[jdom-interest] EntityResolver to resolve relative DTDs

[Pramodh Peddi]

Hi,

I know there are already a few requests and responses on this topic (and a
couple of them helped me in few instances), but I couldn't get exactly what
I wanted from them. I did not clearly undetsrtand the working of
EntiryResolver. I am using Java1.4's API for Transformation.

Following is what I am doing:
My source xml (which is to be transformed) has a relative URI specified for
DTD(<!DOCTYPE FEED_DTD SYSTEM "myDTD.dtd">). I want the parser to use an
absolute URL ("http://myserver.com/dtd/myDTD.dtd", and not the "myDTD.dtd"
specified inside the source xml) for the DTD to validate the source xml.

And this is the code I am using to do the above work:
**********************************CODE**************************************
******************
public class SELTransformationService{

private String stylesheet;
private String fileName = null;
private String dtdURL = null;

private void transform(String metadata, String stylesheet, ServiceRequest
req, ServiceResponse res){

//System.out.println("TRANSFORM USING STRING");

try {

TransformerFactory tFactory = TransformerFactory.newInstance();

Transformer transformer = tFactory.newTransformer(new StreamSource(new
URL(stylesheet).openStream()));

InputStream inputStream = req.getContentObject().getMetadataInputStream();

OutputStream outputStream =
req.getContentObject().getMetadataOutputStream();

SAXParserFactory pfactory= SAXParserFactory.newInstance();

pfactory.setValidating(true);

// Get an XMLReader.

XMLReader reader = pfactory.newSAXParser().getXMLReader();

//create a resolver to resolve the DTD in the source xml

//EntityResolver resolver = new DTDResolver();

reader.setEntityResolver(new DTDResolver());

DTDResolver resolver = (DTDResolver)reader.getEntityResolver();

resolver.setPublicId(this.dtdURL);

SAXSource source = new SAXSource(reader,

new InputSource(new InputStreamReader(inputStream)));

source.setSystemId("SonyDAMAssetMetadata.dtd");

transformer.transform(source, new StreamResult(new
OutputStreamWriter(outputStream, "iso-8859-1")));

outputStream.close();

req.getContentObject().commit();

res.send(req.getContentObject());

log.info(fileName + " - OBJECT SENT OUT OF TRANSFORMATION-2");

} catch (Exception ex) {

res.error(req.getContentObject(), "Exception sending message to bus.");

ex.printStackTrace();

}

}//end SELTransformationService classclass DTDResolver implements
EntityResolver {

String publicId = null;

public void setPublicId(String publicId){

this.publicId = publicId;

System.out.println("Setting publicID");

}

public String getPublicId(){

System.out.println("Getting publicId");

return this.publicId;

}

public InputSource resolveEntity (String publicId, String systemId){

InputStream inputStream = null;

InputSource source = null;

try{

System.out.println("publicID is: " + this.publicId);

if(StringUtils.isNotEmpty(this.publicId)){

URL url = new URL(this.publicId);

inputStream = url.openStream();

System.out.println("got the inputstream");

source = new InputSource(new InputStreamReader(inputStream));

}else{

System.out.println("publicId is not specified!!!");

}

}catch(Exception e){

}

return source;

}

}//end DTDResolver class

******************************************************CODE******************
*****************************

I am sending the publicID (which is the absolute URL i am asking the parser
to for the DTD) and using that for validation.

It is not throwing any exception, but the transformed stream doesn't have
anything except the header "<?xml version="1.0" encoding="UTF-8"?>". The
header is also not supposed to have UTF-8, it is supposed to be
"iso-8859-1". Can anyone tell me what my mistake is and what I am missing?

And few questions about the working of EntityResolver ( I went thru the API,
but its not completely clear to me. I was wondering if I can more clear
concepts about EntityResolver):

1. What is the significance of systemID and publicID?

2. What should the InputSource return symantically?



Sorry for the long email and a bunch of questions. Hope i get some help out
of it.

Thanks,

Pramodh.