/*
 * Created on Mar 15, 2005
 *
 * TODO To change the template for this generated file go to
 * Window - Preferences - Java - Code Style - Code Templates
 */
package cc.glsn;

import java.io.File;
import java.io.FileInputStream;
import java.util.ArrayList;

/**
 * @author fireduck
 *
 * TODO To change the template for this generated type comment go to
 * Window - Preferences - Java - Code Style - Code Templates
 */
public class HTMLTableExtractor
{
    
    public boolean RemoveBraces;
    
    public HTMLTableExtractor()
    {
        RemoveBraces=true;
        
    }
    
    /**
     * 
     * @param HTMLDoc
     * @return An arraylist of html table objects
     */
    public ArrayList extract(String HTMLDoc)
    {
        ArrayList Output=new ArrayList();
        
        Stato S=new Stato(HTMLDoc,0,HTMLDoc.length());
        
        
        
        while(true)
        {
            ArrayList newTable=extractTable(S);
            
            if (newTable==null) return Output;
            
            Output.add(newTable);
            
            
            
            
        }
        
    }
    
    public ArrayList extractFile(String Filename) throws java.io.IOException
    {
        File F=new File(Filename);
        int len=new Long(F.length()).intValue();
        
        byte Buff[]=new byte[len];
        
        FileInputStream FIS=new FileInputStream(F);
        
        
        int r=0;
        while(r<len)
        {
            int delta=FIS.read(Buff,r,len-r);
            if (delta<0) throw new java.io.IOException("Unable to read more.");
            r+=delta;
            
        }
        
        
        String S=new String(Buff);
        
        return extract(S);
        
        
        
    }
    
    
    private Stato extractEntry(Stato S, String Start, String End)
    {
        
        
        int sidx=S.Str.indexOf(Start,S.start);
        if (sidx<0) return null;
        if (sidx>=S.end) return null;
        
        int eidx=S.Str.indexOf(End,sidx);
        if (eidx<0) return null;
        if (eidx>=S.end) return null;
        
        Stato S2=new Stato(S.Str,sidx,eidx);
        
        
        S.start=eidx+End.length();
        
        
        return S2;
        
        
    }
    
    private ArrayList extractTable(Stato S)
    {
        Stato TableData=extractEntry(S,"<table","</table>");
        if (TableData==null) return null;
        
        ArrayList Table=new ArrayList();
        
        
        while(true)
        {
            ArrayList Row=extractTableRow(TableData);
            if (Row==null) return Table;
            
            Table.add(Row);
            
            
        }
        
    }

    private ArrayList extractTableRow(Stato S)
    {
        Stato RowData=extractEntry(S,"<tr","</tr>");
        if (RowData==null) return null;
        
        ArrayList Row=new ArrayList();
        
        
        while(true)
        {
            String D=extractTableData(RowData);
            if (D==null) return Row;
            
            Row.add(D);
            
            
        }        
    }

    private String extractTableData(Stato S)
    {
        Stato TDA=extractEntry(S,"<td","</td>");
        Stato TDB=extractEntry(S,"<th","</th>");
        
        Stato TD=TDA;
        
        if ((TD==null) || ((TDB != null) && (TDB.start < TDA.start)))
        {
            TD=TDB;
        }
        
        if (TD==null) return null;
        
        String strTD=TD.Str.substring(TD.start,TD.end);
        if (RemoveBraces)
        {
            strTD=removeBraces(strTD);
        }
        
        //System.out.println("Record: " + strTD);
        
        return strTD;
        
    }
    
    private String removeBraces(String S)
    {
        int sidx=S.indexOf("<");
        int eidx=S.indexOf(">");
        if ((sidx >=0) && (sidx < eidx))
        {
            String S2=S.substring(0,sidx) + S.substring(eidx+1);
            
            return removeBraces(S2);
            
            
        }
        
        return S;
        
        
    }
    


    
    
    private class Stato
    {
        int start;
        int end;
        String Str;
        
        public Stato(String H, int Start, int End)
        {
            Str=H;
            start=Start;
            end=End;
            
            if (end-start < 2048)
            {
                Str=H.substring(Start,End);
                start=0;
                end=Str.length();
                
            }
            
        }
        
        
        
    }
    

}