Package cc.mallet.types

Examples of cc.mallet.types.Token

112
113
114
115
116
117
118
119
120
121
122
        while (lexer.hasNext()) {
          lexer.next ();
          int tokStart = textStart + lexer.getStartOffset ();
          int tokEnd = textStart + lexer.getEndOffset ();
          dataTokens.add (new StringSpan (string, tokStart, tokEnd));
          targetTokens.add (new Token (tag));
        }
      }
      textStart = nextStart;
      tag = nextTag;
    }
View Full Code Here
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
      String header = ts.get(1).getText();
      if (header.equals("PRESS"))        // Don't bother with "PRESS DIGEST" headers
        return carrier;
      String featureName = "HEADER="+header;
      for (int i = 0; i < ts.size(); i++) {
        Token t = ts.get(i);
        // Only apply this feature to capitalized words, because if we apply it to everything
        // we easily get an immense number of possible feature conjunctions, (e.g. every word
        // with each of these HEADER= features.
        if (t.getText().matches("^[A-Z].*"))
          t.setFeatureValue (featureName, 1.0);
      }
    }
    return carrier;
  }
View Full Code Here
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
      StringBuffer buf = new StringBuffer ();
      TokenSequence ts = (TokenSequence) data;
      StringTokenization spans = new StringTokenization (buf);  // I can use a StringBuffer as the doc! Awesome!

      for (int i = 0; i < ts.size(); i++) {
        Token token = ts.get(i);

        int start = buf.length ();
        buf.append (token.getText());
        int end = buf.length();

        StringSpan span = new StringSpan (buf, start, end);
        span.setFeatures (token.getFeatures ());
        span.setProperties (token.getProperties ());

        spans.add (span);
        buf.append (" ");
      }
View Full Code Here
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
    String newTerm = null;
    TokenSequence tmpTS = new TokenSequence();
    TokenSequence ts = (TokenSequence) carrier.getData();

    for (int i = 0; i < ts.size(); i++) {
      Token t = ts.get(i);
      for(int j = 0; j < gramSizes.length; j++) {
        int len = gramSizes[j];
        if (len <= 0 || len > (i+1)) continue;
        if (len == 1) { tmpTS.add(t); continue; }
        newTerm = new String(t.getText());
        for(int k = 1; k < len; k++)
          newTerm = ts.get(i-k).getText() + "_" + newTerm;
        tmpTS.add(newTerm);
      }
    }
View Full Code Here
96
97
98
99
100
101
102
103
104
105
106
107
        textEnd = m.start();
      }
      if (textEnd - textStart > 0) {
        lexer.setCharSequence (string.subSequence (textStart, textEnd));
        while (lexer.hasNext()) {
          dataTokens.add (new Token ((String) lexer.next()));
          targetTokens.add (new Token (tag));
        }
      }
      textStart = nextStart;
      tag = nextTag;
    }
View Full Code Here
99
100
101
102
103
104
105
106
107
108
109
110
  {
    TokenSequence ts = (TokenSequence) carrier.getData();
    int tsSize = ts.size();
    PropertyList[] newFeatures = new PropertyList[tsSize];
    for (int i = 0; i < tsSize; i++) {
      Token t = ts.get (i);
      PropertyList pl = t.getFeatures();
      newFeatures[i] = pl;
      for (int position = i + leftBoundary; position < i + rightBoundary; position++) {
        if (position == i && !includeCurrentToken)
          continue;
        PropertyList pl2;
View Full Code Here
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
  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());
          }
          break;
        }
        if (firstMentionName != null)
          t.setFeatureValue (firstMentionName, 1.0);
      }
    }
    return carrier;
  }
View Full Code Here
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
 
  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.
            t.setFeatureValue ("Token="+values[j], 1.0);
          }
          else {
            t.setFeatureValue (nameAndValue[0], Double.parseDouble (nameAndValue[1]));           
          }
        }
        else if (realValued) {
          t.setFeatureValue ("Feature#" + j, Double.parseDouble (values[j]));
        }
        else
          t.setFeatureValue (values[j], 1.0);         
      }
    }
    carrier.setData (ts);
    return carrier;
  }
View Full Code Here
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
  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());
      }
    }
    carrier.setTarget(targetTokenSeq);
    carrier.setData(ts);
    return carrier;
View Full Code Here
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
    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);
    return carrier;
  }
View Full Code Here
TOP

Related Classes of cc.mallet.types.Token

  • cc.mallet.extract.pipe.TokenSequence2Tokenization
  • cc.mallet.fst.tests.TestCRF$TestCRFTokenSequenceRemoveSpaces
  • cc.mallet.pipe.SelectiveSGML2TokenSequence
  • cc.mallet.pipe.SGML2TokenSequence
  • cc.mallet.pipe.SourceLocation2TokenSequence
  • cc.mallet.pipe.TokenSequenceLowercase
  • cc.mallet.pipe.TokenSequenceMatchDataAndTarget
  • cc.mallet.pipe.TokenSequenceNGrams
  • cc.mallet.pipe.TokenSequenceParseFeatureString
  • cc.mallet.pipe.TokenSequenceRemoveNonAlpha
    • 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.