Package cc.mallet.types

Examples of cc.mallet.types.TokenSequence

70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
    this (true, conjunctions);
  }
 
  public Instance pipe (Instance carrier)
  {
    TokenSequence ts = (TokenSequence) carrier.getData();
    int tsSize = ts.size();
    PropertyList[] oldfs = null;
    PropertyList[] newfs = null;
    try {
      oldfs = new PropertyList[ts.size()];
    }
    catch (Exception e) {
      System.err.println("Exception allocating oldfs: " + e);
    }
    try {
      newfs = new PropertyList[ts.size()];
    }
    catch (Exception e) {
      System.err.println("Exception allocating newfs: " + e);
    }
   
    for (int i = 0; i < tsSize; i++)
      oldfs[i] = ts.get(i).getFeatures ();
    if (includeOriginalSingletons)
      for (int i = 0; i < tsSize; i++)
        newfs[i] = ts.get(i).getFeatures ();

    for (int i = 0; i < tsSize; i++) {
      for (int j = 0; j < conjunctions.length; j++) {       
        // allow conjunction offsets of length n - awc
        PropertyList.Iterator[] iters = getOffsetIters (conjunctions, j, tsSize, i, oldfs);
        if (iters == null)
          continue;
        int[] iterIndices = new int[iters.length];
        for (int ii=0; ii < iterIndices.length; ii++)
          iterIndices[ii] = -1;
        newfs[i] = makeConjunctions (iters, 0, conjunctions, j, tsSize, newfs[i], i, oldfs, iterIndices);
      }
    }
    // Put the new PropertyLists in place
    for (int i = 0; i < ts.size(); i++)
      ts.get(i).setFeatures (newfs[i]);
    return carrier;
  }   
View Full Code Here
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
    noMtLst.addThruPipe (new ArrayIterator (doc1));

    Instance mtInst = mtLst.get (0);
    Instance noMtInst = noMtLst.get (0);

    TokenSequence mtTs = (TokenSequence) mtInst.getData ();
    TokenSequence noMtTs = (TokenSequence) noMtInst.getData ();

    assertEquals (6, mtTs.size ());
    assertEquals (6, noMtTs.size ());

    assertEquals (1.0, mtTs.get (3).getFeatureValue ("time"), 1e-15);
    assertEquals (1.0, noMtTs.get (3).getFeatureValue ("time"), 1e-15);
    assertEquals (1.0, mtTs.get (4).getFeatureValue ("time"), 1e-15);
    assertEquals (0.0, noMtTs.get (4).getFeatureValue ("time"), 1e-15);
  }
View Full Code Here
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
    this (true, conjunctions);
  }
 
  public Instance pipe (Instance carrier)
  {
    TokenSequence ts = (TokenSequence) carrier.getData();
    int tsSize = ts.size();
    PropertyList[] oldfs = new PropertyList[ts.size()];
    PropertyList[] newfs = new PropertyList[ts.size()];
    for (int i = 0; i < tsSize; i++)
      oldfs[i] = ts.get(i).getFeatures ();
    if (includeOriginalSingletons)
      for (int i = 0; i < tsSize; i++)
        newfs[i] = ts.get(i).getFeatures ();

    for (int i = 0; i < ts.size(); i++) {
      //System.out.println ("OffsetPropertyConjunctions: ts index="+i+", conjunction =");
      conjunctionList: for (int j = 0; j < conjunctions.length; j++) {
        // Make sure that the offsets in the conjunction are all available at this position
        for (int k = 0; k < conjunctions[j].length; k++) {
          if (conjunctions[j][k] + i < 0
              || conjunctions[j][k] + i > tsSize-1
              || oldfs[i+conjunctions[j][k]] == null)
            continue conjunctionList;
          //System.out.print (" "+conjunctions[j][k]);
        }
        //System.out.print ("\n");

        // Add the features for this conjunction
        if (conjunctions[j].length == 1) {
          int offset = conjunctions[j][0];
          if (offset == 0 && includeOriginalSingletons)
            throw new IllegalArgumentException ("Original singletons already there.");
          PropertyList.Iterator iter = oldfs[i+offset].iterator();
          while (iter.hasNext()) {
            iter.next();
            if (propertyKey != null && !propertyKey.equals(iter.getKey()))
              continue;
            String key = iter.getKey() + (offset==0 ? "" : "@"+offset);
            newfs[i] = PropertyList.add (key, iter.getNumericValue(), newfs[i]);
          }

        } else if (conjunctions[j].length == 2) {
          //System.out.println ("token="+ts.getToken(i).getText()+" conjunctionIndex="+j);
          int offset0 = conjunctions[j][0];
          int offset1 = conjunctions[j][1];
          PropertyList.Iterator iter0 = oldfs[i+offset0].iterator();
          int iter0i = -1;
          while (iter0.hasNext()) {
            iter0i++;
            iter0.next();
            if (propertyKey != null && !propertyKey.equals(iter0.getKey()))
              continue;
            PropertyList.Iterator iter1 = oldfs[i+offset1].iterator();
            int iter1i = -1;
            while (iter1.hasNext()) {
              iter1i++;
              iter1.next();
              if (propertyKey != null && !propertyKey.equals(iter1.getKey()))
                continue;
              // Avoid redundant doubling of feature space; include only upper triangle
              //System.out.println ("off0="+offset0+" off1="+offset1+" iter0i="+iter0i+" iter1i="+iter1i);
              if (offset0 == offset1 && iter1i <= iter0i) continue;
              //System.out.println (">off0="+offset0+" off1="+offset1+" iter0i="+iter0i+" iter1i="+iter1i);
              String key = iter0.getKey() + (offset0==0 ? "" : "@"+offset0)
                           +"&"+iter1.getKey() + (offset1==0 ? "" : "@"+offset1);
              newfs[i] = PropertyList.add (key, iter0.getNumericValue() * iter1.getNumericValue(), newfs[i]);
            }
          }

        } else if (conjunctions[j].length == 3) {
          int offset0 = conjunctions[j][0];
          int offset1 = conjunctions[j][1];
          int offset2 = conjunctions[j][2];
          PropertyList.Iterator iter0 = oldfs[i+offset0].iterator();
          int iter0i = -1;
          while (iter0.hasNext()) {
            iter0i++;
            iter0.next();
            if (propertyKey != null && !propertyKey.equals(iter0.getKey()))
              continue;
            PropertyList.Iterator iter1 = oldfs[i+offset1].iterator();
            int iter1i = -1;
            while (iter1.hasNext()) {
              iter1i++;
              iter1.next();
              if (propertyKey != null && !propertyKey.equals(iter1.getKey()))
                continue;
              // Avoid redundant doubling of feature space; include only upper triangle
              if (offset0 == offset1 && iter1i <= iter0i) continue;
              PropertyList.Iterator iter2 = oldfs[i+offset2].iterator();
              int iter2i = -1;
              while (iter2.hasNext()) {
                iter2i++;
                iter2.next();
                if (propertyKey != null && !propertyKey.equals(iter2.getKey()))
                  continue;
                // Avoid redundant doubling of feature space; include only upper triangle
                if (offset1 == offset2 && iter2i <= iter1i) continue;
                String key = iter0.getKey() + (offset0==0 ? "" : "@"+offset0)
                             +"&"+iter1.getKey() + (offset1==0 ? "" : "@"+offset1)
                             +"&"+iter2.getKey() + (offset2==0 ? "" : "@"+offset2);
                newfs[i] = PropertyList.add (key, iter0.getNumericValue() * iter1.getNumericValue()
                                             * iter2.getNumericValue(), newfs[i]);
              }
            }
          }
        } else {
          throw new UnsupportedOperationException ("Conjunctions of length 4 or more not yet implemented.");
        }
      }
    }

    // Put the new PropertyLists in place
    for (int i = 0; i < ts.size(); i++)
      ts.get(i).setFeatures (newfs[i]);
    return carrier;
  }
View Full Code Here
101
102
103
104
105
106
107
108
109
110
    Pipe mtPipe = (Pipe) TestSerializable.cloneViaSerialization (origPipe);
    InstanceList mtLst = new InstanceList (mtPipe);
    mtLst.addThruPipe (new ArrayIterator (doc1));
    Instance mtInst = mtLst.get (0);
    TokenSequence mtTs = (TokenSequence) mtInst.getData ();
    assertEquals (6, mtTs.size ());
    assertEquals (1.0, mtTs.get (3).getFeatureValue ("time"), 1e-15);
    assertEquals (1.0, mtTs.get (4).getFeatureValue ("time"), 1e-15);
  }
View Full Code Here
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
    return isNonNegated;
  }

  public Instance pipe (Instance carrier)
  {
    TokenSequence ts = (TokenSequence) carrier.getData();
    int tsSize = ts.size();
    for (int t = 0; t < tsSize; t++) {
      // Check whether the conjunction is true at time step t
      boolean passes = true;
      for (int fnum = 0; fnum < featurePatterns.length; fnum++) {
        int pos = t + offsets[fnum];
        if (!(pos >= 0 && pos < tsSize)) {
          passes = false;
          break;
        }
        boolean featurePresent = hasMatchingFeature (ts.get(pos), featurePatterns [fnum]);
        if (featurePresent != isNonNegated [fnum]) {
          passes = false;
          break;
        }
      }
      if (passes) {
        if (tagAllTimesteps) {
          for (int fnum = 0; fnum < featurePatterns.length; fnum++) {
            int pos = t + offsets[fnum];
            ts.get(pos).setFeatureValue (thisFeatureName, 1.0);
          }
        } else {
          ts.get(t).setFeatureValue (thisFeatureName, 1.0);
        }
      }
    }

    return carrier;
View Full Code Here
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
    this (namePrefix, null);
  }
 
  public Instance pipe (Instance carrier)
  {
    TokenSequence ts = (TokenSequence) carrier.getData();
    int tsSize = ts.size();
    for (int i = tsSize-1; i >= 0; i--) {
      Token t = ts.get (i);
      String text = t.getText();
      if (featureRegex != null && !featureRegex.matcher(text).matches())
        continue;
      for (int j = 0; j < i; j++) {
        if (ts.get(j).getText().equals(text)) {
          PropertyList.Iterator iter = ts.get(j).getFeatures().iterator();
          while (iter.hasNext()) {
            iter.next();
            String key = iter.getKey();
            if (filterRegex == null || (filterRegex.matcher(key).matches() ^ !includeFiltered))
              t.setFeatureValue (namePrefix+key, iter.getNumericValue());
View Full Code Here
133
134
135
136
137
138
139
140
141
142
    URI uri = null;
    try { uri = new URI ("random:" + classNames[currentClassIndex] + "/" + currentInstanceIndex); }
    catch (Exception e) {e.printStackTrace(); throw new IllegalStateException (); }
    //xxx Producing small numbers? int randomSize = r.nextPoisson (featureVectorSizePoissonLambda);
    int randomSize = (int)featureVectorSizePoissonLambda;
    TokenSequence ts = classCentroid[currentClassIndex].randomTokenSequence (r, randomSize);
    //logger.fine ("FeatureVector "+currentClassIndex+" "+currentInstanceIndex); fv.print();
    currentInstanceIndex--;
    return new Instance (ts, classNames[currentClassIndex], uri, null);
  }
View Full Code Here
85
86
87
88
89
90
91
92
93
94
95
96
97
    this (_realValued, false, "=");
  }

 
  public Instance pipe (Instance carrier) {
    TokenSequence ts = (TokenSequence) carrier.getData ();
    for (int i=0; i < ts.size(); i++) {
      Token t = ts.get (i);
      String[] values = t.getText().split("\\s+");
      for (int j=0; j < values.length; j++) {
        if (specifyFeatureNames) {
          String[] nameAndValue = values[j].split(nameValueSeparator);           
          if (nameAndValue.length != 2) { // no feature name. use token as feature.
View Full Code Here
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
    this (Pattern.compile (regex), dataGroup, targetGroup);
  }

  public Instance pipe (Instance carrier)
  {
    TokenSequence ts = (TokenSequence) carrier.getData();
    TokenSequence targetTokenSeq = new TokenSequence (ts.size());
    for (int i = 0; i < ts.size(); i++) {
      Token t = ts.get(i);
      Matcher matcher = regex.matcher (t.getText());
      if (matcher.matches()) {
        targetTokenSeq.add (matcher.group(targetGroup));
        t.setText (matcher.group (dataGroup));
      } else {
        logger.warning ("Skipping token: No match of "+regex.pattern()
                        +" at token #"+i+" with text "+t.getText());
      }
View Full Code Here
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
    this (false);
  }

  public Instance pipe (Instance carrier)
  {
    TokenSequence ts = (TokenSequence) carrier.getData();
    // xxx This doesn't seem so efficient.  Perhaps have TokenSequence
    // use a LinkedList, and remove Tokens from it? -?
    // But a LinkedList implementation of TokenSequence would be quite inefficient -AKM
    TokenSequence ret = new TokenSequence ();
    Token prevToken = null;
    for (int i = 0; i < ts.size(); i++) {
      Token t = ts.get(i);
      String s = t.getText();
      if (CharSequenceLexer.LEX_ALPHA.matcher(s).matches()) {
        ret.add (t);
        prevToken = t;
      }  else if (markDeletions && prevToken != null)
        prevToken.setProperty (FeatureSequenceWithBigrams.deletionMark, t.getText());
    }
    carrier.setData(ret);
View Full Code Here
TOP

Related Classes of cc.mallet.types.TokenSequence

  • cc.mallet.extract.pipe.TokenSequence2Tokenization
  • cc.mallet.fst.MultiSegmentationEvaluator
  • cc.mallet.fst.SegmentationEvaluator
  • cc.mallet.fst.tests.TestCRF$TestCRFTokenSequenceRemoveSpaces
  • cc.mallet.fst.ViterbiWriter
  • cc.mallet.pipe.CharSequence2CharNGrams
  • cc.mallet.pipe.CharSequence2TokenSequence
  • cc.mallet.pipe.CharSequenceArray2TokenSequence
  • cc.mallet.pipe.iterator.RandomTokenSequenceIterator
  • cc.mallet.pipe.LineGroupString2TokenSequence
    • Copyright © 2018 www.massapicom. All rights reserved.
    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.