使用 Apache POI 在 word (.docx) 中添加乳胶类型方程答案

【问题标题】：add latex type equation in word (.docx) using Apache POI使用 Apache POI 在 word (.docx) 中添加乳胶类型方程
【发布时间】：2017-10-07 18:16:17
【问题描述】：

我正在尝试使用 Apache POI 创建一个自动 (.docx) MS Word 文件。 Java 程序的输入包含文本、图像和 LaTeX 风格的方程式（嵌入在 $$ 或 [ ] 中）。

我的问题是如何在 Word 中添加这个 LaTeX 样式方程，以便在 MS Word 中编辑 .docx 文件时，它会将方程识别为 MS Word 样式方程（OMML 类型）

注意：我认为应该将 LaTeX 方程转换为 MathML。如果是这样，那么如何将 MathML 添加到 .docx 中？

【问题讨论】：

标签： java ms-word apache-poi

【解决方案1】：

Microsoft 提供 XSLT 样式表，用于将 OMML 转换为 MathML (OMML2MML.XSL) 以及使用 XSLT 将 MathML 转换为 OMML (MML2OMML.XSL)。

如果您安装了Microsoft Office，您将在Office 程序目录中找到这些文件。在我的系统中：

使用它，我们可以使用 XSLT 将 MathML 转换为 OMML。

例子：

import java.io.*;
import org.apache.poi.xwpf.usermodel.*;

import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTP;
import org.openxmlformats.schemas.officeDocument.x2006.math.CTOMath;
import org.openxmlformats.schemas.officeDocument.x2006.math.CTOMathPara;
import org.openxmlformats.schemas.officeDocument.x2006.math.CTR;

import javax.xml.transform.Transformer;
import javax.xml.transform.TransformerFactory;
import javax.xml.transform.stream.StreamSource;
import javax.xml.transform.stream.StreamResult;

import org.apache.xmlbeans.XmlCursor;

/*
needs the full ooxml-schemas-*.jar or poi-ooxml-full-5.0.0.jar as mentioned in https://poi.apache.org/faq.html#faq-N10025
*/

public class CreateWordFormulaFromMathML {

 static File stylesheet = new File("MML2OMML.XSL");
 static TransformerFactory tFactory = TransformerFactory.newInstance();
 static StreamSource stylesource = new StreamSource(stylesheet); 

 static CTOMath getOMML(String mathML) throws Exception {
  Transformer transformer = tFactory.newTransformer(stylesource);

  StringReader stringreader = new StringReader(mathML);
  StreamSource source = new StreamSource(stringreader);

  StringWriter stringwriter = new StringWriter();
  StreamResult result = new StreamResult(stringwriter);
  transformer.transform(source, result);

  String ooML = stringwriter.toString();
  stringwriter.close();

  CTOMathPara ctOMathPara = CTOMathPara.Factory.parse(ooML);
  CTOMath ctOMath = ctOMathPara.getOMathArray(0);

  //for making this to work with Office 2007 Word also, special font settings are necessary
  XmlCursor xmlcursor = ctOMath.newCursor();
  while (xmlcursor.hasNextToken()) {
   XmlCursor.TokenType tokentype = xmlcursor.toNextToken();
   if (tokentype.isStart()) {
    if (xmlcursor.getObject() instanceof CTR) {
     CTR cTR = (CTR)xmlcursor.getObject();
     cTR.addNewRPr2().addNewRFonts().setAscii("Cambria Math");
     cTR.getRPr2().getRFonts().setHAnsi("Cambria Math"); // up to apache poi 4.1.2
     //cTR.getRPr2().getRFontsArray(0).setHAnsi("Cambria Math"); // since apache poi 5.0.0
    }
   }
  }

  return ctOMath;
 }

 public static void main(String[] args) throws Exception {

  XWPFDocument document = new XWPFDocument();

  XWPFParagraph paragraph = document.createParagraph();
  XWPFRun run = paragraph.createRun();
  run.setText("The Pythagorean theorem: ");

  String mathML = 
    "<math xmlns=\"http://www.w3.org/1998/Math/MathML\">" 
   +"<mrow>"
   +"<msup><mi>a</mi><mn>2</mn></msup><mo>+</mo><msup><mi>b</mi><mn>2</mn></msup><mo>=</mo><msup><mi>c</mi><mn>2</mn></msup>"
   +"</mrow>"
   +"</math>";

  CTOMath ctOMath = getOMML(mathML);
System.out.println(ctOMath);

  CTP ctp = paragraph.getCTP();
  ctp.setOMathArray(new CTOMath[]{ctOMath});

  paragraph = document.createParagraph();
  run = paragraph.createRun();
  run.setText("The Quadratic Formula: ");

  mathML = 
    "<math xmlns=\"http://www.w3.org/1998/Math/MathML\">"
   +"<mrow>" 
   +"<mi>x</mi><mo>=</mo><mfrac><mrow><mrow><mo>-</mo><mi>b</mi></mrow><mo>±</mo><msqrt><mrow><msup><mi>b</mi><mn>2</mn></msup><mo>-</mo><mrow><mn>4</mn><mo>⁢</mo><mi>a</mi><mo>⁢</mo><mi>c</mi></mrow></mrow></msqrt></mrow><mrow><mn>2</mn><mo>⁢</mo><mi>a</mi></mrow></mfrac>"
   +"</mrow>"
   +"</math>";

  ctOMath = getOMML(mathML);
System.out.println(ctOMath);

  ctp = paragraph.getCTP();
  ctp.setOMathArray(new CTOMath[]{ctOMath});
  
  FileOutputStream out = new FileOutputStream("CreateWordFormulaFromMathML.docx");
  document.write(out);
  out.close();
  document.close();

 }
}

请注意，此代码需要完整的ooxml-schemas-*.jar 或poi-ooxml-full-5.0.0.jar，如https://poi.apache.org/faq.html#faq-N10025 中所述。

当然有 Java 库可用于将 LaTeX 转换为 MathML。例如：http://www.fmath.info/java/download.jsp。

已下载：fmath-mathml-java-test-project-b1124.zip 并在类路径中有 /lib/fmath-mathml-java.jar 和 /lib/jdom-2.0.6.jar，以下工作：

import java.io.*;
import org.apache.poi.xwpf.usermodel.*;

import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTP;
import org.openxmlformats.schemas.officeDocument.x2006.math.CTOMath;
import org.openxmlformats.schemas.officeDocument.x2006.math.CTOMathPara;
import org.openxmlformats.schemas.officeDocument.x2006.math.CTR;

import javax.xml.transform.Transformer;
import javax.xml.transform.TransformerFactory;
import javax.xml.transform.stream.StreamSource;
import javax.xml.transform.stream.StreamResult;

import org.apache.xmlbeans.XmlCursor;

/*
needs the full ooxml-schemas-1.3.jar as mentioned in https://poi.apache.org/faq.html#faq-N10025
*/

public class CreateWordFormulaFromLaTeX {

 static File stylesheet = new File("MML2OMML.XSL");
 static TransformerFactory tFactory = TransformerFactory.newInstance();
 static StreamSource stylesource = new StreamSource(stylesheet); 

 static CTOMath getOMML(String mathML) throws Exception {
  Transformer transformer = tFactory.newTransformer(stylesource);

  StringReader stringreader = new StringReader(mathML);
  StreamSource source = new StreamSource(stringreader);

  StringWriter stringwriter = new StringWriter();
  StreamResult result = new StreamResult(stringwriter);
  transformer.transform(source, result);

  String ooML = stringwriter.toString();
  stringwriter.close();

  CTOMathPara ctOMathPara = CTOMathPara.Factory.parse(ooML);
  CTOMath ctOMath = ctOMathPara.getOMathArray(0);

  //for making this to work with Office 2007 Word also, special font settings are necessary
  XmlCursor xmlcursor = ctOMath.newCursor();
  while (xmlcursor.hasNextToken()) {
   XmlCursor.TokenType tokentype = xmlcursor.toNextToken();
   if (tokentype.isStart()) {
    if (xmlcursor.getObject() instanceof CTR) {
     CTR cTR = (CTR)xmlcursor.getObject();
     cTR.addNewRPr2().addNewRFonts().setAscii("Cambria Math");
     cTR.getRPr2().getRFonts().setHAnsi("Cambria Math");
    }
   }
  }

  return ctOMath;
 }

 public static void main(String[] args) throws Exception {

  XWPFDocument document = new XWPFDocument();

  XWPFParagraph paragraph = document.createParagraph();
  XWPFRun run = paragraph.createRun();
  run.setText("The Pythagorean theorem: ");

  String latex = "$a^2 + b^2 = c^2$";

  String mathML = fmath.conversion.ConvertFromLatexToMathML.convertToMathML(latex);
  mathML = mathML.replaceFirst("<math ", "<math xmlns=\"http://www.w3.org/1998/Math/MathML\" ");
System.out.println(mathML);

  CTOMath ctOMath = getOMML(mathML);
System.out.println(ctOMath);

  CTP ctp = paragraph.getCTP();
  ctp.setOMathArray(new CTOMath[]{ctOMath});


  paragraph = document.createParagraph();
  run = paragraph.createRun();
  run.setText("The Quadratic Formula: ");

  latex = "$x=\\frac{-b\\pm\\sqrt{b^2-4ac}}{2a}$";

  mathML = fmath.conversion.ConvertFromLatexToMathML.convertToMathML(latex);
  mathML = mathML.replaceFirst("<math ", "<math xmlns=\"http://www.w3.org/1998/Math/MathML\" ");
  mathML = mathML.replaceAll("&plusmn;", "±");
System.out.println(mathML);

  ctOMath = getOMML(mathML);
System.out.println(ctOMath);

  ctp = paragraph.getCTP();
  ctp.setOMathArray(new CTOMath[]{ctOMath});

  document.write(new FileOutputStream("CreateWordFormulaFromLaTeX.docx"));
  document.close();

 }
}

但每次转换都包含可能的错误。所以 LaTeX -> MathML -> OMML 将比仅 MathML -> OMML 更容易出错。

在这种情况下，fmath.conversion.ConvertFromLatexToMathML.convertToMathML 会导致 Math XML 没有名称空间。但是由于XSLT需要这个，所以必须手动添加。

而fmath.conversion.ConvertFromLatexToMathML.convertToMathML 使用HTML 实体，MML2OMML.XSL 不知道。所以示例中的“±”必须替换为“±”。

也许SnuggleTeX 会是更好的库？

下载它并在类路径中有snuggletex-core-1.2.2.jar，我在上一个示例中的以下代码更改有效：

...
  String latex = "$a^2 + b^2 = c^2$";

  uk.ac.ed.ph.snuggletex.SnuggleEngine engine = new uk.ac.ed.ph.snuggletex.SnuggleEngine();
  uk.ac.ed.ph.snuggletex.SnuggleSession session = engine.createSession();
  uk.ac.ed.ph.snuggletex.SnuggleInput input = new uk.ac.ed.ph.snuggletex.SnuggleInput(latex);
  session.parseInput(input);
  String mathML = session.buildXMLString();
System.out.println(mathML);

/*
  String mathML = fmath.conversion.ConvertFromLatexToMathML.convertToMathML(latex);
  mathML = mathML.replaceFirst("<math ", "<math xmlns=\"http://www.w3.org/1998/Math/MathML\" ");
System.out.println(mathML);
*/

  CTOMath ctOMath = getOMML(mathML);
System.out.println(ctOMath);

...

  latex = "$x=\\frac{-b\\pm\\sqrt{b^2-4ac}}{2a}$";

  engine = new uk.ac.ed.ph.snuggletex.SnuggleEngine();
  session = engine.createSession();
  input = new uk.ac.ed.ph.snuggletex.SnuggleInput(latex);
  session.parseInput(input);
  mathML = session.buildXMLString();
System.out.println(mathML);

/*
  mathML = fmath.conversion.ConvertFromLatexToMathML.convertToMathML(latex);
  mathML = mathML.replaceFirst("<math ", "<math xmlns=\"http://www.w3.org/1998/Math/MathML\" ");
  mathML = mathML.replaceAll("&plusmn;", "±");
System.out.println(mathML);
*/

  ctOMath = getOMML(mathML);
System.out.println(ctOMath);
...

无需手动干预。至少不使用给定的 LaTeX 示例。

【讨论】：

谢谢！似乎工作得很好。有没有办法通过java将LaTeX风格的方程转换为MathML？
为了使它也可以与 Office 2007 Word 一起使用，需要特殊的字体设置，请参阅我的代码补充。对于您关于通过 java 将 LaTeX 样式方程转换为 MathML 的任何方法的问题：您对此的研究结果是什么？当然有任何方法，但每次转换都包含可能的错误。所以 LaTeX -> MathML -> OMML 将比仅 MathML -> OMML 更容易出错。所以自己做试验/错误。
@Axel Richter 以 Latex 风格提供输入要容易得多。另外，我只发现 MathJax 能够将 Latex 转换为 MathML。但它是一个 Web 应用程序，而不是 java。
@ashwin bande：当然有 Java 库可用于将 LaTeX 转换为 MathML。我目前发现SnuggleTeX 是最好的。看我的补充。但你必须自己测试。
SnuggleTex 是完美的解决方案。甚至比 fmath 还要好。再次感谢！