Extracting XMP data and JAXB

Programming Questions

phillarby

in Programming Questions • 1 year ago

Hi all,

I'm currently trying to use JAXB to read XMP data extracted from image files. I'm basing this on the tutorial here: http://wiki.processing.org/w/XML_parsing_with_JAXB

Getting the XMP data from the file is fine. An example of the data I'm getting is:

<x:xmpmeta xmlns:x="adobe:ns:meta/"><rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:Description rdf:about="uuid:faf5bdd5-ba3d-11da-ad31-d33d75182f1b" xmlns:dc="http://purl.org/dc/elements/1.1/"><dc:subject><rdf:Bag xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:li>Hong Kong</rdf:li><rdf:li>Asia</rdf:li><rdf:li>Market</rdf:li><rdf:li>Night</rdf:li></rdf:Bag></dc:subject></rdf:Description><rdf:Description rdf:about="uuid:faf5bdd5-ba3d-11da-ad31-d33d75182f1b" xmlns:dc="http://purl.org/dc/elements/1.1/"><dc:creator><rdf:Seq xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:li>Phil Larby</rdf:li></rdf:Seq></dc:creator><dc:title><rdf:Alt xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:li xml:lang="x-default">Temple Street Night Market</rdf:li></rdf:Alt></dc:title><dc:description><rdf:Alt xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:li xml:lang="x-default">Temple Street Night Market</rdf:li></rdf:Alt></dc:description></rdf:Description><rdf:Description xmlns:MicrosoftPhoto="http://ns.microsoft.com/photo/1.0/"><MicrosoftPhoto:LastKeywordXMP><rdf:Bag xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:li>Hong Kong</rdf:li><rdf:li>Asia</rdf:li><rdf:li>Market</rdf:li><rdf:li>Night</rdf:li></rdf:Bag></MicrosoftPhoto:LastKeywordXMP></rdf:Description></rdf:RDF></x:xmpmeta>

However when I try to unmarshal this into objects it fails due to the namespace prefixes. The error I get is:

javax.xml.bind.UnmarshalException: unexpected element (uri:"adobe:ns:meta/", local:"xmpmeta"). Expected elements are <{}xmpmeta>

If I manipulate the XMP data as a string string to remove the namespace prefixes e.g. <x:xmpmeta /> become <xmpmeta /> then it works, but this is pretty unreliable as I keep encountering new namespaces as I try different files and basically this just doesn't seem to be the 'right' way to go about it.

Any pointers on how to handle the namespace prefixes in JAXB, as I'm failing to get to the bottom of it at the moment!

Cheers
Phil

Replies(5)

PhiLho

Re: Extracting XMP data and JAXB

1 year ago

Can you show the code you have currently to read this file?

phillarby

Re: Extracting XMP data and JAXB

1 year ago

Here are snips of some of the relevant bits. Can provide more, or even the full project, if it will help.

The method that inititates processing of the Xml - snipped to only show processing one data hierarchy:

private void extractXMPDetails()
{
    //Trying to parse xml datra using JAXB as per http://wiki.processing.org/w/XML_parsing_with_JAXB
    AppConfig config;

    try
    {
      // setup object mapper using the AppConfig class
      JAXBContext context = JAXBContext.newInstance(AppConfig.class);

      String sXML = removeXMPNamespaces(LoadXMPData());
       //Load XML string into stream for processing using JAXB
      ByteArrayInputStream input = new ByteArrayInputStream(sXML.getBytes());
      println(sXML);

      if (LoadXMPData().getBytes().length > 0)
      {
        // parse the XML and return an instance of the AppConfig class
        config = (AppConfig) context.createUnmarshaller().unmarshal(input);

        //Iterate through elements adding to detail
        for (XMPRDFElement r : config.rdf)
        {
          println("Found RDF TAG");

          for (XMPDescriptionElement d : r.description)
          {
            println("Found Description TAG");

            for (XMPTitleElement t : d.title)
            {
              println("Found Title TAG");

              for (XMPAltElement a : t.alt)
              {
                println("Found Alt TAG");

                for (XMPLiElement l : a.li)
                {
                  println("Found li TAG");

                  //Add detail to image
                  addDetail("Title", l.value);
                }
              }
            }
          }
        }
      }
    }
    catch(JAXBException e)
    {
      // if things went wrong...
      println("error parsing xml: ");
      e.printStackTrace();
      // force quit
      System.exit(1);
    }
}

The AppConfig.java file

//This class has beened added as a .java file to prevent processing from adding it as
//an inner class to the main applicaton when compiling. This is required as JAXB does
//not support non-static inner classes

//This class is based on the example found here: http://wiki.processing.org/w/XML_parsing_with_JAXB

//This class defines the structure of the XMP data, as defined in the XMP data model found here:
//http://wwwimages.adobe.com/www.adobe.com/content/dam/Adobe/en/devnet/xmp/pdfs/XMPSpecificationPart1.pdf

import java.util.*;
import javax.xml.bind.annotation.*;

// this annoation marks the AppConfig class to be able to act as
// an XML root element. The name parameter is only needed since
// our XML element name is different from the class name:
// <config> vs. AppConfig

@XmlRootElement(name="xmpmeta")
public class AppConfig {

// now we simply annotate the different variables
// depending if they are XML elements/nodes or node attributes
// the mapping to the actual data type is done automatically
@XmlAttribute(name="xmlns:x")
String x;

// here we also specify default values, which are used
// if there's no matching data for this variable in the XML
//@XmlElement
// String websites="Window title";

// one of the best things in JAXB is the ability to map entire
// class hierarchies and collections of data
// in this case each <url> element will be added to this list
// the actual MyURL class is defined in its own tab in the Processing PDE
@XmlElement(name="RDF")
List<XMPRDFElement> rdf = new ArrayList<XMPRDFElement>();
}

XMPDescription.java

//This class has beened added as a .java file to prevent processing from adding it as
//an inner class to the main applicaton when compiling. This is required as JAXB does
//not support non-static inner classes

//This class is based on the example found here: http://wiki.processing.org/w/XML_parsing_with_JAXB
import java.util.*;
import javax.xml.bind.annotation.*;

// this class is a simple pairing of 2 strings:
// a url and its description
public class XMPDescriptionElement {

@XmlElement(name="creator")
    List<XMPCreatorElement> creator = new ArrayList<XMPCreatorElement>();

@XmlElement(name="title")
    List<XMPTitleElement> title = new ArrayList<XMPTitleElement>();
}

XMPTitleElement.java

//This class has beened added as a .java file to prevent processing from adding it as
//an inner class to the main applicaton when compiling. This is required as JAXB does
//not support non-static inner classes

//This class is based on the example found here: http://wiki.processing.org/w/XML_parsing_with_JAXB
import java.util.*;
import javax.xml.bind.annotation.*;

// this class is a simple pairing of 2 strings:
// a url and its description
public class XMPTitleElement {

@XmlElement(name="Alt")
List<XMPAltElement> alt = new ArrayList<XMPAltElement>();
}

I won't post the full list of these classes as they're all very similar.

If I pass the XMP data through this method before loading it into JAXB then it works and I can naviate the object hierarchy. However if I don't include this step, or a new namespace is encountered (which happens fairly regularly when adding new image files) JAXB falls over with the previously posted error. I'm pretty uncomfortable with this as a solution as just doesn't feel like the way this should be done!

private String removeXMPNamespaces(String input)
{
   String xmp = input;
   xmp = xmp.replace("dc:", "");
   xmp = xmp.replace("x:", "");
   xmp = xmp.replace("rdf:", "");
   xmp = xmp.replace("xmlns:", "");
   xmp = xmp.replace("xml:", "");
   xmp = xmp.replace("xmp:", "");
   xmp = xmp.replace("tiff:", "");
   xmp = xmp.replace("MicrosoftPhoto:", "");
   return xmp;
}

Any advice much appreciated :)

Phil

PhiLho

Re: Extracting XMP data and JAXB

1 year ago

Mmm, I don't know enough about JAXB to help you, but a slightly smarter way to write your removeXMPNamespaces, working with all kinds of name spaces, could be:

private String removeXMPNamespaces(String input)
{
return input.replaceAll("<\\w+:", "");
}

Warning: untested, and a bit brutal.
The risk is to have two identical tags in different namespaces to be treated identically: that's the point of namespaces, to allow synonymic tags while keeping their meaning separate.

phillarby

Re: Extracting XMP data and JAXB

1 year ago

Thanks phi.lho. I think the code should be:

private String removeXMPNamespaces(String input)
{
return input.replaceAll("\\w+:", "");
}

as you wouldn't want to remove the leading < from the tag, but then this introduces more scope for problems!

Anyway, I'm sure I'll get the the bottom of it eventally - It's probably just a case of needing to put in a bit more effort to learn more about JAXB. I'll dig out the textbooks again!

I'll be sure to post an update if I get it working.

Cheers
Phil

PhiLho

Re: Extracting XMP data and JAXB

1 year ago

You are right, but let's make it: return input.replaceAll("<\\w+:", "<");
as it is more selective (you can find xxx: in text too).
And yes, JAXB probably offers some facilities to handle efficiently various namespaces.

Top Reply