import java.io.*;
import org.w3c.dom.*;
import org.w3c.dom.html.*;
import javax.xml.parsers.*;
import javax.xml.transform.*;
import javax.xml.transform.stream.*;
import javax.xml.transform.dom.*;
import org.xml.sax.SAXException;
import org.xml.sax.SAXParseException;

/**
 * This class repairs invalid HTML & javascript code produced by 
 * IHMC Concept Map Tools, v2.9.1.
 *
 * This code is experimental and is recommended only for testing purposes.
 *
 * @author Miika Nurminen (minurmin@cc.jyu.fi)
 * @version 0.1 / 2003-06-16
 */
public class Transform {
  // Global value so it can be ref'd by the tree-adapter
  Document document;
  
  
  /**
   * Creates a DOM representation of an XML document given in filename.
   *
   * @param filename input file
   */
  public void fromFile(String filename) throws SAXException,ParserConfigurationException,IOException {
    DocumentBuilderFactory factory =
    DocumentBuilderFactory.newInstance();
    DocumentBuilder builder = factory.newDocumentBuilder();
    document = builder.parse( new File(filename) );
  }
  
  /**
   * Perform actual transforming & element cleanup. 
   * Cleanup is done using W3C DOM interfaces.
   *
   * Javascript code is moved to a separate file that is referenced from 
   * HTML. Most formatting information is omitted or simplified.
   * Inline pictures (filenames starting with 'in_' are embedded directly to
   * HTML page, no links.
   */
  public void doTransform() {
    NodeList nl;
    Element e;
    Node n;
    
    e = (Element)(document.getElementsByTagName("script").item(0));
    
    // cleaning up SCRIPT element
    nl=e.getChildNodes();
    while(nl.getLength()>0) {
      e.removeChild(nl.item(0));
    }
    e.setAttribute("language","JavaScript");
    e.setAttribute("src","cmpopup.js");
    e.appendChild(this.document.createComment("content for Opera 7"));
    
    // removing bgcolor & onclick from body
    e = (Element)(document.getElementsByTagName("body").item(0));
    e.removeAttribute("bgcolor");
    e.removeAttribute("onclick");
    
    // clearing presentation-specific (and partially invalid) *-attributes
    nl = document.getElementsByTagName("table");
    for (int i=0; i<nl.getLength(); i++) {
      e = (Element)(nl.item(i));
      e.removeAttribute("width");
      e.removeAttribute("height");
    }
    
    // clearing invalid valign-attribute
    nl = document.getElementsByTagName("td");
    for (int i=0; i<nl.getLength(); i++) {
      e = (Element)(nl.item(i));
      e.removeAttribute("valign");
    }
    
    //clearing FONT elements away
    nl = document.getElementsByTagName("font");
    while(nl.getLength()>0) {
      n = nl.item(nl.getLength()-1).getParentNode();
      NodeList nl2 = nl.item(nl.getLength()-1).getChildNodes();
      for (int i=0; i<nl2.getLength(); i++) {
        n.insertBefore(nl2.item(i), nl.item(nl.getLength()-1));
      }
      n.removeChild(nl.item(nl.getLength()-1));
    }
    
    // clearing AREA elements
    nl = document.getElementsByTagName("area");
    if (nl.getLength()>0) for (int i=nl.getLength()-1; i>=0; i--) { // reverse because we are removing one element
      e = (Element)(nl.item(i));
      if (e.getAttribute("shape").equals("default")) {
        e.getParentNode().removeChild(e);
      }
      else {
        String s = e.getAttribute("onmouseover");
        if (!s.equals("")) {
          int j = s.indexOf("\\'")+2;
          int k = s.indexOf("\\'",j);
          e.removeAttribute("onmouseover");
          e.setAttribute("alt", s.substring(j,k));
        }
      }
    }
    
    // clearing STYLE attributes
    nl = document.getElementsByTagName("div");
    for (int i=0; i<nl.getLength(); i++) {
      e = (Element)(nl.item(i));
      //      int j = e.getAttribute("style").indexOf("visibility");
      e.setAttribute("style",e.getAttribute("style").replaceAll("visibility: hidden","display: none"));
      e.setAttribute("style",e.getAttribute("style").replaceAll("VISIBILITY: hidden","display: none")); // uppercase in linux
    }
    
    // clearing onclick-events from links
    nl = document.getElementsByTagName("a");
    if (nl.getLength()>0) for (int i=nl.getLength()-1; i>=0; i--) { // reverse because we are removing some links
      e = (Element)(nl.item(i));
      e.removeAttribute("onclick");
      if ((e.getAttribute("href").indexOf("gif")>=0) || (e.getAttribute("href").indexOf("jpg")>=0)) {
        String desc = e.getChildNodes().item(0).getNodeValue();
        if (desc.indexOf("in_")!=0) continue; // if description does start with in_, image is placed inline
        desc=desc.substring(3);
        Element e2 = document.createElement("img");
        e2.setAttribute("src", e.getAttribute("href"));
        e2.setAttribute("alt", desc);
        e.getParentNode().insertBefore(e2, e);
        e.getParentNode().removeChild(e);
      }
    }
    
    // adding dummy ALT tags to get it validated & clearing invalid align attrs
    nl = document.getElementsByTagName("img");
    for (int i=0; i<nl.getLength(); i++) {
      e = (Element)(nl.item(i));
      if (e.getAttribute("alt").equals("")) {
        e.setAttribute("alt","imagemap");
      }
      if (e.getAttribute("align").equals("middle"))
        e.removeAttribute("align");
    }
  }
  
  /**
   * Transforms DOM representation to stream using XHTML 1.0 Transitional doctype.
   * 
   * @param os Writer output stream writer 
   */
  public void toStream(Writer os) throws TransformerException {
    this.document.normalize();
    DOMSource domSource = new DOMSource(this.document);
    StreamResult streamResult = new StreamResult(os);
    TransformerFactory tf = TransformerFactory.newInstance();
    // XSL transformation is required to set correct doctype.
    Transformer serializer = tf.newTransformer(new StreamSource("transformoutput.xsl"));
    serializer.setOutputProperty(OutputKeys.ENCODING,"ISO-8859-1");
    serializer.setOutputProperty(OutputKeys.DOCTYPE_SYSTEM,"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd");
    serializer.setOutputProperty(OutputKeys.DOCTYPE_PUBLIC,"-//W3C//DTD XHTML 1.0 Transitional//EN");
    serializer.setOutputProperty(OutputKeys.INDENT,"yes");
    serializer.transform(domSource, streamResult);
  }
  
  
  /**
   * Main program.
   * @param argv Input arguments. First argument is used as input file, second as output file.
   */
  public static void main(String argv[]) {
    
    if (argv.length != 2) {
      System.err.println("Usage: java Transform srcfile targetfile");
      System.exit(1);
    }
    
    Transform t = new Transform();
    
    try {
      t.fromFile(argv[0]);
      t.doTransform();
      
      PrintWriter out = new PrintWriter(new FileWriter(new File(argv[1])));
      
      t.toStream(out);
      System.out.println("Conversion succesful.");
    }
    catch (Exception e) {
    /* Potential Exception types (omitted for clarity): 
        -SAXException
        -ParserConfigurationException
        -TransformerException
        -IOException */

      System.out.println(e.getMessage());
      e.printStackTrace();
    }
  }
}