Plutext

Posted: **Thu Aug 11, 2022 7:41 pm**

Hi,
Nice work you have done with Docx4j!

I need to convert html with MathML to docx but during conversion MathML tags are striped.

How could I make this work?
Hope you can help me.

Thanks
Milqn

Posted: **Fri Aug 12, 2022 7:50 pm**

There is XSLT for converting OMML to MathML, but maybe not for the converse.

Google reveals python and npm libraries called mathml2omml (and there may be others); if one of these work well enough on your MathML, you could integrate them into docx4j-ImportXHTML.

Posted: **Fri Aug 12, 2022 10:06 pm**

Thank you for your answer!

I also thought that my issue could arise due to the Ooml.

MathML tags are not preserved and hence, LibreOffice also can't display math.

Posted: **Tue Aug 16, 2022 7:37 am**

Please don't mind but I need to ask what would be the best way to include MML2OMML?
I have find MML2OMML.XSL stylesheet.

How would you do it?

Thanks a lot

Posted: **Fri Aug 19, 2022 4:41 am**

Any assistance would be gratefully appreciated.

jason wrote:There is XSLT for converting OMML to MathML, but maybe not for the converse.

Google reveals python and npm libraries called mathml2omml (and there may be others); if one of these work well enough on your MathML, you could integrate them into docx4j-ImportXHTML.

Posted: **Tue Aug 23, 2022 9:57 pm**

Here's a sketch of how to do it to help you get started, mostly untested, but based on importing the following sample XHTML:

Code: Select all: <?xml version="1.0" encoding="UTF-8"?> <html xmlns="http://www.w3.org/1999/xhtml"> <head> <title>MathML in XHTML</title> </head> <body> <p> Follows: <math xmlns="http://www.w3.org/1998/Math/MathML"> <mfrac> <mn>1</mn> <msqrt> <mn>2</mn> </msqrt> </mfrac> </math> </p> </body> </html>

In XHTMLImporterImpl, at approx line 1383, add and flesh out the following block of code:

Syntax: [ Download ] [ Hide ]

Using java Syntax Highlighting

}elseif(e.getNodeName().equals("math")){

// handle me
System.out.println("TODO: Handle mathml \n\r"+ XmlUtils.w3CDomNodeToString(e));

// Prepare to transform Element e

                        Templates xslt =null;// your mathml2omml.xslt 

// Use constructor which takes Unmarshaller, rather than JAXBContext,
// so we can set JaxbValidationEventHandler

                        JAXBContext jc =Context.jc;

                        Unmarshaller u;
try{

                                        u = jc.createUnmarshaller();

                                u.setEventHandler(new org.docx4j.jaxb.JaxbValidationEventHandler());

                                jakarta.xml.bind.util.JAXBResult result =new jakarta.xml.bind.util.JAXBResult(u );

                                XmlUtils.transform(new DOMSource(e), xslt, null, result);

// What happened?
Object o = result.getResult();

// Attach it to the document
this.contentContextStack.peek().getContent().add();// or addAll

}catch(JAXBException e1){

                                        e1.printStackTrace();
}

return;

Parsed in 0.017 seconds,  using GeSHi 1.0.8.4

Posted: **Wed Aug 24, 2022 10:41 pm**

The following proof of concept code works on my sample XHTML:

Syntax: [ Download ] [ Hide ]

Using java Syntax Highlighting

}elseif(e.getNodeName().equals("math")){

// handle me
System.out.println("Handling mathml \n\r"+ XmlUtils.w3CDomNodeToString(e));

try{
// Prepare to transform Element e

                                Source xsltSource  =new StreamSource(

                                                ResourceUtils.getResource(
"mml2omml.xsl")
);// https://raw.githubusercontent.com/Marti ... l2omml.xsl
/* You need to add the template:

                                 * 

                                                  <xsl:template match="/|*">

                                                    <oMath>

                                                      <xsl:apply-templates mode="mml"  />

                                                    </oMath>

                                                  </xsl:template>

                                 */

                                Templates xslt  = XmlUtils.getTransformerTemplate(xsltSource);

// Use constructor which takes Unmarshaller, rather than JAXBContext,
// so we can set JaxbValidationEventHandler

                                JAXBContext jc =Context.jc;

                                Unmarshaller u = jc.createUnmarshaller();

                                u.setEventHandler(new org.docx4j.jaxb.JaxbValidationEventHandler());

                                jakarta.xml.bind.util.JAXBResult result =new jakarta.xml.bind.util.JAXBResult(u );

                                XmlUtils.transform(new DOMSource(e), xslt, null, result);

// What happened?

                                org.docx4j.math.CTOMath math =(org.docx4j.math.CTOMath)XmlUtils.unwrap(result.getResult());

                                org.docx4j.math.ObjectFactory mathObjectFactory =new org.docx4j.math.ObjectFactory();
// Create object for oMathPara (wrapped in JAXBElement) 

                        CTOMathPara omathpara = mathObjectFactory.createCTOMathPara();

                        JAXBElement<org.docx4j.math.CTOMathPara> omathparaWrapped = mathObjectFactory.createOMathPara(omathpara);

                        omathpara.getOMath().add(math);

                        P wP =new P();

                        wP.getContent().add(omathparaWrapped);

// Attach it to the document
this.contentContextStack.peek().getContent().add(wP);

}catch(Exception e1){
thrownew Docx4JException("Error processing MathML", e1);
}

return;

Parsed in 0.016 seconds,  using GeSHi 1.0.8.4

Note that the heavy lifting is done by https://raw.githubusercontent.com/Marti ... l2omml.xsl

Posted: **Thu Aug 25, 2022 9:03 am**

That's what I'm looking for. Great work.
Thanks a lot, Jason.

Posted: **Fri Feb 07, 2025 3:20 am**

Hello,

The issue is now fixed with the following:
- use later version of docx4j-importxhtml (I only needed the core library so I only referenced this one in my project) :

Code: Select all: <dependency> <groupId>org.docx4j</groupId> <artifactId>docx4j-ImportXHTML-core</artifactId> <version>11.5.0</version> </dependency>

Note: You can visit the docx4j-importxhtml GitHub repository https://github.com/plutext/docx4j-ImportXHTML/tree/VERSION_11_4_8/docx4j-ImportXHTML-core and check the XHTMLImporterImpl file on the version that you are interested in. Make sure you find the right code if you search by "math" (including the quotes).
- use the correct mml2omml.xlsZ file (make sure this is the right name and extension, as otherwise you will get an error while trying to decode, that the file wasn't found on classpath). The correct one was shared earlier by jason :

Note that the heavy lifting is done by https://raw.githubusercontent.com/Marti ... l2omml.xsl

In which you must add :

Code: Select all: <xsl:template match="/|*"> <oMath> <xsl:apply-templates mode="mml" /> </oMath> </xsl:template>

I added it right after

Code: Select all: <xsl:variable name="StrLCAlphabet">abcdefghijklmnopqrstuvwxyz</xsl:variable>

Cheers,
A.

Posted: **Thu Sep 03, 2026 11:16 am**

An update: as of docx4j-ImportXHTML 17.0.4, MathML in your XHTML converts to native Word equations out of the box — no mml2omml.xsl, no XSLT template to add, no stylesheet renaming.

The conversion is now a native Java converter (MathML → OMML), so a <math> element in the input becomes a real m:oMath in the output docx — an actual Word equation, editable in Word's equation editor, not an image or stripped content. The mml2omml.xsl route this thread describes (and 11.5.0 required you to wire up manually) has become the opt-in legacy path; by default nothing needs configuring.

If you're also going the other direction (docx → HTML), the matching converter emits native MathML there too, since docx4j 17.0.4.

Details: CR-007 (OMML ⇄ MathML, both directions).

Plutext

Issue with MathML

Issue with MathML

Re: Issue with MathML

Re: Issue with MathML

Re: Issue with MathML

Re: Issue with MathML

Re: Issue with MathML

Re: Issue with MathML

Re: Issue with MathML

Re: Issue with MathML

Re: Issue with MathML