Examples of cc.mallet.pipe.Pipe

cc.mallet.pipe.Pipe
The abstract superclass of all Pipes, which transform one data type to another. Pipes are most often used for feature extraction.
Although Pipe does not have any "abstract methods", in order to use a Pipe subclass you must override either the {@link pipe} method or the {@link newIteratorFrom} method.The former is appropriate when the pipe's processing of an Instance is strictly one-to-one. For every Instance coming in, there is exactly one Instance coming out. The later is appropriate when the pipe's processing may result in more or fewer Instances than arrive through its source iterator.
A pipe operates on an {@link cc.mallet.types.Instance}, which is a carrier of data. A pipe reads from and writes to fields in the Instance when it is requested to process the instance. It is up to the pipe which fields in the Instance it reads from and writes to, but usually a pipe will read its input from and write its output to the "data" field of an instance.
A pipe doesn't have any direct notion of input or output - it merely modifies instances that are handed to it. A set of helper classes, which implement the interface {@link Iterator}, iterate over commonly encountered input data structures and feed the elements of these data structures to a pipe as instances.
A pipe is frequently used in conjunction with an {@link cc.mallet.types.InstanceList} As instances are addedto the list, they are processed by the pipe associated with the instance list and the processed Instance is kept in the list.
In one common usage, a {@link cc.mallet.pipe.iterator.FileIterator} is given a list of directories to operate over.The FileIterator walks through each directory, creating an instance for each file and putting the data from the file in the data field of the instance. The directory of the file is stored in the target field of the instance. The FileIterator feeds instances to an InstanceList, which processes the instances through its associated pipe and keeps the results.
Pipes can be hierachically composed. In a typical usage, a SerialPipe is created, which holds other pipes in an ordered list. Piping an instance through a SerialPipe means piping the instance through each of the child pipes in sequence.
A pipe holds two separate Alphabets: one for the symbols (feature names) encountered in the data fields of the instances processed through the pipe, and one for the symbols (e.g. class labels) encountered in the target fields.
@author Andrew McCallum mccallum@cs.umass.edu

    TestOptimizable.testGetSetParameters(omemm);
  }


  public void testSpaceMaximizable ()
  {
    Pipe p = makeSpacePredictionPipe ();
    InstanceList training = new InstanceList (p);
//    String[] data = { TestMEMM.data[0], }; // TestMEMM.data[1], TestMEMM.data[2], TestMEMM.data[3], };
//    String[] data = { "ab" };
    training.addThruPipe (new ArrayIterator (data));

View Full Code Here

    TestOptimizable.testValueAndGradient (mcrf);
  }


  public void testSpaceSerializable () throws IOException, ClassNotFoundException
  {
    Pipe p = makeSpacePredictionPipe ();
    InstanceList training = new InstanceList (p);
    training.addThruPipe (new ArrayIterator (data));


    MEMM memm = new MEMM (p, null);
    memm.addFullyConnectedStatesForLabels ();

View Full Code Here

  }




  public void doTestSpacePrediction(boolean testValueAndGradient)
  {
    Pipe p = makeSpacePredictionPipe ();
    Pipe p2 = new TestMEMM2String();


    InstanceList instances = new InstanceList(p);
    instances.addThruPipe(new ArrayIterator(data));
    InstanceList[] lists = instances.split(new double[]{.5, .5});
    MEMM memm = new MEMM(p, p2);

View Full Code Here


  public void doTestSpacePrediction(boolean testValueAndGradient,
                                    boolean useSaved,
                                    boolean useSparseWeights)
  {
    Pipe p = makeSpacePredictionPipe ();


    MEMM savedCRF;
    File f = new File("TestObject.obj");
    InstanceList instances = new InstanceList(p);
    instances.addThruPipe(new ArrayIterator(data));
    InstanceList[] lists = instances.split(new double[]{.5, .5});
    MEMM crf = new MEMM(p.getDataAlphabet(), p.getTargetAlphabet());
    crf.addFullyConnectedStatesForLabels();
    if (useSparseWeights)
      crf.setWeightsDimensionAsIn(lists[0]);
    else
      crf.setWeightsDimensionDensely();

View Full Code Here

  }




  public static  Pipe makeSpacePredictionPipe ()
  {
    Pipe p = new SerialPipes(new Pipe[]{
      new CharSequence2TokenSequence("."),
      new TokenSequenceLowercase(),
      new TestMEMMTokenSequenceRemoveSpaces(),
      new TokenText(),
      new OffsetConjunctions(true,

View Full Code Here

  }




  public void disabledtestAddOrderNStates ()
  {
    Pipe p = makeSpacePredictionPipe ();


    InstanceList instances = new InstanceList (p);
    instances.addThruPipe (new ArrayIterator(data));
    InstanceList[] lists = instances.split (new java.util.Random (678), new double[]{.5, .5});


    // Compare 3 CRFs trained with addOrderNStates, and make sure
    // that having more features leads to a higher likelihood


    MEMM crf1 = new MEMM(p.getDataAlphabet(), p.getTargetAlphabet());
    crf1.addOrderNStates (lists [0],
                         new int[] { 1, },
                         new boolean[] { false, },
                         "START",
                         null,
                         null,
                         false);
    crf1.setWeightsDimensionAsIn(lists[0]);
    MEMMTrainer memmt1 = new MEMMTrainer (crf1);
    memmt1.train(lists [0]);




    MEMM crf2 = new MEMM(p.getDataAlphabet(), p.getTargetAlphabet());
    crf2.addOrderNStates (lists [0],
                           new int[] { 1, 2, },
                           new boolean[] { false, true },
                           "START",
                           null,
                           null,
                           false);
    crf2.setWeightsDimensionAsIn(lists[0]);
    MEMMTrainer memmt2 = new MEMMTrainer (crf2);
    memmt2.train(lists [0]);




    MEMM crf3 = new MEMM(p.getDataAlphabet(), p.getTargetAlphabet());
    crf3.addOrderNStates (lists [0],
                         new int[] { 1, 2, },
                         new boolean[] { false, false },
                         "START",
                         null,

View Full Code Here

    doTestSpacePrediction(false, true, false);
  }


  public void disabledtestPrint ()
  {
    Pipe p = new SerialPipes (new Pipe[] {
       new CharSequence2TokenSequence("."),
       new TokenText(),
       new TestMEMM.TestMEMMTokenSequenceRemoveSpaces(),
       new TokenSequence2FeatureVectorSequence(),
       new PrintInputAndTarget(),

View Full Code Here

        constraintsFile = new FileReader(new File(args[restArgs+1]));
      }
    } else 
      testFile = new FileReader(new File(args[restArgs]));


    Pipe p = null;
    CRF crf = null;
    TransducerEvaluator eval = null;
    if (continueTrainingOption.value || !trainOption.value) {
      if (modelOption.value == null)
      {
        commandOptions.printUsage(true);
        throw new IllegalArgumentException("Missing model file option");
      }
      ObjectInputStream s =
        new ObjectInputStream(new FileInputStream(modelOption.value));
      crf = (CRF) s.readObject();
      s.close();
      p = crf.getInputPipe();
    }
    else {
      p = new SimpleTaggerSentence2FeatureVectorSequence();
      p.getTargetAlphabet().lookupIndex(defaultOption.value);
    }




    if (trainOption.value)
    {
      p.setTargetProcessing(true);
      trainingData = new InstanceList(p);
      trainingData.addThruPipe(
          new LineGroupIterator(trainingFile,
            Pattern.compile("^\\s*$"), true));
      logger.info
        ("Number of features in training data: "+p.getDataAlphabet().size());
      if (testOption.value != null)
      {
        if (testFile != null)
        {
          testData = new InstanceList(p);
          testData.addThruPipe(
              new LineGroupIterator(testFile,
                Pattern.compile("^\\s*$"), true));
        } else
        {
          Random r = new Random (randomSeedOption.value);
          InstanceList[] trainingLists =
            trainingData.split(
                r, new double[] {trainingFractionOption.value,
                  1-trainingFractionOption.value});
          trainingData = trainingLists[0];
          testData = trainingLists[1];
        }
      }
    } else if (testOption.value != null)
    {
      p.setTargetProcessing(true);
      testData = new InstanceList(p);
      testData.addThruPipe(
          new LineGroupIterator(testFile,
            Pattern.compile("^\\s*$"), true));
    } else
    {
      p.setTargetProcessing(false);
      testData = new InstanceList(p);
      testData.addThruPipe(
          new LineGroupIterator(testFile,
            Pattern.compile("^\\s*$"), true));
    }
    logger.info ("Number of predicates: "+p.getDataAlphabet().size());
    
    
    if (testOption.value != null)
    {
      if (testOption.value.startsWith("lab"))
        eval = new TokenAccuracyEvaluator(new InstanceList[] {trainingData, testData}, new String[] {"Training", "Testing"});
      else if (testOption.value.startsWith("seg="))
      {
        String[] pairs = testOption.value.substring(4).split(",");
        if (pairs.length < 1)
        {
          commandOptions.printUsage(true);
          throw new IllegalArgumentException(
              "Missing segment start/continue labels: " + testOption.value);
        }
        String startTags[] = new String[pairs.length];
        String continueTags[] = new String[pairs.length];
        for (int i = 0; i < pairs.length; i++)
        {
          String[] pair = pairs[i].split("\\.");
          if (pair.length != 2)
          {
            commandOptions.printUsage(true);
            throw new
              IllegalArgumentException(
                  "Incorrectly-specified segment start and end labels: " +
                  pairs[i]);
          }
          startTags[i] = pair[0];
          continueTags[i] = pair[1];
        }
        eval = new MultiSegmentationEvaluator(new InstanceList[] {trainingData, testData}, new String[] {"Training", "Testing"}, 
            startTags, continueTags);
      }
      else
      {
        commandOptions.printUsage(true);
        throw new IllegalArgumentException("Invalid test option: " +
            testOption.value);
      }
    }
    
    
    
    if (p.isTargetProcessing())
    {
      Alphabet targets = p.getTargetAlphabet();
      StringBuffer buf = new StringBuffer("Labels:");
      for (int i = 0; i < targets.size(); i++)
        buf.append(" ").append(targets.lookupObject(i).toString());
      logger.info(buf.toString());
    }

View Full Code Here

    }
    else {
      testFile = new FileReader(new File(args[restArgs]));
    }


    Pipe p = null;
    CRF crf = null;
    TransducerEvaluator eval = null;


    if (continueTrainingOption.value || !trainOption.value) {
      if (modelOption.value == null) {
        commandOptions.printUsage(true);
        throw new IllegalArgumentException("Missing model file option");
      }
      ObjectInputStream s =
        new ObjectInputStream(new FileInputStream(modelOption.value));
      crf = (CRF) s.readObject();
      s.close();
      p = crf.getInputPipe();
    }
    else {
      p = new SimpleTaggerSentence2FeatureVectorSequence();
      p.getTargetAlphabet().lookupIndex(defaultOption.value);
    }




    if (trainOption.value) {
      p.setTargetProcessing(true);
      trainingData = new InstanceList(p);
      trainingData.addThruPipe(new LineGroupIterator(trainingFile,
                               Pattern.compile("^\\s*$"), true));


      logger.info("Number of features in training data: "+p.getDataAlphabet().size());


      if (testOption.value != null) {
        if (testFile != null) {
          testData = new InstanceList(p);
          testData.addThruPipe(new LineGroupIterator(testFile,
                                 Pattern.compile("^\\s*$"), true));
        }
        else {
          Random r = new Random (randomSeedOption.value);
          InstanceList[] trainingLists =
            trainingData.split(r, new double[] {trainingFractionOption.value,
                              1 - trainingFractionOption.value});
          trainingData = trainingLists[0];
          testData = trainingLists[1];
        }
      }
    }
    else if (testOption.value != null) {
      p.setTargetProcessing(true);
      testData = new InstanceList(p);
      testData.addThruPipe(new LineGroupIterator(testFile,
                             Pattern.compile("^\\s*$"), true));
    }
    else {
        p.setTargetProcessing(false);
        testData = new InstanceList(p);
        testData.addThruPipe(
                   new LineGroupIterator(testFile,
                               Pattern.compile("^\\s*$"), true));
    }
    logger.info ("Number of predicates: "+p.getDataAlphabet().size());
    
    
    if (testOption.value != null) {
      if (testOption.value.startsWith("lab")) {
          eval = new TokenAccuracyEvaluator(new InstanceList[] {trainingData, testData}, new String[] {"Training", "Testing"});
      }
      else if (testOption.value.startsWith("seg=")) {
        String[] pairs = testOption.value.substring(4).split(",");
        if (pairs.length < 1) {
          commandOptions.printUsage(true);
          throw new IllegalArgumentException
            ("Missing segment start/continue labels: " + testOption.value);
        }
        String startTags[] = new String[pairs.length];
        String continueTags[] = new String[pairs.length];


        for (int i = 0; i < pairs.length; i++) {
          String[] pair = pairs[i].split("\\.");
          if (pair.length != 2) {
            commandOptions.printUsage(true);
            throw new
              IllegalArgumentException
              ("Incorrectly-specified segment start and end labels: " + pairs[i]);
          }
          startTags[i] = pair[0];
          continueTags[i] = pair[1];
        }
        eval = new MultiSegmentationEvaluator(new InstanceList[] {trainingData, testData}, new String[] {"Training", "Testing"}, 
                            startTags, continueTags);
      }
      else {
          commandOptions.printUsage(true);
          throw new IllegalArgumentException("Invalid test option: " +
                             testOption.value);
      }
    }




    
    if (p.isTargetProcessing()) {
      Alphabet targets = p.getTargetAlphabet();
      StringBuffer buf = new StringBuffer("Labels:");
      for (int i = 0; i < targets.size(); i++)
        buf.append(" ").append(targets.lookupObject(i).toString());
      logger.info(buf.toString());
    }

View Full Code Here

  {
    return pipes.size();
  }


  public Pipe getPipe (int index) {
    Pipe retPipe = null;
    try {
      retPipe = pipes.get(index);
    }
    catch (Exception e) {
      System.err.println("Error getting pipe. Index = " + index + ".  " + e.getMessage());

View Full Code Here

0 1 2 3 4 5

TOP

Related Classes of cc.mallet.pipe.Pipe

cc.mallet.classify.tui.SvmLight2Classify

cc.mallet.classify.WinnowTrainer

cc.mallet.cluster.examples.FirstOrderClusterExample

cc.mallet.cluster.tui.Clusterings2Clusterer

cc.mallet.cluster.tui.Clusterings2Clusterings

cc.mallet.extract.CRFExtractor

cc.mallet.extract.test.TestDocumentViewer

cc.mallet.extract.test.TestLatticeViewer

cc.mallet.fst.semi_supervised.tui.SimpleTaggerWithConstraints

cc.mallet.fst.SimpleTagger

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.