source: branches/OKBJavaConnector/ECJClient/src/ec/gp/GPIndividual.java @ 8614

/*
  Copyright 2006 by Sean Luke
  Licensed under the Academic Free License version 3.0
  See the file "LICENSE" for more information
*/


package ec.gp;
import ec.*;
import ec.util.*;
import java.io.*;

/* 
 * GPIndividual.java
 * 
 * Created: Fri Aug 27 17:07:45 1999
 * By: Sean Luke
 */

/**
 * GPIndividual is an Individual used for GP evolution runs.
 * GPIndividuals contain, at the very least, a nonempty array of GPTrees.
 * You can use GPIndividual directly, or subclass it to extend it as
 * you see fit.
 
 * <P>GPIndividuals have two clone methods: clone() and lightClone().  clone() is
 * a deep clone method as usual.  lightClone() is a light clone which does not copy
 * the trees.
 *
 * <p>In addition to serialization for checkpointing, Individuals may read and write themselves to streams in three ways.
 *
 * <ul>
 * <li><b>writeIndividual(...,DataOutput)/readIndividual(...,DataInput)</b>&nbsp;&nbsp;&nbsp;This method
 * transmits or receives an individual in binary.  It is the most efficient approach to sending
 * individuals over networks, etc.  These methods write the evaluated flag and the fitness, then
 * call <b>readGenotype/writeGenotype</b>, which you must implement to write those parts of your 
 * Individual special to your functions-- the default versions of readGenotype/writeGenotype throw errors.
 * You don't need to implement them if you don't plan on using read/writeIndividual.
 *
 * <li><b>printIndividual(...,PrintWriter)/readIndividual(...,LineNumberReader)</b>&nbsp;&nbsp;&nbsp;This
 * approach transmits or receives an indivdual in text encoded such that the individual is largely readable
 * by humans but can be read back in 100% by ECJ as well.  Because GPIndividuals are often very large,
 * <b>GPIndividual has overridden these methods -- they work differently than in Individual (the superclass).</b>  In specific:
 * <b>readIndividual</b> by default reads in the fitness and the evaluation flag, then calls <b>parseGenotype</b> 
 * to read in the trees (via GPTree.readTree(...)).
 * However <b>printIndividual</b> by default prints the fitness and evaluation flag, and prints all the trees
 * by calling GPTree.printTree(...).  It does not call <b>genotypeToString</b> at all.  This
 * is because it's very wasteful to build up a large string holding the printed form of the GPIndividual 
 * just to pump it out a stream once.
 *
 * <li><b>printIndividualForHumans(...,PrintWriter)</b>&nbsp;&nbsp;&nbsp;This
 * approach prints an individual in a fashion intended for human consumption only. Because GPIndividuals are often very large,
 * <b>GPIndividual has overridden this methods -- it works differently than in Individual (the superclass).</b>  In specific:
 * <b>printIndividual</b> by default prints the fitness and evaluation flag, and prints all the trees
 * by calling GPTree.printTreeForHumans(...).  It does not call <b>genotypeToStringForHumans</b> at all.  This
 * is because it's very wasteful to build up a large string holding the printed form of the GPIndividual 
 * just to pump it out a stream once.
 *
 * <p>In general, the various readers and writers do three things: they tell the Fitness to read/write itself,
 * they read/write the evaluated flag, and they read/write the GPTree array (by having each GPTree read/write
 * itself).  If you add instance variables to GPIndividual, you'll need to read/write those variables as well.


 <p><b>Parameters</b><br>
 <table>
 <tr><td valign=top><i>base</i>.<tt>numtrees</tt><br>
 <font size=-1>int &gt;= 1</font></td>
 <td valign=top>(number of trees in the GPIndividual)</td></tr>

 <tr><td valign=top><i>base</i>.<tt>tree.</tt><i>n</i><br>
 <font size=-1>classname, inherits or = ec.gp.GPTree</font></td>
 <td valign=top>(class of tree <i>n</i> in the individual)</td></tr>
 </table>

 <p><b>Default Base</b><br>
 gp.individual

 <p><b>Parameter bases</b><br>
 <table>
 <tr><td valign=top><i>base</i>.<tt>tree.</tt><i>n</i></td>
 <td>tree <i>n</i> in the individual</td></tr>
 </table>

 *
 * @author Sean Luke
 * @version 1.0 
 */

public class GPIndividual extends Individual
    {
    public static final String P_NUMTREES = "numtrees";
    public static final String P_TREE = "tree";
    
    public GPTree[] trees;
    
    public Parameter defaultBase()
        {
        return GPDefaults.base().push(P_INDIVIDUAL);
        }

    public boolean equals(Object ind)
        {
        if (!(this.getClass().equals(ind.getClass()))) return false;  // GPIndividuals are special.
        GPIndividual i = (GPIndividual)ind;
        if (trees.length != i.trees.length) return false;
        // this default version works fine for most GPIndividuals.
        for(int x=0;x<trees.length;x++)
            if (!(trees[x].treeEquals(i.trees[x]))) return false;
        return true;
        }
    
    public int hashCode()
        {
        // stolen from GPNode.  It's a decent algorithm.
        int hash = this.getClass().hashCode();
        
        for(int x=0;x<trees.length;x++)
            hash =
                // Rotate hash and XOR
                (hash << 1 | hash >>> 31 ) ^
                trees[x].treeHashCode();
        return hash;
        }

    /** Sets up a prototypical GPIndividual with those features which it
        shares with other GPIndividuals in its species, and nothing more. */

    public void setup(final EvolutionState state, final Parameter base)
        {
        super.setup(state,base);  // actually unnecessary (Individual.setup() is empty)

        Parameter def = defaultBase();

        // set my evaluation to false
        evaluated = false;

        // how many trees?
        int t = state.parameters.getInt(base.push(P_NUMTREES),def.push(P_NUMTREES),1);  // at least 1 tree for GP!
        if (t <= 0) 
            state.output.fatal("A GPIndividual must have at least one tree.",
                base.push(P_NUMTREES),def.push(P_NUMTREES));
        
        // load the trees
        trees = new GPTree[t];

        for (int x=0;x<t;x++)
            {
            Parameter p = base.push(P_TREE).push(""+x);
            trees[x] = (GPTree)(state.parameters.getInstanceForParameterEq(
                    p,def.push(P_TREE).push(""+x),GPTree.class));
            trees[x].owner = this;
            trees[x].setup(state,p);
            }
        
        // now that our function sets are all associated with trees,
        // give the nodes a chance to determine whether or not this is
        // going to work for them (especially the ADFs).
        GPInitializer initializer = ((GPInitializer)state.initializer);
        for (int x=0;x<t;x++)
            {
            for(int w = 0;w < trees[x].constraints(initializer).functionset.nodes.length;w++)
                {
                GPNode[] gpfi = trees[x].constraints(initializer).functionset.nodes[w];
                for (int y = 0;y<gpfi.length;y++)
                    gpfi[y].checkConstraints(state,x,this,base);
                }
            }
        // because I promised with checkConstraints(...)
        state.output.exitIfErrors();
        }


    /** Verification of validity of the GPIndividual -- strictly for debugging purposes only */
    public final void verify(EvolutionState state)
        {
        if (!(state.initializer instanceof GPInitializer))
            { state.output.error("Initializer is not a GPInitializer"); return; }
            
        GPInitializer initializer = (GPInitializer)(state.initializer);

        if (trees==null) 
            { state.output.error("Null trees in GPIndividual."); return; }
        for(int x=0;x<trees.length;x++) if (trees[x]==null) 
                                            { state.output.error("Null tree (#"+x+") in GPIndividual."); return; }
        for(int x=0;x<trees.length;x++)
            trees[x].verify(state);
        state.output.exitIfErrors();
        }

    /** Prints just the trees of the GPIndividual.  Broken out like this to be used by GEIndividual to avoid
        re-printing the fitness and evaluated premables. */
    public void printTrees(final EvolutionState state, final int log)
        {
        for(int x=0;x<trees.length;x++)
            {
            state.output.println("Tree " + x + ":",log);
            trees[x].printTreeForHumans(state,log);
            }
        }

    public void printIndividualForHumans(final EvolutionState state, final int log)
        {
        state.output.println(EVALUATED_PREAMBLE + (evaluated ? "true" : "false"), log);
        fitness.printFitnessForHumans(state,log);
        printTrees(state,log);
        }

    public void printIndividual(final EvolutionState state, final int log)
        {
        state.output.println(EVALUATED_PREAMBLE + Code.encode(evaluated), log);
        fitness.printFitness(state,log);
        for(int x=0;x<trees.length;x++)
            {
            state.output.println("Tree " + x + ":",log);
            trees[x].printTree(state,log);
            }   
        }
            
    public void printIndividual(final EvolutionState state,
        final PrintWriter writer)
        {
        writer.println(EVALUATED_PREAMBLE + Code.encode(evaluated));
        fitness.printFitness(state,writer);
        for(int x=0;x<trees.length;x++)
            {
            writer.println("Tree " + x + ":");
            trees[x].printTree(state,writer);
            }   
        }
        
    /** Overridden for the GPIndividual genotype. */
    public void writeGenotype(final EvolutionState state,
        final DataOutput dataOutput) throws IOException
        {
        dataOutput.writeInt(trees.length);
        for(int x=0;x<trees.length;x++)
            trees[x].writeTree(state,dataOutput);
        }

    /** Overridden for the GPIndividual genotype. */
    public void readGenotype(final EvolutionState state,
        final DataInput dataInput) throws IOException
        {
        int treelength = dataInput.readInt();
        if (trees == null || treelength != trees.length) // wrong size!
            state.output.fatal("Number of trees differ in GPIndividual when reading from readGenotype(EvolutionState, DataInput).");
        for(int x=0;x<trees.length;x++)
            trees[x].readTree(state,dataInput);
        }

    public void parseGenotype(final EvolutionState state,
        final LineNumberReader reader) throws IOException
        {
        // Read my trees
        for(int x=0;x<trees.length;x++)
            {
            reader.readLine();  // throw it away -- it's the tree indicator
            trees[x].readTree(state,reader);
            }
        }

    /** Deep-clones the GPIndividual.  Note that you should not deep-clone the prototypical GPIndividual
        stored in GPSpecies: they contain blank GPTrees with null roots, and this method,
        which calls GPTree.clone(), will produce a NullPointerException as a result. Instead, you probably
        want to use GPSpecies.newIndividual(...) if you're thinking of playing with the prototypical
        GPIndividual. */
        
    public Object clone()
        {
        // a deep clone
                
        GPIndividual myobj = (GPIndividual)(super.clone());

        // copy the tree array
        myobj.trees = new GPTree[trees.length];
        for(int x=0;x<trees.length;x++)
            {
            myobj.trees[x] = (GPTree)(trees[x].clone());  // force a deep clone
            myobj.trees[x].owner = myobj;  // reset owner away from me
            }
        return myobj;
        }

    /** Like clone(), but doesn't force the GPTrees to deep-clone themselves. */
    public GPIndividual lightClone()
        {
        // a light clone
        GPIndividual myobj = (GPIndividual)(super.clone());
        
        // copy the tree array
        myobj.trees = new GPTree[trees.length];
        for(int x=0;x<trees.length;x++)
            {
            myobj.trees[x] = (GPTree)(trees[x].lightClone());  // note light-cloned!
            myobj.trees[x].owner = myobj;  // reset owner away from me
            }
        return myobj;
        }

    /** Returns the "size" of the individual, namely, the number of nodes
        in all of its subtrees.  */
    public long size()
        {
        long size = 0;
        for(int x=0;x<trees.length;x++)
            size += trees[x].child.numNodes(GPNode.NODESEARCH_ALL);
        return size;
        }

    }