// This file was automatically generated with the command: 

 //     java org.galagosearch.tupleflow.typebuilder.TypeBuilderMojo ...

 package org.galagosearch.core.types;

 import org.galagosearch.tupleflow.Utility;

 import org.galagosearch.tupleflow.ArrayInput;

 import org.galagosearch.tupleflow.ArrayOutput;

 import org.galagosearch.tupleflow.Order;   

 import org.galagosearch.tupleflow.OrderedWriter;

 import org.galagosearch.tupleflow.Type; 

 import org.galagosearch.tupleflow.TypeReader;

 import org.galagosearch.tupleflow.Step; 

 import org.galagosearch.tupleflow.IncompatibleProcessorException;

 import org.galagosearch.tupleflow.ReaderSource;

 import java.io.IOException;             

 import java.io.EOFException;

 import java.io.UnsupportedEncodingException;

 import java.util.ArrayList;

 import java.util.Arrays;   

 import java.util.Comparator;

 import java.util.PriorityQueue;

 import java.util.Collection;

 public class DocumentWordProbability implements Type<DocumentWordProbability> {

     public String document;

     public byte[] word;

     public double probability; 

     public DocumentWordProbability() {}

     public DocumentWordProbability(String document, byte[] word, double probability) {

         this.document = document;

         this.word = word;

         this.probability = probability;

     }  

     public String toString() {

         try {

             return String.format("%s,%s,%f",

                                    document, new String(word, "UTF-8"), probability);

         } catch(UnsupportedEncodingException e) {

             throw new RuntimeException("Couldn't convert string to UTF-8.");

         }

     } 

     public Order<DocumentWordProbability> getOrder(String... spec) {

         if (Arrays.equals(spec, new String[] { "+document", "+word" })) {

             return new DocumentWordOrder();

         }

         if (Arrays.equals(spec, new String[] { "+word" })) {

             return new WordOrder();

         }

         if (Arrays.equals(spec, new String[] { "+document" })) {

             return new DocumentOrder();

         }

         return null;

     } 

     public interface Processor extends Step, org.galagosearch.tupleflow.Processor<DocumentWordProbability> {

         public void process(DocumentWordProbability object) throws IOException;

         public void close() throws IOException;

     }                        

     public interface Source extends Step {

     }

     public static class DocumentWordOrder implements Order<DocumentWordProbability> {

         public int hash(DocumentWordProbability object) {

             int h = 0;

             h += Utility.hash(object.document);

             h += Utility.hash(object.word);

             return h;

         } 

         public Comparator<DocumentWordProbability> greaterThan() {

             return new Comparator<DocumentWordProbability>() {

                 public int compare(DocumentWordProbability one, DocumentWordProbability two) {

                     int result = 0;

                     do {

                         result = + Utility.compare(one.document, two.document);

                         if(result != 0) break;

                         result = + Utility.compare(one.word, two.word);

                         if(result != 0) break;

                     } while (false);

                     return -result;

                 }

             };

         }     

         public Comparator<DocumentWordProbability> lessThan() {

             return new Comparator<DocumentWordProbability>() {

                 public int compare(DocumentWordProbability one, DocumentWordProbability two) {

                     int result = 0;

                     do {

                         result = + Utility.compare(one.document, two.document);

                         if(result != 0) break;

                         result = + Utility.compare(one.word, two.word);

                         if(result != 0) break;

                     } while (false);

                     return result;

                 }

             };

         }     

         public TypeReader<DocumentWordProbability> orderedReader(ArrayInput _input) {

             return new ShreddedReader(_input);

         }    

         public TypeReader<DocumentWordProbability> orderedReader(ArrayInput _input, int bufferSize) {

             return new ShreddedReader(_input, bufferSize);

         }    

         public OrderedWriter<DocumentWordProbability> orderedWriter(ArrayOutput _output) {

             ShreddedWriter w = new ShreddedWriter(_output);

             return new OrderedWriterClass(w); 

         }                                    

         public static class OrderedWriterClass extends OrderedWriter< DocumentWordProbability > {

             DocumentWordProbability last = null;

             ShreddedWriter shreddedWriter = null; 

             public OrderedWriterClass(ShreddedWriter s) {

                 this.shreddedWriter = s;

             }

             public void process(DocumentWordProbability object) throws IOException {

                boolean processAll = false;

                if (processAll || last == null || 0 != Utility.compare(object.document, last.document)) { processAll = true; shreddedWriter.processDocument(object.document); }

                if (processAll || last == null || 0 != Utility.compare(object.word, last.word)) { processAll = true; shreddedWriter.processWord(object.word); }

                shreddedWriter.processTuple(object.probability);

                last = object;

             }           

             public void close() throws IOException {

                 shreddedWriter.close();

             }

             public Class<DocumentWordProbability> getInputClass() {

                 return DocumentWordProbability.class;

             }

         } 

         public ReaderSource<DocumentWordProbability> orderedCombiner(Collection<TypeReader<DocumentWordProbability>> readers, boolean closeOnExit) {

             ArrayList<ShreddedReader> shreddedReaders = new ArrayList();

             for (TypeReader<DocumentWordProbability> reader : readers) {

                 shreddedReaders.add((ShreddedReader)reader);

             }

             return new ShreddedCombiner(shreddedReaders, closeOnExit);

         }                  

         public DocumentWordProbability clone(DocumentWordProbability object) {

             DocumentWordProbability result = new DocumentWordProbability();

             if (object == null) return result;

             result.document = object.document; 

             result.word = object.word; 

             result.probability = object.probability; 

             return result;

         }                 

         public Class<DocumentWordProbability> getOrderedClass() {

             return DocumentWordProbability.class;

         }                           

         public String[] getOrderSpec() {

             return new String[] {"+document", "+word"};

         }

         public static String getSpecString() {

             return "+document +word";

         }

         public interface ShreddedProcessor extends Step {

             public void processDocument(String document) throws IOException;

             public void processWord(byte[] word) throws IOException;

             public void processTuple(double probability) throws IOException;

             public void close() throws IOException;

         }    

         public interface ShreddedSource extends Step {

         }                                              

         public static class ShreddedWriter implements ShreddedProcessor {

             ArrayOutput output;

             ShreddedBuffer buffer = new ShreddedBuffer();

             String lastDocument;

             byte[] lastWord;

             boolean lastFlush = false;

             public ShreddedWriter(ArrayOutput output) {

                 this.output = output;

             }                        

             public void close() throws IOException {

                 flush();

             }

             public void processDocument(String document) {

                 lastDocument = document;

                 buffer.processDocument(document);

             }

             public void processWord(byte[] word) {

                 lastWord = word;

                 buffer.processWord(word);

             }

             public final void processTuple(double probability) throws IOException {

                 if (lastFlush) {

                     if(buffer.documents.size() == 0) buffer.processDocument(lastDocument);

                     if(buffer.words.size() == 0) buffer.processWord(lastWord);

                     lastFlush = false;

                 }

                 buffer.processTuple(probability);

                 if (buffer.isFull())

                     flush();

             }

             public final void flushTuples(int pauseIndex) throws IOException {

                 while (buffer.getReadIndex() < pauseIndex) {

                     output.writeDouble(buffer.getProbability());

                     buffer.incrementTuple();

                 }

             }  

             public final void flushDocument(int pauseIndex) throws IOException {

                 while (buffer.getReadIndex() < pauseIndex) {

                     int nextPause = buffer.getDocumentEndIndex();

                     int count = nextPause - buffer.getReadIndex();

                     output.writeString(buffer.getDocument());

                     output.writeInt(count);

                     buffer.incrementDocument();

                     flushWord(nextPause);

                     assert nextPause == buffer.getReadIndex();

                 }

             }

             public final void flushWord(int pauseIndex) throws IOException {

                 while (buffer.getReadIndex() < pauseIndex) {

                     int nextPause = buffer.getWordEndIndex();

                     int count = nextPause - buffer.getReadIndex();

                     output.writeBytes(buffer.getWord());

                     output.writeInt(count);

                     buffer.incrementWord();

                     flushTuples(nextPause);

                     assert nextPause == buffer.getReadIndex();

                 }

             }

             public void flush() throws IOException { 

                 flushDocument(buffer.getWriteIndex());

                 buffer.reset(); 

                 lastFlush = true;

             }                           

         }

         public static class ShreddedBuffer {

             ArrayList<String> documents = new ArrayList();

             ArrayList<byte[]> words = new ArrayList();

             ArrayList<Integer> documentTupleIdx = new ArrayList();

             ArrayList<Integer> wordTupleIdx = new ArrayList();

             int documentReadIdx = 0;

             int wordReadIdx = 0;

             double[] probabilitys;

             int writeTupleIndex = 0;

             int readTupleIndex = 0;

             int batchSize;

             public ShreddedBuffer(int batchSize) {

                 this.batchSize = batchSize;

                 probabilitys = new double[batchSize];

             }                              

             public ShreddedBuffer() {    

                 this(10000);

             }                                                                                                                    

             public void processDocument(String document) {

                 documents.add(document);

                 documentTupleIdx.add(writeTupleIndex);

             }                                      

             public void processWord(byte[] word) {

                 words.add(word);

                 wordTupleIdx.add(writeTupleIndex);

             }                                      

             public void processTuple(double probability) {

                 assert documents.size() > 0;

                 assert words.size() > 0;

                 probabilitys[writeTupleIndex] = probability;

                 writeTupleIndex++;

             }

             public void resetData() {

                 documents.clear();

                 words.clear();

                 documentTupleIdx.clear();

                 wordTupleIdx.clear();

                 writeTupleIndex = 0;

             }                  

             public void resetRead() {

                 readTupleIndex = 0;

                 documentReadIdx = 0;

                 wordReadIdx = 0;

             } 

             public void reset() {

                 resetData();

                 resetRead();

             } 

             public boolean isFull() {

                 return writeTupleIndex >= batchSize;

             }

             public boolean isEmpty() {

                 return writeTupleIndex == 0;

             }                          

             public boolean isAtEnd() {

                 return readTupleIndex >= writeTupleIndex;

             }           

             public void incrementDocument() {

                 documentReadIdx++;  

             }                                                                                              

             public void autoIncrementDocument() {

                 while (readTupleIndex >= getDocumentEndIndex() && readTupleIndex < writeTupleIndex)

                     documentReadIdx++;

             }                 

             public void incrementWord() {

                 wordReadIdx++;  

             }                                                                                              

             public void autoIncrementWord() {

                 while (readTupleIndex >= getWordEndIndex() && readTupleIndex < writeTupleIndex)

                     wordReadIdx++;

             }                 

             public void incrementTuple() {

                 readTupleIndex++;

             }                    

             public int getDocumentEndIndex() {

                 if ((documentReadIdx+1) >= documentTupleIdx.size())

                     return writeTupleIndex;

                 return documentTupleIdx.get(documentReadIdx+1);

             }

             public int getWordEndIndex() {

                 if ((wordReadIdx+1) >= wordTupleIdx.size())

                     return writeTupleIndex;

                 return wordTupleIdx.get(wordReadIdx+1);

             }

             public int getReadIndex() {

                 return readTupleIndex;

             }   

             public int getWriteIndex() {

                 return writeTupleIndex;

             } 

             public String getDocument() {

                 assert readTupleIndex < writeTupleIndex;

                 assert documentReadIdx < documents.size();

                 return documents.get(documentReadIdx);

             }

             public byte[] getWord() {

                 assert readTupleIndex < writeTupleIndex;

                 assert wordReadIdx < words.size();

                 return words.get(wordReadIdx);

             }

             public double getProbability() {

                 assert readTupleIndex < writeTupleIndex;

                 return probabilitys[readTupleIndex];

             }                                         

             public void copyTuples(int endIndex, ShreddedProcessor output) throws IOException {

                 while (getReadIndex() < endIndex) {

                    output.processTuple(getProbability());

                    incrementTuple();

                 }

             }                                                                           

             public void copyUntilIndexDocument(int endIndex, ShreddedProcessor output) throws IOException {

                 while (getReadIndex() < endIndex) {

                     output.processDocument(getDocument());

                     assert getDocumentEndIndex() <= endIndex;

                     copyUntilIndexWord(getDocumentEndIndex(), output);

                     incrementDocument();

                 }

             } 

             public void copyUntilIndexWord(int endIndex, ShreddedProcessor output) throws IOException {

                 while (getReadIndex() < endIndex) {

                     output.processWord(getWord());

                     assert getWordEndIndex() <= endIndex;

                     copyTuples(getWordEndIndex(), output);

                     incrementWord();

                 }

             }  

             public void copyUntilDocument(ShreddedBuffer other, ShreddedProcessor output) throws IOException {

                 while (!isAtEnd()) {

                     if (other != null) {   

                         assert !other.isAtEnd();

                         int c = + Utility.compare(getDocument(), other.getDocument());

                         if (c > 0) {

                             break;   

                         }

                         output.processDocument(getDocument());

                         if (c < 0) {

                             copyUntilIndexWord(getDocumentEndIndex(), output);

                         } else if (c == 0) {

                             copyUntilWord(other, output);

                             autoIncrementDocument();

                             break;

                         }

                     } else {

                         output.processDocument(getDocument());

                         copyUntilIndexWord(getDocumentEndIndex(), output);

                     }

                     incrementDocument();  

                 }

             }

             public void copyUntilWord(ShreddedBuffer other, ShreddedProcessor output) throws IOException {

                 while (!isAtEnd()) {

                     if (other != null) {   

                         assert !other.isAtEnd();

                         int c = + Utility.compare(getWord(), other.getWord());

                         if (c > 0) {

                             break;   

                         }

                         output.processWord(getWord());

                         copyTuples(getWordEndIndex(), output);

                     } else {

                         output.processWord(getWord());

                         copyTuples(getWordEndIndex(), output);

                     }

                     incrementWord();  

                     if (getDocumentEndIndex() <= readTupleIndex)

                         break;   

                 }

             }

             public void copyUntil(ShreddedBuffer other, ShreddedProcessor output) throws IOException {

                 copyUntilDocument(other, output);

             }

         }                         

         public static class ShreddedCombiner implements ReaderSource<DocumentWordProbability>, ShreddedSource {   

             public ShreddedProcessor processor;

             Collection<ShreddedReader> readers;       

             boolean closeOnExit = false;

             boolean uninitialized = true;

             PriorityQueue<ShreddedReader> queue = new PriorityQueue<ShreddedReader>();

             public ShreddedCombiner(Collection<ShreddedReader> readers, boolean closeOnExit) {

                 this.readers = readers;                                                       

                 this.closeOnExit = closeOnExit;

             }

             public void setProcessor(Step processor) throws IncompatibleProcessorException {  

                 if (processor instanceof ShreddedProcessor) {

                     this.processor = new DuplicateEliminator((ShreddedProcessor) processor);

                 } else if (processor instanceof DocumentWordProbability.Processor) {

                     this.processor = new DuplicateEliminator(new TupleUnshredder((DocumentWordProbability.Processor) processor));

                 } else if (processor instanceof org.galagosearch.tupleflow.Processor) {

                     this.processor = new DuplicateEliminator(new TupleUnshredder((org.galagosearch.tupleflow.Processor<DocumentWordProbability>) processor));

                 } else {

                     throw new IncompatibleProcessorException(processor.getClass().getName() + " is not supported by " + this.getClass().getName());                                                                       

                 }

             }                                

             public Class<DocumentWordProbability> getOutputClass() {

                 return DocumentWordProbability.class;

             }

             public void initialize() throws IOException {

                 for (ShreddedReader reader : readers) {

                     reader.fill();                                        

                     if (!reader.getBuffer().isAtEnd())

                         queue.add(reader);

                 }   

                 uninitialized = false;

             }

             public void run() throws IOException {

                 initialize();

                 while (queue.size() > 0) {

                     ShreddedReader top = queue.poll();

                     ShreddedReader next = null;

                     ShreddedBuffer nextBuffer = null; 

                     assert !top.getBuffer().isAtEnd();

                     if (queue.size() > 0) {

                         next = queue.peek();

                         nextBuffer = next.getBuffer();

                         assert !nextBuffer.isAtEnd();

                     }

                     top.getBuffer().copyUntil(nextBuffer, processor);

                     if (top.getBuffer().isAtEnd())

                         top.fill();                 

                     if (!top.getBuffer().isAtEnd())

                         queue.add(top);

                 }              

                 if (closeOnExit)

                     processor.close();

             }

             public DocumentWordProbability read() throws IOException {

                 if (uninitialized)

                     initialize();

                 DocumentWordProbability result = null;

                 while (queue.size() > 0) {

                     ShreddedReader top = queue.poll();

                     result = top.read();

                     if (result != null) {

                         if (top.getBuffer().isAtEnd())

                             top.fill();

                         queue.offer(top);

                         break;

                     } 

                 }

                 return result;

             }

         } 

         public static class ShreddedReader implements Step, Comparable<ShreddedReader>, TypeReader<DocumentWordProbability>, ShreddedSource {      

             public ShreddedProcessor processor;

             ShreddedBuffer buffer;

             DocumentWordProbability last = new DocumentWordProbability();         

             long updateDocumentCount = -1;

             long updateWordCount = -1;

             long tupleCount = 0;

             long bufferStartCount = 0;  

             ArrayInput input;

             public ShreddedReader(ArrayInput input) {

                 this.input = input; 

                 this.buffer = new ShreddedBuffer();

             }                               

             public ShreddedReader(ArrayInput input, int bufferSize) { 

                 this.input = input;

                 this.buffer = new ShreddedBuffer(bufferSize);

             }

             public final int compareTo(ShreddedReader other) {

                 ShreddedBuffer otherBuffer = other.getBuffer();

                 if (buffer.isAtEnd() && otherBuffer.isAtEnd()) {

                     return 0;                 

                 } else if (buffer.isAtEnd()) {

                     return -1;

                 } else if (otherBuffer.isAtEnd()) {

                     return 1;

                 }

                 int result = 0;

                 do {

                     result = + Utility.compare(buffer.getDocument(), otherBuffer.getDocument());

                     if(result != 0) break;

                     result = + Utility.compare(buffer.getWord(), otherBuffer.getWord());

                     if(result != 0) break;

                 } while (false);                                             

                 return result;

             }

             public final ShreddedBuffer getBuffer() {

                 return buffer;

             }                

             public final DocumentWordProbability read() throws IOException {

                 if (buffer.isAtEnd()) {

                     fill();             

                     if (buffer.isAtEnd()) {

                         return null;

                     }

                 }

                 assert !buffer.isAtEnd();

                 DocumentWordProbability result = new DocumentWordProbability();

                 result.document = buffer.getDocument();

                 result.word = buffer.getWord();

                 result.probability = buffer.getProbability();

                 buffer.incrementTuple();

                 buffer.autoIncrementDocument();

                 buffer.autoIncrementWord();

                 return result;

             }           

             public final void fill() throws IOException {

                 try {   

                     buffer.reset();

                     if (tupleCount != 0) {

                         if(updateDocumentCount - tupleCount > 0) {

                             buffer.documents.add(last.document);

                             buffer.documentTupleIdx.add((int) (updateDocumentCount - tupleCount));

                         }                              

                         if(updateWordCount - tupleCount > 0) {

                             buffer.words.add(last.word);

                             buffer.wordTupleIdx.add((int) (updateWordCount - tupleCount));

                         }

                         bufferStartCount = tupleCount;

                     }

                     while (!buffer.isFull()) {

                         updateWord();

                         buffer.processTuple(input.readDouble());

                         tupleCount++;

                     }

                 } catch(EOFException e) {}

             }

             public final void updateDocument() throws IOException {

                 if (updateDocumentCount > tupleCount)

                     return;

                 last.document = input.readString();

                 updateDocumentCount = tupleCount + input.readInt();

                 buffer.processDocument(last.document);

             }

             public final void updateWord() throws IOException {

                 if (updateWordCount > tupleCount)

                     return;

                 updateDocument();

                 last.word = input.readBytes();

                 updateWordCount = tupleCount + input.readInt();

                 buffer.processWord(last.word);

             }

             public void run() throws IOException {

                 while (true) {

                     fill();

                     if (buffer.isAtEnd())

                         break;

                     buffer.copyUntil(null, processor);

                 }      

                 processor.close();

             }

             public void setProcessor(Step processor) throws IncompatibleProcessorException {  

                 if (processor instanceof ShreddedProcessor) {

                     this.processor = new DuplicateEliminator((ShreddedProcessor) processor);

                 } else if (processor instanceof DocumentWordProbability.Processor) {

                     this.processor = new DuplicateEliminator(new TupleUnshredder((DocumentWordProbability.Processor) processor));

                 } else if (processor instanceof org.galagosearch.tupleflow.Processor) {

                     this.processor = new DuplicateEliminator(new TupleUnshredder((org.galagosearch.tupleflow.Processor<DocumentWordProbability>) processor));

                 } else {

                     throw new IncompatibleProcessorException(processor.getClass().getName() + " is not supported by " + this.getClass().getName());                                                                       

                 }

             }                                

             public Class<DocumentWordProbability> getOutputClass() {

                 return DocumentWordProbability.class;

             }                

         }

         public static class DuplicateEliminator implements ShreddedProcessor {

             public ShreddedProcessor processor;

             DocumentWordProbability last = new DocumentWordProbability();

             boolean documentProcess = true;

             boolean wordProcess = true;

             public DuplicateEliminator() {}

             public DuplicateEliminator(ShreddedProcessor processor) {

                 this.processor = processor;

             }

             public void setShreddedProcessor(ShreddedProcessor processor) {

                 this.processor = processor;

             }

             public void processDocument(String document) throws IOException {  

                 if (documentProcess || Utility.compare(document, last.document) != 0) {

                     last.document = document;

                     processor.processDocument(document);

             resetWord();

                     documentProcess = false;

                 }

             }

             public void processWord(byte[] word) throws IOException {  

                 if (wordProcess || Utility.compare(word, last.word) != 0) {

                     last.word = word;

                     processor.processWord(word);

                     wordProcess = false;

                 }

             }  

             public void resetDocument() {

                  documentProcess = true;

             resetWord();

             }                                                

             public void resetWord() {

                  wordProcess = true;

             }                                                

             public void processTuple(double probability) throws IOException {

                 processor.processTuple(probability);

             } 

             public void close() throws IOException {

                 processor.close();

             }                    

         }

         public static class TupleUnshredder implements ShreddedProcessor {

             DocumentWordProbability last = new DocumentWordProbability();

             public org.galagosearch.tupleflow.Processor<DocumentWordProbability> processor;                               

             public TupleUnshredder(DocumentWordProbability.Processor processor) {

                 this.processor = processor;

             }         

             public TupleUnshredder(org.galagosearch.tupleflow.Processor<DocumentWordProbability> processor) {

                 this.processor = processor;

             }

             public DocumentWordProbability clone(DocumentWordProbability object) {

                 DocumentWordProbability result = new DocumentWordProbability();

                 if (object == null) return result;

                 result.document = object.document; 

                 result.word = object.word; 

                 result.probability = object.probability; 

                 return result;

             }                 

             public void processDocument(String document) throws IOException {

                 last.document = document;

             }   

             public void processWord(byte[] word) throws IOException {

                 last.word = word;

             }   

             public void processTuple(double probability) throws IOException {

                 last.probability = probability;

                 processor.process(clone(last));

             }               

             public void close() throws IOException {

                 processor.close();

             }

         }     

         public static class TupleShredder implements Processor {

             DocumentWordProbability last = new DocumentWordProbability();

             public ShreddedProcessor processor;

             public TupleShredder(ShreddedProcessor processor) {

                 this.processor = processor;

             }                              

             public DocumentWordProbability clone(DocumentWordProbability object) {

                 DocumentWordProbability result = new DocumentWordProbability();

                 if (object == null) return result;

                 result.document = object.document; 

                 result.word = object.word; 

                 result.probability = object.probability; 

                 return result;

             }                 

             public void process(DocumentWordProbability object) throws IOException {                                                                                                                                                   

                 boolean processAll = false;

                 if(last == null || Utility.compare(last.document, object.document) != 0 || processAll) { processor.processDocument(object.document); processAll = true; }

                 if(last == null || Utility.compare(last.word, object.word) != 0 || processAll) { processor.processWord(object.word); processAll = true; }

                 processor.processTuple(object.probability);                                         

             }

             public Class<DocumentWordProbability> getInputClass() {

                 return DocumentWordProbability.class;

             }

             public void close() throws IOException {

                 processor.close();

             }                     

         }

     } 

     public static class WordOrder implements Order<DocumentWordProbability> {

         public int hash(DocumentWordProbability object) {

             int h = 0;

             h += Utility.hash(object.word);

             return h;

         } 

         public Comparator<DocumentWordProbability> greaterThan() {

             return new Comparator<DocumentWordProbability>() {

                 public int compare(DocumentWordProbability one, DocumentWordProbability two) {

                     int result = 0;

                     do {

                         result = + Utility.compare(one.word, two.word);

                         if(result != 0) break;

                     } while (false);

                     return -result;

                 }

             };

         }     

         public Comparator<DocumentWordProbability> lessThan() {

             return new Comparator<DocumentWordProbability>() {

                 public int compare(DocumentWordProbability one, DocumentWordProbability two) {

                     int result = 0;

                     do {

                         result = + Utility.compare(one.word, two.word);

                         if(result != 0) break;

                     } while (false);

                     return result;

                 }

             };

         }     

         public TypeReader<DocumentWordProbability> orderedReader(ArrayInput _input) {

             return new ShreddedReader(_input);

         }    

         public TypeReader<DocumentWordProbability> orderedReader(ArrayInput _input, int bufferSize) {

             return new ShreddedReader(_input, bufferSize);

         }    

         public OrderedWriter<DocumentWordProbability> orderedWriter(ArrayOutput _output) {

             ShreddedWriter w = new ShreddedWriter(_output);

             return new OrderedWriterClass(w); 

         }                                    

         public static class OrderedWriterClass extends OrderedWriter< DocumentWordProbability > {

             DocumentWordProbability last = null;

             ShreddedWriter shreddedWriter = null; 

             public OrderedWriterClass(ShreddedWriter s) {

                 this.shreddedWriter = s;

             }

             public void process(DocumentWordProbability object) throws IOException {

                boolean processAll = false;

                if (processAll || last == null || 0 != Utility.compare(object.word, last.word)) { processAll = true; shreddedWriter.processWord(object.word); }

                shreddedWriter.processTuple(object.document, object.probability);

                last = object;

             }           

             public void close() throws IOException {

                 shreddedWriter.close();

             }

             public Class<DocumentWordProbability> getInputClass() {

                 return DocumentWordProbability.class;

             }

         } 

         public ReaderSource<DocumentWordProbability> orderedCombiner(Collection<TypeReader<DocumentWordProbability>> readers, boolean closeOnExit) {

             ArrayList<ShreddedReader> shreddedReaders = new ArrayList();

             for (TypeReader<DocumentWordProbability> reader : readers) {

                 shreddedReaders.add((ShreddedReader)reader);

             }

             return new ShreddedCombiner(shreddedReaders, closeOnExit);

         }                  

         public DocumentWordProbability clone(DocumentWordProbability object) {

             DocumentWordProbability result = new DocumentWordProbability();

             if (object == null) return result;

             result.document = object.document; 

             result.word = object.word; 

             result.probability = object.probability; 

             return result;

         }                 

         public Class<DocumentWordProbability> getOrderedClass() {

             return DocumentWordProbability.class;

         }                           

         public String[] getOrderSpec() {

             return new String[] {"+word"};

         }

         public static String getSpecString() {

             return "+word";

         }

         public interface ShreddedProcessor extends Step {

             public void processWord(byte[] word) throws IOException;

             public void processTuple(String document, double probability) throws IOException;

             public void close() throws IOException;

         }    

         public interface ShreddedSource extends Step {

         }                                              

         public static class ShreddedWriter implements ShreddedProcessor {

             ArrayOutput output;

             ShreddedBuffer buffer = new ShreddedBuffer();

             byte[] lastWord;

             boolean lastFlush = false;

             public ShreddedWriter(ArrayOutput output) {

                 this.output = output;

             }                        

             public void close() throws IOException {

                 flush();

             }

             public void processWord(byte[] word) {

                 lastWord = word;

                 buffer.processWord(word);

             }

             public final void processTuple(String document, double probability) throws IOException {

                 if (lastFlush) {

                     if(buffer.words.size() == 0) buffer.processWord(lastWord);

                     lastFlush = false;

                 }

                 buffer.processTuple(document, probability);

                 if (buffer.isFull())

                     flush();

             }

             public final void flushTuples(int pauseIndex) throws IOException {

                 while (buffer.getReadIndex() < pauseIndex) {

                     output.writeString(buffer.getDocument());

                     output.writeDouble(buffer.getProbability());

                     buffer.incrementTuple();

                 }

             }  

             public final void flushWord(int pauseIndex) throws IOException {

                 while (buffer.getReadIndex() < pauseIndex) {

                     int nextPause = buffer.getWordEndIndex();

                     int count = nextPause - buffer.getReadIndex();

                     output.writeBytes(buffer.getWord());

                     output.writeInt(count);

                     buffer.incrementWord();

                     flushTuples(nextPause);

                     assert nextPause == buffer.getReadIndex();

                 }

             }

             public void flush() throws IOException { 

                 flushWord(buffer.getWriteIndex());

                 buffer.reset(); 

                 lastFlush = true;

             }                           

         }

         public static class ShreddedBuffer {

             ArrayList<byte[]> words = new ArrayList();

             ArrayList<Integer> wordTupleIdx = new ArrayList();

             int wordReadIdx = 0;

             String[] documents;

             double[] probabilitys;

             int writeTupleIndex = 0;

             int readTupleIndex = 0;

             int batchSize;

             public ShreddedBuffer(int batchSize) {

                 this.batchSize = batchSize;

                 documents = new String[batchSize];

                 probabilitys = new double[batchSize];

             }                              

             public ShreddedBuffer() {    

                 this(10000);

             }                                                                                                                    

             public void processWord(byte[] word) {

                 words.add(word);

                 wordTupleIdx.add(writeTupleIndex);

             }                                      

             public void processTuple(String document, double probability) {

                 assert words.size() > 0;

                 documents[writeTupleIndex] = document;

                 probabilitys[writeTupleIndex] = probability;

                 writeTupleIndex++;

             }

             public void resetData() {

                 words.clear();

                 wordTupleIdx.clear();

                 writeTupleIndex = 0;

             }                  

             public void resetRead() {

                 readTupleIndex = 0;

                 wordReadIdx = 0;

             } 

             public void reset() {

                 resetData();

                 resetRead();

             } 

             public boolean isFull() {

                 return writeTupleIndex >= batchSize;

             }

             public boolean isEmpty() {

                 return writeTupleIndex == 0;

             }                          

             public boolean isAtEnd() {

                 return readTupleIndex >= writeTupleIndex;

             }           

             public void incrementWord() {

                 wordReadIdx++;  

             }                                                                                              

             public void autoIncrementWord() {

                 while (readTupleIndex >= getWordEndIndex() && readTupleIndex < writeTupleIndex)

                     wordReadIdx++;

             }                 

             public void incrementTuple() {

                 readTupleIndex++;

             }                    

             public int getWordEndIndex() {

                 if ((wordReadIdx+1) >= wordTupleIdx.size())

                     return writeTupleIndex;

                 return wordTupleIdx.get(wordReadIdx+1);

             }

             public int getReadIndex() {

                 return readTupleIndex;

             }   

             public int getWriteIndex() {

                 return writeTupleIndex;

             } 

             public byte[] getWord() {

                 assert readTupleIndex < writeTupleIndex;

                 assert wordReadIdx < words.size();

                 return words.get(wordReadIdx);

             }

             public String getDocument() {

                 assert readTupleIndex < writeTupleIndex;

                 return documents[readTupleIndex];

             }                                         

             public double getProbability() {

                 assert readTupleIndex < writeTupleIndex;

                 return probabilitys[readTupleIndex];

             }                                         

             public void copyTuples(int endIndex, ShreddedProcessor output) throws IOException {

                 while (getReadIndex() < endIndex) {

                    output.processTuple(getDocument(), getProbability());

                    incrementTuple();

                 }

             }                                                                           

             public void copyUntilIndexWord(int endIndex, ShreddedProcessor output) throws IOException {

                 while (getReadIndex() < endIndex) {

                     output.processWord(getWord());

                     assert getWordEndIndex() <= endIndex;

                     copyTuples(getWordEndIndex(), output);

                     incrementWord();