problem with document created by Google Docs

[sfmorais]

Hi,

When I run this code:

Code: Select all

WordprocessingMLPackage wordprocessingMLPackage = WordprocessingMLPackage.load(new File("c:\\document.docx"));
System.out.println(XmlUtils.marshaltoString(wordprocessingMLPackage.getMainDocumentPart().getJaxbElement(), true, true));


throw an exception

Code: Select all

INFO [org.docx4j.utils.XPathFactoryUtil:22] - xpath implementation: org.apache.xpath.jaxp.XPathFactoryImpl
INFO [org.docx4j.openpackaging.io3.Load3:180] - package read;  elapsed time: 4635 ms
INFO [org.docx4j.openpackaging.parts.JaxbXmlPart:129] - Lazily unmarshalling /word/document.xml
INFO [org.docx4j.openpackaging.parts.JaxbXmlPartXPathAware:299] - For org.docx4j.openpackaging.parts.WordprocessingML.MainDocumentPart, unmarshall via binder
java.lang.NumberFormatException: For input string: "10206.0"
   at java.lang.NumberFormatException.forInputString(NumberFormatException.java:48)
   at java.lang.Integer.parseInt(Integer.java:458)
   at java.math.BigInteger.<init>(BigInteger.java:316)
   at java.math.BigInteger.<init>(BigInteger.java:451)
   at com.sun.xml.bind.DatatypeConverterImpl._parseInteger(DatatypeConverterImpl.java:88)


I view the document.xml and really somewhere there is

Code: Select all

... <w:tblW w:w="10206.0" w:type="dxa"/> ...


I can conclude that the cause of error is the 10206.0 value of <w:tblW w:w and the docx4j expect an integer instead of a bigdecimal

I tested with docx4j-3.0.1 bu problem is the same

Note: the document.docx was created in the Google Docs and downloaded as docx

Attached is the document.docx


Many Thanks

Regards
Sérgio

[jason]

A quick look suggests Google Docs should be writing an integer.

Syntax: [ Download ] [ Hide ]

Syntax: [ Download ] [ Show ]

Using xml Syntax Highlighting

        <xsd:simpleType name="ST_DecimalNumber">
                <xsd:annotation>
                        <xsd:documentation>Decimal Number Value</xsd:documentation>
                </xsd:annotation>
                <xsd:restriction base="xsd:integer"/>
        </xsd:simpleType>
 

Parsed in 0.001 seconds, using GeSHi 1.0.8.4

See http://webapp.docx4java.org/OnlineDemo/ ... blW_2.html for @w:

The possible values for this attribute are defined by the ST_DecimalNumber simple type

http://webapp.docx4java.org/OnlineDemo/ ... umber.html

This simple type specifies that its contents will contain a whole decimal number (positive or negative)

Perhaps you could raise it with Google at https://productforums.google.com/forum/ ... ories/docs

[sfmorais]

Thanks, Jason

If I understand, you said that directly the 'docx4j API' can do nothing to solve it. The problem is the conversion done by google docs that does not respect the xsd, isn´t it ?

As you suggest I sent a topic in google forum, but it is difficult (or can take long time) obtain a reply due the quantity of posts everyday.

How can i turn this around in your opinion ?

Thanks
Sérgio

[jason]

Does it later say "encountered unexpected content; pre-processing"?

It should .. If so, you could modify "org/docx4j/jaxb/mc-preprocessor.xslt" to replace the decimal with an integer.

If you are comfortable with XSLT, I'd be happy to accept an appropriate template rule as a contrib. If not, let me know...

[sfmorais]

Jason,

it does not say "encountered unexpected content; pre-processing".

The full stracktrace is

Code: Select all

...
INFO [org.docx4j.openpackaging.contenttype.ContentTypeManager:802] - Detected WordProcessingML package
INFO [org.docx4j.openpackaging.io3.Load3:161] - Instantiated package of type org.docx4j.openpackaging.packages.WordprocessingMLPackage
INFO [org.docx4j.utils.XPathFactoryUtil:22] - xpath implementation: org.apache.xpath.jaxp.XPathFactoryImpl
INFO [org.docx4j.openpackaging.io3.Load3:180] - package read;  elapsed time: 4504 ms
INFO [org.docx4j.openpackaging.parts.JaxbXmlPart:129] - Lazily unmarshalling /word/document.xml
INFO [org.docx4j.openpackaging.parts.JaxbXmlPartXPathAware:299] - For org.docx4j.openpackaging.parts.WordprocessingML.MainDocumentPart, unmarshall via binder
java.lang.NumberFormatException: For input string: "10206.0"
   at java.lang.NumberFormatException.forInputString(NumberFormatException.java:48)
   at java.lang.Integer.parseInt(Integer.java:458)
   at java.math.BigInteger.<init>(BigInteger.java:316)
   at java.math.BigInteger.<init>(BigInteger.java:451)
   at com.sun.xml.bind.DatatypeConverterImpl._parseInteger(DatatypeConverterImpl.java:88)
   at com.sun.xml.bind.v2.model.impl.RuntimeBuiltinLeafInfoImpl$22.parse(RuntimeBuiltinLeafInfoImpl.java:733)
   at com.sun.xml.bind.v2.model.impl.RuntimeBuiltinLeafInfoImpl$22.parse(RuntimeBuiltinLeafInfoImpl.java:736)
   at com.sun.xml.bind.v2.runtime.reflect.TransducedAccessor$CompositeTransducedAccessorImpl.parse(TransducedAccessor.java:241)
   at com.sun.xml.bind.v2.runtime.unmarshaller.StructureLoader.startElement(StructureLoader.java:202)
   at com.sun.xml.bind.v2.runtime.unmarshaller.UnmarshallingContext._startElement(UnmarshallingContext.java:449)
   at com.sun.xml.bind.v2.runtime.unmarshaller.UnmarshallingContext.startElement(UnmarshallingContext.java:427)
   at com.sun.xml.bind.v2.runtime.unmarshaller.InterningXmlVisitor.startElement(InterningXmlVisitor.java:71)
   at com.sun.xml.bind.v2.runtime.unmarshaller.SAXConnector.startElement(SAXConnector.java:137)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:240)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:277)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:246)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:277)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:246)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:277)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:246)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:277)
   at com.sun.xml.bind.unmarshaller.DOMScanner.visit(DOMScanner.java:246)
   at com.sun.xml.bind.unmarshaller.DOMScanner.scan(DOMScanner.java:123)
   at com.sun.xml.bind.unmarshaller.DOMScanner.scan(DOMScanner.java:106)
   at com.sun.xml.bind.unmarshaller.DOMScanner.scan(DOMScanner.java:99)
   at com.sun.xml.bind.v2.runtime.BinderImpl.associativeUnmarshal(BinderImpl.java:156)
   at com.sun.xml.bind.v2.runtime.BinderImpl.unmarshal(BinderImpl.java:127)
   at org.docx4j.openpackaging.parts.JaxbXmlPartXPathAware.unmarshal(JaxbXmlPartXPathAware.java:316)
   at org.docx4j.openpackaging.parts.JaxbXmlPart.getJaxbElement(JaxbXmlPart.java:130)
   at exp.siga.aaa.docx4j.testes.Main.main(Main.java:36)


So, I can conclude that modify the mc-preprocessor.xslt cannot solve my problem, it is not ?

Thanks

[jason]

Hi Sérgio,

Here is what happens in org.docx4j.openpackaging.parts.JaxbXmlPartXPathAware.unmarshal

Syntax: [ Download ] [ Hide ]

Syntax: [ Download ] [ Show ]

Using java Syntax Highlighting

                        try {
                                jaxbElement =  (E) XmlUtils.unwrap(binder.unmarshal( doc ));
                        } catch (UnmarshalException ue) {
                                        // try the mc-preprocessor.xslt stuff
 

Parsed in 0.015 seconds, using GeSHi 1.0.8.4

In this case the exception is NumberFormatException (as opposed to UnmarshalException) - interesting.

We can catch that and do the same sort of thing. It may be expedient to do it in mc-preprocessor.xslt; if we don't, then the exception handled there could still occur (we'd need to experiment to see what order they happen in).

[sfmorais]

Thanks in advance Jason

I think that I must something like this:

- Create in my project a JaxbXmlPartXPathAware that overrides the jar and

Code: Select all

         try {
            jaxbElement =  (E) XmlUtils.unwrap(binder.unmarshal( doc ));
               // Unwrap, so we have eg CTEndnotes, not JAXBElement
         }
         catch (NumberFormatException numberFormatException) {
            
            DOMResult result = new DOMResult();
            Templates mcPreprocessorXslt = JaxbValidationEventHandler.getMcPreprocessor();
            
            XmlUtils.transform(doc, mcPreprocessorXslt, null, result);
            doc = (org.w3c.dom.Document) result.getNode();
            try {
               jaxbElement = (E) XmlUtils.unwrap(binder.unmarshal(doc));
            }
            catch (ClassCastException e) {
               Unmarshaller u = jc.createUnmarshaller();
               jaxbElement = (E) XmlUtils.unwrap(u.unmarshal( doc ));      
            }
         }
         catch (UnmarshalException ue) {
                           ... // no changes
                        }


Now, my difficult is understand the mc-preprocessor.xslt and adapt to my issue.
Am I the right way ?


My other way to around the problem is to get the content of 'document.xml' like string text (like XMLUtil.marshalToString() does, but without xsd validation). So, with regex pattern I could replace the values of w:w= to integer (trunc the decimal places). Is there any way to get the string content ?


Thanks

[jason]

Please see now https://github.com/plutext/docx4j/commi ... bd41be7bc3

I'll upload a nightly incorporating this later today.

I guess the question is: on what other elements/attributes does Google Docs make the same error? Please add to this thread if you find any others...

My other way to around the problem is to get the content of 'document.xml' like string text

So far we've managed to avoid any optional string manipulation step prior to unmarshalling; that may need to change in the future..

[jason]

Please try http://www.docx4java.org/docx4j/docx4j- ... 140225.jar

[sfmorais]

Many many thanks jason in advance

In a best analysis in the 'document.xml', I found (for my simple docx - in the first post) other tags with w:w= attribute of other different nodes like: <w:pgSz, <w:right, <w:bottom, <w:left, <w:top, <w:gridCol. For other docx more complex can exists more (I tell you when I find).

But, some w:w= appear in integer like <w:pgSz. (I don´t know what is the criteria of the Google to put some w:w= in integer other with decimal places).

The best is prevent the two possible values for all w:w (integer and decimal) and round or trunc if decimal

Your last jar solve me the first occurence of w:w (<w:tblW w:w="10206.0") but now appear the same exception for a "100.0"

Attached is the document.xml

Many Thanks

Regards,
Sergio

[jason]

OK please see now https://github.com/plutext/docx4j/commi ... e00d5ca257

By the way, could you please add a link to your post in the Google Docs forum, to make it easier to monitor for any replies?

[sfmorais]

Hi Jason

I did some tests and the problem is solved.
Really your solution in xstl is the best to control all w:w atributes for any node

The related link in Google Docs forum is
https://productforums.google.com/forum/#!msg/docs/fcFovKihMtw/Ii7_mf1bO9wJ

Many thanks for your solution and nice.

Regards,
Sérgio