source: trunk/gsdl3/src/java/org/greenstone/gsdl3/gs3build/util/HTMLParser.java@ 6700

package org.greenstone.gsdl3.gs3build.util;

import java.applet.*;

public class HTMLParser
{ HTMLDoc document;
    int         pos;
    int         lastpos;

    public HTMLParser(HTMLDoc document)
    { this.document = document;
        this.pos = 0;
    }

    public void startParse()
    { this.pos = 0;
        this.lastpos = -1;
//      System.out.println("Starting "+document.urlString());
    }

    public int atParse()
    { return this.pos;
    }

    public int lastParse()
    { return this.lastpos;
    }

  // do a full text/tag parse
  public String fullParse()
  { String reply = null;
    int      end;
    int      start;

    this.lastpos = this.pos;

    if (this.pos >= this.document.getContent().length())
    { return reply;
    }

    if (this.document.getContent().charAt(this.pos) == '<')
    {
      start = this.pos;
      
      // if we're not at the end of the document,
      // read the rest of the tag
      if (this.pos == this.document.getContent().length() - 1)
      { this.pos = this.document.getContent().length();
        return reply;
      }

      // if the tag is a comment
      if (this.pos < this.document.getContent().length() - 3 &&
      this.document.getContent().substring(this.pos, this.pos+4).equals("<!--"))
      { end = this.document.getContent().substring(this.pos).indexOf("-->") + 3 + this.pos;
        reply = this.document.getContent().substring(this.pos, end);
        this.pos = end;
      }
      else 
      { // read up to the end of the tag
    end = this.pos + 1;
    while (end < this.document.getContent().length() &&
           this.document.getContent().charAt(end) != '>')
    { end ++;
    }

    // get the whole of the tag into 'reply', and
    // set the current pos to immediately after the tag
    if (end < this.document.getContent().length())
    { reply = this.document.getContent().substring(this.pos, end + 1);
      this.pos = end + 1;
    }
    // patch the trailing > onto the tag string
    else
    { reply = this.document.getContent().substring(this.pos, end) + ">";
      this.pos = end;
    }
      }
    }
    else
    { // hunt for the beginning of the next tag
      start = this.pos;
      while ((this.pos < this.document.getContent().length()) &&
         (this.document.getContent().charAt(this.pos) != '<'))
      { this.pos ++;
      }

      // return everything up to that tag
      reply = this.document.getContent().substring(start, this.pos);
    }
    return reply;
  }

    // Get the next tag to parse
    public String nextParse()
    { String reply  = null;
        int      end;

        this.lastpos    = this.pos;

        if (this.document.getContent() == null)
        {   return null;
        }

        if (this.pos >= this.document.getContent().length())
        { return reply;
        }

        // hunt for the beginning of the next tag
        while ((this.pos < this.document.getContent().length()) &&
                        (this.document.getContent().charAt(this.pos) != '<'))
        { this.pos ++;
        }

        // if we're not at the end of the document,
        // read the rest of the tag
        if (this.pos < this.document.getContent().length())
        { if (this.pos == this.document.getContent().length() - 1)
            { this.pos = this.document.getContent().length();
              return reply;
            }

          end = this.pos + 1;
            while ( end < this.document.getContent().length() &&
                    this.document.getContent().charAt(end) != '>')
            { end ++;
            }

            // get the whole of the tag into 'reply', and
            // set the current pos to immediately after the tag
            if (end < this.document.getContent().length())
            { reply = this.document.getContent().substring(this.pos, end + 1);
                this.pos = end + 1;
            }
            else
            { this.pos = end;
            }
        }
        return reply;
    }


    /* --
         -- return next HREF value we come across in the document
         --
         -- NB: the checking of the quotes etc is rather lazy - and should be tidied
         --
    */
    public String nextHREF()
    { String reply;
        int         start, end;
        boolean quoted;
        HTMLTag tag;

        reply = this.nextParse();
        while (this.pos < this.document.getContent().length())
        { tag = new HTMLTag(reply);

            if (tag.tagName().equals("a"))
            { start = reply.indexOf("href");
                if (start == -1)
                { start = reply.indexOf("HREF");
                }

                if (start >= 0)
                { start += 4;
                    quoted = false;
                    while   (reply.charAt(start) == ' ' ||
                                 reply.charAt(start) == '=' ||
                                 reply.charAt(start) == '"')
                    { if (reply.charAt(start) == '"')
                        { quoted = true;
                        }
                        start ++;
                    }

                    end = -1;
                    if (quoted)
                    { end = reply.indexOf('"', start+1);
                    }
                    if (end == -1)
                    { end = reply.indexOf(' ', start+1);
                        if (end == -1)
                        { end = reply.length() - 1;
                        }
                    }

                    reply = reply.substring(start, end);
                  return reply;
                }
            }
            reply = this.nextParse();
        }
        return null;
    }

    // Return the next link in the page
    public String nextLink(AppletContext ac)
    { String reply, reply2;
        int         start, end;
        boolean quoted;
        HTMLTag tag;

        if (this.document.getContent() == null)
        {   return null;
        }

        reply = this.nextParse();
        while (this.pos < this.document.getContent().length())
        { tag = new HTMLTag(reply);
            if (tag.tagName().equals("a"))
            { reply = tag.idValue("href");
            }
            else if (tag.tagName().equals("frame")) // image tag
            { reply = tag.idValue("src");
            }
            else if (tag.tagName().equals("area")) // image map area
            { reply = tag.idValue("href");
            }
            else if (tag.tagName().equals("frame"))
            {   reply = tag.idValue("src");
            }
            else
            {   reply   = null;
            }

            if (reply != null)
            {   return reply;
            }

            reply = this.nextParse();
        }
        return null;
    }
}