PDF Images lost at xdocreport 1.0.0 in felix osgi container

GoogleCodeExporter commented 9 years ago

What steps will reproduce the problem?

  Docx to PDF conversion of images loses image.

What is the expected output? What do you see instead?

  The PDF contains no image: the image should be in the PDF.

What version of the product are you using? On what operating system?

  xdocreport 1.0.0 in felix OSGi on Windows 7: some jars are re-wrapped.

Please provide any additional information below.

  When generating from docx to pdf, images are not being written to the generated pdf, but only when I am running in a felix osgi container. Running outside of this enviroment generates images successfully when using the same jars. So there appears to be a config issue with the bundle wrappers. Some of the bundles I am using are wrapped by you, some by me.

I was previously running xdocreport 0.9.8 as and wrapped several bundles in 
order to get up and running in osgi, as per: 
http://code.google.com/p/xdocreport/issues/detail?id=159. Images were being 
successfully generated.

Migrating to xdocreport 1.0.0 involved:

    . modifying:

    - my existing poi wrapper which wraps:
        org.apache.poi:poi:3.8
        org.apache.poi:poi-ooxml:3.8
        org.apache.poi:ooxml-schemas:1.1
      in order to export org.openxmlformats.schemas.drawingml.x2006.wordprocessingDrawing.

    . wrapping (to modify imports/exports to coexist with my config):

    - fr.opensagres.xdocreport.converter-1.0.0.jar
    - fr.opensagres.xdocreport.converter.docx.xwpf-1.0.0.jar (I also wrapped this at 0.9.8)

    . using unmodified:

    - fr.opensagres.xdocreport.core-1.0.0.jar
    - fr.opensagres.xdocreport.document-1.0.0.jar
    - fr.opensagres.xdocreport.document.docx-1.0.0.jar
    - fr.opensagres.xdocreport.itext.extension-1.0.0.jar
    - fr.opensagres.xdocreport.template-1.0.0.jar
    - fr.opensagres.xdocreport.template.freemarker-1.0.0.jar
    - org.apache.poi.xwpf.converter.core-1.0.0.jar
    - org.apache.poi.xwpf.converter.pdf-1.0.0.jar
    - org.apache.poi.xwpf.converter.xhtml-1.0.0.jar

The problem is related to xmlbeans representing the image as XmlAnyTypeImpl 
object when it should be a CTPicture.

The line it fails on is 
org.apache.poi.xwpf.converter.core.XWPFDocumentVisitor.visitGraphicalObject(..) 
line 1223 (org.apache.poi.xwpf.converer.core-1.0.0.jar). The object is an 
XmlAnyTypeImpl object rather than a CTPicture object so it skips some 
significant code. I cannot debug the getObject() call.

    XmlObject o = c.getObject();                    
    if ( o instanceof CTPicture ) { // test fails as o is XmlAnyType
        CTPicture picture = (CTPicture) o;                        
        // extract the picture if needed                        
        IImageExtractor extractor = getImageExtractor();

Note that prior to this in 
org.apache.poi.xwpf.usermodel.XWPFRun.getCTPictures(...) line 150 it caters for 
an XmlAnyTypeImpl object and creates a CTPicture from it.

    if(pict instanceof XmlAnyTypeImpl) {               
      // Pesky XmlBeans bug - see Bugzilla #49934                
      try {                    
        pict = CTPicture.Factory.parse(pict.toString());                
      } catch(XmlException e) {                    
        throw new POIXMLException(e);                
      }
    }

Applying the same check and fix in XWPFDocumentVisitor.visitGraphicalObject(..) 
resolves my problem. However, given that this only occurs in the osgi 
environment (I get a CTPicture 

"org.openxmlformats.schemas.drawingml.x2006.picture.impl.CTPictureImpl" at both 
points when running outside osgi) I believe this is not the correct fix - there 
must be some additional config required. There is nothing in the log files 
which supplies any clues regarding what to do.

I noticed that the pom for your org.apache.poi.xwpf.converter.core-1.0.0.jar 
contains an exclusion which hints at some problems in this area:

        <dependency>
            <groupId>org.apache.poi</groupId>
            <artifactId>poi-ooxml</artifactId>
            <version>3.8</version>
            <exclusions>
                <!-- surprisingly this artifect contains a subset of org.apache.poi:ooxml-schemas:1.1 -->
                <exclusion>
                    <artifactId>poi-ooxml-schemas</artifactId>
                    <groupId>org.apache.poi</groupId>
                </exclusion>
            </exclusions>
        </dependency>

And there is also the issue referenced in the commment in the XWPFRun code 
which again hints at similar problems: 
https://issues.apache.org/bugzilla/show_bug.cgi?id=49934.

I have tinkered with the import/exports in the bundles that I have wrapped to 
try expose the packages which contain the CTPictures related class files but 
without success. I'm basically working by guesswork.

Can you provide any suggestions regarding what is occurring here and how to 
resolve it?

Original issue reported on code.google.com by Mr.M.McM...@googlemail.com on 21 Mar 2013 at 9:10

GoogleCodeExporter commented 9 years ago

Hi, 
At first, why you have wrapping XDocReport bundles :
- fr.opensagres.xdocreport.converter-1.0.0.jar
- fr.opensagres.xdocreport.converter.docx.xwpf-1.0.0.jar (I also wrapped this 
at 0.9.8)

If there are a problem with thoses JARs in OSGi context, it should be fixed.

@Pascal: could you see this issue please?

Many thank's

Original comment by angelo.z...@gmail.com on 21 Mar 2013 at 9:28

Changed state: Accepted

GoogleCodeExporter commented 9 years ago

At 0.9.8 I followed the advice at 
http://code.google.com/p/xdocreport/issues/detail?id=159 to run in OSGi. 
Specifically the wrapper is needed, as reported in 159, to avoid:

    java.lang.NoClassDefFoundError: fr/opensagres/xdocreport/core/utils/StringUtils

At 1.0.0 for similar reasons I again wrapped the jar, ie in order to import 
fr.opensagres.xdocreport.core.utils which is missing from your version of the 
wrapper.

Note that this is only required when executing the conversion via:

    IXDocReport report = ...
    report.convert(context, Options.getTo(ConverterTypeTo.PDF).via(ConverterTypeVia.XWPF), os);

but it is not required when executing the conversion via:

    XWPFDocument document = ...
    PdfConverter.getInstance().convert(document, os, PdfOptions.create())

The other wrapper was modified in error by me. In accounting for the addition 
of the import of org.apache.poi.xwpf.converter.core to 
fr.opensagres.xdocreport.converter.docx.xwpf-1.0.0, I wrapped 
fr.opensagres.xdocreport.converter-1.0.0 and imported it there. So there was no 
need to wrap fr.opensagres.xdocreport.converter-1.0.0.

I have now rectified this situation so that I only now wrap 
fr.opensagres.xdocreport.converter.docx.xwpf-1.0.0.jar to reolve the 
StringUtils problem. However, the same lost image issue occurs when converting 
a docx to a PDF: ie  XmlAnyTypeImpl object rather than a CTPicture object so 
image is lost.

I also followed the wrapping advice from 
http://code.google.com/p/xdocreport/issues/detail?id=159 for POI (using the 
that issue's attached pom.xml). When reusing this with your 
org.apache.poi.xwpf.converter.core 1.0.0 there was a problem starting that 
bundle:

    org.osgi.framework.BundleException: Unresolved constraint in bundle org.apache.poi.xwpf.converter.core [89]: package;
    (package=org.openxmlformats.schemas.drawingml.x2006.wordprocessingDrawing)

hence, my exposing 
org.openxmlformats.schemas.drawingml.x2006.wordprocessingDrawing in my POI 
wrapper.

My suspicion is that more of these org.openxmlformats.schemas are required. I 
tried exporting the CTPictures related ones but it made no difference. Should 
the org.apache.poi.xwpf.converter.core wrapper perhaps import some of these 
other schema packages?

Can you provide any suggestions regarding what is occurring here and how to 
resolve it?

Original comment by Mr.M.McM...@googlemail.com on 21 Mar 2013 at 4:18

GoogleCodeExporter commented 9 years ago

Hi there.

Do you have any thoughts about what is happening here or how to resolve this 
issue?

Thanks.

Original comment by Mr.M.McM...@googlemail.com on 28 Mar 2013 at 9:01

GoogleCodeExporter commented 9 years ago

Hi,

on the dev branch i add the missing package on the converter fragments 
(fr.opensagres.xdocreport.core.utils).
If you want to test, our jars are deployed to 
https://oss.sonatype.org/content/repositories/snapshots/.

Not sure I get your point with POI packages but I suspect Bnd may miss 
Import-package on POI.
I suggest to add
DynamicImport-Package: org.apache.poi.* on xdocreport XMPF converter.

what do u think ?

Original comment by pascal.leclercq on 30 Mar 2013 at 10:39

GoogleCodeExporter commented 9 years ago

I tried the DynamicImport-Package as suggested but the image was still lost.

Do you understand the behaviour? Is the image objcet created as an XmlAnyType 
rather than a CTPicture because some package is incorrectly exported in the POI 
wrapper (from http://code.google.com/p/xdocreport/issues/detail?id=159) or 
because it is incorrectly imported by xdocreport 1.0.0? Why did it work with 
xdocreport 0.9.8?

Original comment by Mr.M.McM...@googlemail.com on 3 Apr 2013 at 9:13

GoogleCodeExporter commented 9 years ago

I have doen a big refactoring in XDocReport 1.0.0 for docx->pdf converter for 
images to improve it (manage aboslute image, etc)

In 0.9.8, pictures are visited like this : 

---------------------------------------------------------------------
protected void visitPictures( XWPFRun run, T parentContainer )
        throws Exception
    {
        List<XWPFPicture> embeddedPictures = run.getEmbeddedPictures();
        for ( XWPFPicture picture : embeddedPictures )
        {
            visitPicture( picture, parentContainer );
        }
    }
---------------------------------------------------------------------

See 
https://code.google.com/p/xdocreport/source/browse/thirdparties-extension/org.ap
ache.poi.xwpf.converter/src/main/java/org/apache/poi/xwpf/converter/internal/XWP
FElementVisitor.java?name=xdocreport-0.9.8

In 1.0.0, it is visited like this : 

------------------------------------------------------------------
private void visitGraphicalObject( T parentContainer, CTGraphicalObject 
graphic, Float offsetX,
                                       STRelFromH.Enum relativeFromH, Float offsetY, STRelFromV.Enum relativeFromV )
        throws Exception
    {
        if ( graphic != null )
        {
            CTGraphicalObjectData graphicData = graphic.getGraphicData();
            if ( graphicData != null )
            {
                XmlCursor c = graphicData.newCursor();
                c.selectPath( "./*" );
                while ( c.toNextSelection() )
                {
                    XmlObject o = c.getObject();
                    if ( o instanceof CTPicture )
                    {
                        CTPicture picture = (CTPicture) o;
                        // extract the picture if needed
                        IImageExtractor extractor = getImageExtractor();
                        if ( extractor != null )
                        {
                            XWPFPictureData pictureData = getPictureData( picture );
                            if ( pictureData != null )
                            {
                                try
                                {
                                    extractor.extract( WORD_MEDIA + pictureData.getFileName(), pictureData.getData() );
                                }
                                catch ( Throwable e )
                                {
                                    LOGGER.log( Level.SEVERE,
                                                "Error while extracting the image " + pictureData.getFileName(), e );
                                }
                            }
                        }
                        // visit the picture.
                        visitPicture( picture, offsetX, relativeFromH, offsetY, relativeFromV, parentContainer );
                    }
                }
                c.dispose();
            }
        }
    }
-----------------------------------------

See 
https://code.google.com/p/xdocreport/source/browse/thirdparties-extension/org.ap
ache.poi.xwpf.converter.core/src/main/java/org/apache/poi/xwpf/converter/core/XW
PFDocumentVisitor.java?name=xdocreport-1.0.0

If I remember I do that to keep the oroder of image, text, image text etc.

Perhaps we should apply your patch : 

-------------------------------------
if(pict instanceof XmlAnyTypeImpl) {               
      // Pesky XmlBeans bug - see Bugzilla #49934                
      try {                    
        pict = CTPicture.Factory.parse(pict.toString());                
      } catch(XmlException e) {                    
        throw new POIXMLException(e);                
      }
    }
-------------------------------------

I don't know?

Original comment by angelo.z...@gmail.com on 3 Apr 2013 at 9:51

GoogleCodeExporter commented 9 years ago

I'm trying to convert docx to pdf, I'm not seeing the images in header section 
converted in pdf. Can you please help? I'm using xdocreport 1.0.0 version

Original comment by gha...@gmail.com on 23 May 2013 at 3:35

GoogleCodeExporter commented 9 years ago

At first last version is 1.0.2. Please test your docx with this version.

Are you in OSGi env? If no, your problem is not linked to this issue, so please 
create a new issue if 1.0.2 doesn't resolve your problem by attaching your docx.

Original comment by angelo.z...@gmail.com on 23 May 2013 at 7:48

GoogleCodeExporter commented 9 years ago

org.apache.poi.xwpf.converter.core.XWPFConverterException: 
java.lang.NullPointerException
        at org.apache.poi.xwpf.converter.pdf.PdfConverter.doConvert(PdfConverter.java:59)
        at org.apache.poi.xwpf.converter.pdf.PdfConverter.doConvert(PdfConverter.java:37)
        at org.apache.poi.xwpf.converter.core.AbstractXWPFConverter.convert(AbstractXWPFConverter.java:45)
        at com.medikm.util.PdfConversion.convertDocToPDF(PdfConversion.java:131)
        at com.medikm.util.PdfConversion.convertFileToPDF(PdfConversion.java:405)
        at com.medikm.servlet.ResourceUploadServlet.doPost(ResourceUploadServlet.java:134)
        at javax.servlet.http.HttpServlet.service(HttpServlet.java:641)
        at javax.servlet.http.HttpServlet.service(HttpServlet.java:722)
        at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(ApplicationFilterChain.java:304)
        at org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationFilterChain.java:210)
        at com.medikm.filter.UserAuthFilter.doFilter(UserAuthFilter.java:43)
        at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(ApplicationFilterChain.java:243)
        at org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationFilterChain.java:210)
        at com.medikm.filter.DispatcheFilter.doFilter(DispatcheFilter.java:133)
        at org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(ApplicationFilterChain.java:243)
        at org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationFilterChain.java:210)
        at org.apache.catalina.core.StandardWrapperValve.invoke(StandardWrapperValve.java:224)
        at org.apache.catalina.core.StandardContextValve.invoke(StandardContextValve.java:169)
        at org.apache.catalina.authenticator.AuthenticatorBase.invoke(AuthenticatorBase.java:472)
        at org.apache.catalina.core.StandardHostValve.invoke(StandardHostValve.java:168)
        at org.apache.catalina.valves.ErrorReportValve.invoke(ErrorReportValve.java:100)
        at org.apache.catalina.valves.AccessLogValve.invoke(AccessLogValve.java:929)
        at org.apache.catalina.core.StandardEngineValve.invoke(StandardEngineValve.java:118)
        at org.apache.catalina.connector.CoyoteAdapter.service(CoyoteAdapter.java:405)
        at org.apache.coyote.http11.AbstractHttp11Processor.process(AbstractHttp11Processor.java:964)
        at org.apache.coyote.AbstractProtocol$AbstractConnectionHandler.process(AbstractProtocol.java:515)
        at org.apache.tomcat.util.net.JIoEndpoint$SocketProcessor.run(JIoEndpoint.java:302)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
        at java.lang.Thread.run(Thread.java:724)
Caused by: java.lang.NullPointerException
        at org.apache.poi.xwpf.converter.core.utils.XWPFTableUtil.getWidth(XWPFTableUtil.java:273)
        at org.apache.poi.xwpf.converter.core.utils.XWPFTableUtil.computeColWidths(XWPFTableUtil.java:284)
        at org.apache.poi.xwpf.converter.core.utils.XWPFTableUtil.computeColWidths(XWPFTableUtil.java:127)
        at org.apache.poi.xwpf.converter.core.XWPFDocumentVisitor.visitTable(XWPFDocumentVisitor.java:828)
        at org.apache.poi.xwpf.converter.core.XWPFDocumentVisitor.visitBodyElements(XWPFDocumentVisitor.java:246)
        at org.apache.poi.xwpf.converter.core.XWPFDocumentVisitor.start(XWPFDocumentVisitor.java:194)
        at org.apache.poi.xwpf.converter.pdf.PdfConverter.doConvert(PdfConverter.java:55)
        ... 29 more

how to solve this problem

Original comment by rohit.mc...@gmail.com on 24 Feb 2014 at 12:00

karurkarthi / xdocreport

PDF Images lost at xdocreport 1.0.0 in felix osgi container #238