source: trunk/gsdl3/src/java/org/greenstone/gsdl3/gs3build/doctypes/HTMLDocument.java@ 5800

package org.greenstone.gsdl3.gs3build.doctypes;

import java.io.File;
import java.net.URL;
import java.net.MalformedURLException;

import java.util.List;
import java.util.Map;
import java.util.ArrayList;

import org.greenstone.gsdl3.gs3build.metadata.METSFile;
import org.greenstone.gsdl3.gs3build.util.*;

public class HTMLDocument extends AbstractDocument
{
    public static final String HTML_DOCUMENT_TYPE = "HTML";

    /**
     *  Create the HTMLDocument from a given URL - the URL may in fact be a reference
     *  to a local file.
     *
     *  @param <code>URL</code> The location from which to load the file
     */
    public HTMLDocument(URL url)
    {   super(url);

        HTMLDoc htmlDoc;
        if (url.toString().startsWith("file://"))
        {   htmlDoc = new HTMLDoc(url, url.toString().substring(7));
        }
        else
        {   htmlDoc = new HTMLDoc(url);
        }

        this._extractDocumentFiles(htmlDoc);
        this._extractDocumentMetadata(htmlDoc);
    }

    private void _extractDocumentMetadata(HTMLDoc htmlDoc)
    {   HTMLBlock codedContent = htmlDoc.getCodedContent();
        boolean inTitle = false;
        StringBuffer title = new StringBuffer();

        for (int e = 0; e < codedContent.size(); e ++)
        {   if (codedContent.elementAt(e) instanceof HTMLTag)
            {   HTMLTag tag = (HTMLTag) codedContent.elementAt(e);

                if (tag.tagName().equals("meta"))
                {   // check that the name of the metadata item exists
                    String name = tag.idValue("name");
                    if (name == null || name.length() == 0) {
                        continue;
                    }

                    // get the value, if it exists
                    String value = tag.idValue("content");
                    if (value != null && value.length() > 0) {
                        System.out.println("  " + value);
                    }
                    // if value does not exist, default it to being the same
                    // as the name.
                    else {
                        value = name;
                    }

                    this.addDocumentMetadata(name, value);
                }
                else if (tag.tagName().equals("title"))
                { inTitle = true;
                }
                else if (tag.tagName().equals("/title"))
                {   inTitle = false;
                }
                // cut off when real body content appears - not a perfect
                // implementation, just cheap & cheerful
                else if (tag.tagName().equals("/head"))
                {   break;
                }
                else if (tag.tagName().equals("body"))
                {   break;
                }
            }
            else if (inTitle == true)
            { title.append(codedContent.elementAt(e).toString());
            }
        }
        if (title.length() > 0)
        { this.addDocumentMetadata("title", title.toString());
        }
    }

    private void _extractDocumentFiles(HTMLDoc htmlDoc)
    { URL homeUrl = this.fileSet.getFile(0).getLocation();

      HTMLBlock codedContent = htmlDoc.getCodedContent();
      for (int e = 0; e < codedContent.size(); e ++)
      { if (codedContent.elementAt(e) instanceof HTMLTag)
    { HTMLTag tag = (HTMLTag) codedContent.elementAt(e);

      if (tag.tagName().equals("img"))
      { String location = tag.idValue("src");

        try 
        { // make the url for the image, and then add it to the document list of
          // 
          URL imgUrl = new URL(homeUrl, location);
          METSFile file = this.fileSet.addFile(imgUrl);
          this.structureSet.getStructure("All").getDivision("All").addFileReference(file.getID());
        }
        catch (MalformedURLException ex)
        { // TODO: report exception/failure to resolve...
        }
      }
    }
      }
    }

    public String getDocumentType()
    {   return HTML_DOCUMENT_TYPE;
    }

    public String getDocumentText()
    { HTMLDoc htmlDoc = new HTMLDoc((URL) this.fileSet.getFile(0).getLocation(), this.fileSet.getFile(0).toString().substring(7));
      return htmlDoc.getContent();
    }
}