml, "/conventer/mml2tex/mmltex.xsl", r);
-        if (latex != null && latex.length() > 1) {
-            latex = latex.substring(1, latex.length() - 1);
-        }
-        return latex;
-    }
-
-    /**
-     * Description: office mathml转为mml
-     * @param xml
-     * @return
-     */
-    public static String convertOMML2MML(String xml) {
-        xml = "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"yes\"?>\n" +
-                "<m:oMath xmlns:m=\"http://schemas.openxmlformats.org/officeDocument/2006/math\">\n" +
-                "    <m:rad>\n" +
-                "        <m:e>\n" +
-                "            <m:sup>\n" +
-                "                <m:e>\n" +
-                "                    <m:r>\n" +
-                "                        <m:t>2</m:t>\n" +
-                "                    </m:r>\n" +
-                "                </m:e>\n" +
-                "                <m:sup>\n" +
-                "                    <m:e>\n" +
-                "                        <m:r>\n" +
-                "                            <m:t>3</m:t>\n" +
-                "                        </m:r>\n" +
-                "                    </m:e>\n" +
-                "                    <m:sup>\n" +
-                "                        <m:e>\n" +
-                "                            <m:r>\n" +
-                "                                <m:t>4</m:t>\n" +
-                "                            </m:r>\n" +
-                "                        </m:e>\n" +
-                "                        <m:sup>\n" +
-                "                            <m:e>\n" +
-                "                                <m:r>\n" +
-                "                                    <m:t>5</m:t>\n" +
-                "                                </m:r>\n" +
-                "                            </m:e>\n" +
-                "                            <m:r>\n" +
-                "                                <m:t>6</m:t>\n" +
-                "                            </m:r>\n" +
-                "                        </m:sup>\n" +
-                "                    </m:sup>\n" +
-                "                </m:sup>\n" +
-                "            </m:sup>\n" +
-                "        </m:e>\n" +
-                "    </m:rad>\n" +
-                "    <m:r>\n" +
-                "        <m:t> + </m:t>\n" +
-                "    </m:r>\n" +
-                "    <m:rad>\n" +
-                "        <m:e>\n" +
-                "            <m:r>\n" +
-                "                <m:t>sin(x)</m:t>\n" +
-                "            </m:r>\n" +
-                "        </m:e>\n" +
-                "    </m:rad>\n" +
-                "</m:oMath>\n";
-        String result = xslConvert(xml, "/conventer/OMML2MML.XSL", null);
-        return result;
-    }
-}

laws-modules/src/main/java/com/jero/modules/split/util/ParagraphChildOrderManager.java

@@ -1,23 +1,14 @@
 package com.jero.modules.split.util;
 
-import cn.hutool.poi.word.WordUtil;
 import lombok.extern.slf4j.Slf4j;
 import org.apache.poi.xwpf.usermodel.XWPFParagraph;
 import org.apache.poi.xwpf.usermodel.XWPFRun;
 import org.apache.xmlbeans.XmlCursor;
-import org.dom4j.Document;
-import org.dom4j.DocumentException;
-import org.dom4j.Element;
-import org.dom4j.Node;
-import org.dom4j.io.SAXReader;
 import org.openxmlformats.schemas.officeDocument.x2006.math.CTOMath;
 import org.openxmlformats.schemas.officeDocument.x2006.math.CTOMathPara;
-import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTP;
+import org.w3c.dom.Node;
+import org.w3c.dom.NodeList;
 
-import javax.xml.transform.*;
-import javax.xml.transform.stream.StreamResult;
-import javax.xml.transform.stream.StreamSource;
-import java.io.*;
 import java.util.ArrayList;
 import java.util.LinkedList;
 import java.util.List;
@@ -60,6 +51,14 @@ public class ParagraphChildOrderManager {
                 xmlcursor.pop();
             }
         }
+        if (paragraph.getCTP().xmlText().contains("m:oMath")) {
+            for (XWPFRun run : paragraph.getRuns()) {
+                String xmlText = run.getCTR().xmlText();
+                if (xmlText.contains("m:oMath")) {
+                    System.out.println(xmlText);
+                }
+            }
+        }
     }
 
     public List<Object> getChildList() {
@@ -97,103 +96,21 @@ public class ParagraphChildOrderManager {
             return runsOrMathList;
         }
     }
-
-    public static String parseParagraph(XWPFParagraph xwpfParagraph) {
-        CTP ctp = xwpfParagraph.getCTP();
-        String xmlText = ctp.xmlText();
-        StringBuilder sb = new StringBuilder();
-//        if (xmlText.contains("<m:oMath>")) {
-
-        //段落文本内容
-        sb.append(xwpfParagraph.getParagraphText());
-        //段落公式解析
-        //得到根节点的值
-        SAXReader saxReader = new SAXReader();
-        //将String类型的字符串转换成XML文本对象
-        Document doc = null;
-        try {
-            doc = saxReader.read(new ByteArrayInputStream(xmlText.getBytes()));
-        } catch (DocumentException e) {
-            e.printStackTrace();
+    private Node getChildNode(Node node, String nodeName) {
+        if (!node.hasChildNodes()) {
+            return null;
         }
-        Element root = doc.getRootElement();
-        // 一个段落多个表达式解析
-        //用xpath得到OMML节点
-        List<Node> nodes = root.selectNodes("//m:oMath");
-        if (nodes != null && !nodes.isEmpty()) {
-            for (Node node : nodes) {
-                /**
-                 * OMML -> MathML -> LaTex
-                 * Office在安装目录中提供了将OMML转为MathML的xsl工具：MML2OMML.XSL
-                 * MathML转LaTex使用网上找到另一个xsl工具mmltex.xsl。
-                 */
-                String xml = node.asXML();
-                //xml转 mathml
-                String mathml = convertOMML2MML(xml);
-                //mathml转latx
-                String latex = convertMML2Latex(mathml);
-                sb.append(latex);
-                log.info("latex表达式：{}",latex);
+        NodeList childNodes = node.getChildNodes();
+        for (int i = 0; i < childNodes.getLength(); i++) {
+            Node childNode = childNodes.item(i);
+            if (nodeName.equals(childNode.getNodeName())) {
+                return childNode;
+            }
+            childNode = getChildNode(childNode, nodeName);
+            if (childNode != null) {
+                return childNode;
             }
         }
-        log.info("公式个数：{},解析内容：{}",nodes.size(),sb.toString());
-        return sb.toString();
-    }
-
-    /**
-     * <p>Description: 将mathml转为latx </p>
-     * @param mml mathml字符串
-     * @return
-     */
-    public static String convertMML2Latex(String mml){
-        mml = mml.substring(mml.indexOf("?>")+2, mml.length()); //去掉xml的头节点
-        URIResolver r = new URIResolver(){  //设置xls依赖文件的路径
-            @Override
-            public Source resolve(String href, String base) throws TransformerException {
-                File f = new File("/conventer/mml2tex/" + href);
-                InputStream inputStream = WordUtil.class.getResourceAsStream("/conventer/mml2tex/" + href);
-                return new StreamSource(inputStream);
-            }
-        };
-        String latex = xslConvert(mml, "/conventer/mml2tex/mmltex.xsl", r);
-        if(latex != null && latex.length() > 1){
-            latex = latex.substring(1, latex.length() - 1);
-        }
-        return latex;
-    }
-    /**
-     * <p>Description: office xml转为mathml </p>
-     * @param xml  公式xml
-     * @return
-     */
-    public static String convertOMML2MML(String xml){
-        // 进行转换的过程中需要借助这个文件,一般来说本机安装office就会有这个文件,找到就可以
-        String result = xslConvert(xml, "/conventer/OMML2MML.XSL", null);
-        return result;
-    }
-
-    /**
-     * Description: xsl转换器</p>
-     * @param s  公式xml字符串
-     * @param xslpath   转换器路径
-     * @param uriResolver xls依赖文件
-     * @return
-     */
-    public static String xslConvert(String s, String xslpath, URIResolver uriResolver){
-        TransformerFactory tFac = TransformerFactory.newInstance();
-        if(uriResolver != null) {
-            tFac.setURIResolver(uriResolver);
-        }
-        StreamSource xslSource = new StreamSource(WordUtil.class.getResourceAsStream(xslpath));
-        StringWriter writer = new StringWriter();
-        try {
-            Transformer t = tFac.newTransformer(xslSource);
-            Source source = new StreamSource(new StringReader(s));
-            Result result = new StreamResult(writer);
-            t.transform(source, result);
-        } catch (TransformerException e) {
-            log.error(e.getMessage(), e);
-        }
-        return writer.getBuffer().toString();
+        return null;
     }
 }