Package org.apache.lucene.analysis.tokenattributes

Examples of org.apache.lucene.analysis.tokenattributes.CharTermAttribute

185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
  }

  private void assertURLDecodedTo(final Tokenizer t, final String encoding, final String uri, final String[] expectedStems, final String[] expectedTypes, final int[] expectedPosIncr)
  throws IOException {
    assertTrue("has CharTermAttribute", t.hasAttribute(CharTermAttribute.class));
    final CharTermAttribute termAtt = t.getAttribute(CharTermAttribute.class);

    assertTrue("has TypeAttribute", t.hasAttribute(TypeAttribute.class));
    final TypeAttribute typeAtt = t.getAttribute(TypeAttribute.class);

    assertTrue("has PositionIncrementAttribute", t.hasAttribute(PositionIncrementAttribute.class));
    final PositionIncrementAttribute posIncrAtt = t.getAttribute(PositionIncrementAttribute.class);

    t.setReader(new StringReader(uri));
    t.reset();

    final URIDecodingFilter filter = new URIDecodingFilter(t, encoding);
    for (int i = 0; i < expectedStems.length; i++) {
        assertTrue("token " + i + " exists", filter.incrementToken());
        assertEquals(expectedStems[i], termAtt.toString());
        if (expectedTypes == null)
          assertEquals(uritype, typeAtt.type());
        else
          assertEquals(expectedTypes[i], typeAtt.type());
        if (expectedPosIncr != null)
View Full Code Here
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
                                 final String[] expectedImages,
                                 final String[] expectedTypes)
   throws Exception {

     assertTrue("has CharTermAttribute", t.hasAttribute(CharTermAttribute.class));
     final CharTermAttribute termAtt = t.getAttribute(CharTermAttribute.class);

     TypeAttribute typeAtt = null;
     if (expectedTypes != null) {
       assertTrue("has TypeAttribute", t.hasAttribute(TypeAttribute.class));
       typeAtt = t.getAttribute(TypeAttribute.class);
     }

     t.setReader(new StringReader(input));
     t.reset();

     final TokenStream filter = new URINormalisationFilter(t);

     for (int i = 0; i < expectedImages.length; i++) {

       assertTrue("token "+i+" exists", filter.incrementToken());

       assertEquals(expectedImages[i], termAtt.toString());

       if (expectedTypes != null) {
         assertEquals(expectedTypes[i], typeAtt.type());
       }
View Full Code Here
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
  }

  private void assertURLDecodedTo(final Tokenizer t, final String uri, final String[] expectedStems, final String[] expectedTypes, final int[] expectedPosIncr)
  throws IOException {
    assertTrue("has CharTermAttribute", t.hasAttribute(CharTermAttribute.class));
    final CharTermAttribute termAtt = t.getAttribute(CharTermAttribute.class);

    assertTrue("has TypeAttribute", t.hasAttribute(TypeAttribute.class));
    final TypeAttribute typeAtt = t.getAttribute(TypeAttribute.class);

    assertTrue("has PositionIncrementAttribute", t.hasAttribute(PositionIncrementAttribute.class));
    final PositionIncrementAttribute posIncrAtt = t.getAttribute(PositionIncrementAttribute.class);

    t.setReader(new StringReader(uri));
    t.reset();

    final TokenFilter filter = new MailtoFilter(t);
    for (int i = 0; i < expectedStems.length; i++) {
        assertTrue("token " + i + " exists", filter.incrementToken());
        assertEquals(expectedStems[i], termAtt.toString());
        if (expectedTypes == null)
          assertEquals(uritype, typeAtt.type());
        else
          assertEquals(expectedTypes[i], typeAtt.type());
        if (expectedPosIncr != null)
View Full Code Here
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
                                final String[] expectedTypes,
                                final int[] expectedPosIncrs)
  throws Exception {

    assertTrue("has TermAttribute", t.hasAttribute(CharTermAttribute.class));
    final CharTermAttribute termAtt = t.getAttribute(CharTermAttribute.class);

    TypeAttribute typeAtt = null;
    if (expectedTypes != null) {
      assertTrue("has TypeAttribute", t.hasAttribute(TypeAttribute.class));
      typeAtt = t.getAttribute(TypeAttribute.class);
    }

    PositionIncrementAttribute posIncrAtt = null;
    if (expectedPosIncrs != null) {
      assertTrue("has PositionIncrementAttribute", t.hasAttribute(PositionIncrementAttribute.class));
      posIncrAtt = t.getAttribute(PositionIncrementAttribute.class);
    }

    t.setReader(new StringReader(input));
    t.reset();

    final URILocalnameFilter filter = new URILocalnameFilter(t);
    filter.setMaxLength(MAX_LENGTH);

    for (int i = 0; i < expectedImages.length; i++) {

      assertTrue("token "+i+" exists", filter.incrementToken());

      assertEquals(expectedImages[i], termAtt.toString());

      if (expectedTypes != null) {
        assertEquals(expectedTypes[i], typeAtt.type());
      }
View Full Code Here
37
38
39
40
41
42
43
44
45
46
47
  public void testTokenTypeFilter1() throws Exception {
    final Reader reader = new StringReader("aaa clés café");
    final TokenStream stream = new WhitespaceTokenizer(TEST_VERSION_CURRENT, reader);
    final ASCIIFoldingExpansionFilter filter = new ASCIIFoldingExpansionFilter(stream);

    final CharTermAttribute termAtt = filter.getAttribute(CharTermAttribute.class);
    final PositionIncrementAttribute posAtt = filter.getAttribute(PositionIncrementAttribute.class);

    filter.reset(); // prepare stream

    this.assertTermEquals("aaa", 1, filter, termAtt, posAtt);
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
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
  throws Exception {
    final TokenStream t = a.tokenStream("", new StringReader(input));
    t.reset();

    assertTrue("has TermAttribute", t.hasAttribute(CharTermAttribute.class));
    final CharTermAttribute termAtt = t.getAttribute(CharTermAttribute.class);

    TypeAttribute typeAtt = null;
    if (expectedTypes != null) {
      assertTrue("has TypeAttribute", t.hasAttribute(TypeAttribute.class));
      typeAtt = t.getAttribute(TypeAttribute.class);
    }

    PositionIncrementAttribute posIncrAtt = null;
    if (expectedPosIncrs != null) {
      assertTrue("has PositionIncrementAttribute", t.hasAttribute(PositionIncrementAttribute.class));
      posIncrAtt = t.getAttribute(PositionIncrementAttribute.class);
    }

    NodeAttribute nodeAtt = null;
    if (expectedNode != null) {
      assertTrue("has NodeAttribute", t.hasAttribute(NodeAttribute.class));
      nodeAtt = t.getAttribute(NodeAttribute.class);
    }

    PositionAttribute posAtt = null;
    if (expectedPos != null) {
      assertTrue("has PositionAttribute", t.hasAttribute(PositionAttribute.class));
      posAtt = t.getAttribute(PositionAttribute.class);
    }

    for (int i = 0; i < expectedImages.length; i++) {

      assertTrue("token "+i+" exists", t.incrementToken());

      assertEquals("i=" + i, expectedImages[i], termAtt.toString());

      if (expectedTypes != null) {
        assertEquals(expectedTypes[i], typeAtt.type());
      }

      if (expectedPosIncrs != null) {
        assertEquals(expectedPosIncrs[i], posIncrAtt.getPositionIncrement());
      }

      if (expectedNode != null) {
        assertEquals(expectedNode[i], nodeAtt.node());
      }

      if (expectedPos != null) {
        assertEquals(expectedPos[i], posAtt.position());
      }
    }

    assertFalse("end of stream, received token " + termAtt.toString(), t.incrementToken());
    t.end();
    t.close();
  }
View Full Code Here
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
                                final int[] expectedPosIncrs,
                                final IntsRef[] expectedNode)
  throws Exception {

    assertTrue("has TermAttribute", t.hasAttribute(CharTermAttribute.class));
    final CharTermAttribute termAtt = t.getAttribute(CharTermAttribute.class);

    TypeAttribute typeAtt = null;
    if (expectedTypes != null) {
      assertTrue("has TypeAttribute", t.hasAttribute(TypeAttribute.class));
      typeAtt = t.getAttribute(TypeAttribute.class);
    }

    DatatypeAttribute dtypeAtt = null;
    if (expectedDatatypes != null) {
      assertTrue("has DatatypeAttribute", t.hasAttribute(DatatypeAttribute.class));
      dtypeAtt = t.getAttribute(DatatypeAttribute.class);
    }

    PositionIncrementAttribute posIncrAtt = null;
    if (expectedPosIncrs != null) {
      assertTrue("has PositionIncrementAttribute", t.hasAttribute(PositionIncrementAttribute.class));
      posIncrAtt = t.getAttribute(PositionIncrementAttribute.class);
    }

    NodeAttribute nodeAtt = null;
    if (expectedNode != null) {
      assertTrue("has NodeAttribute", t.hasAttribute(NodeAttribute.class));
      nodeAtt = t.getAttribute(NodeAttribute.class);
    }

    t.setReader(new StringReader(input));
    t.reset(); // reset the stream for the new reader

    for (int i = 0; i < expectedImages.length; i++) {

      assertTrue("token "+i+" exists", t.incrementToken());

      assertEquals("i=" + i, expectedImages[i], termAtt.toString());

      if (expectedTypes != null) {
        assertEquals("i=" + i, expectedTypes[i], typeAtt.type());
      }
View Full Code Here
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
  public void assertTokenStreamContents(final TokenStream stream,
                                        final String[] expectedImages)
  throws Exception {
    assertTrue("has TermAttribute", stream.hasAttribute(CharTermAttribute.class));
    final CharTermAttribute termAtt = stream.getAttribute(CharTermAttribute.class);

    stream.reset();
    for (int i = 0; i < expectedImages.length; i++) {
      stream.clearAttributes();
      assertTrue("token "+i+" does not exists", stream.incrementToken());

      assertEquals(expectedImages[i], termAtt.toString());
    }

    assertFalse("end of stream", stream.incrementToken());
    stream.end();
    stream.close();
View Full Code Here
115
116
117
118
119
120
121
122
123
124
125
126
127
128
    final String field = ( f == null) ? "contents" : f;
    if ( a == null) a = new StandardAnalyzer(Version.LUCENE_CURRENT);

    // [1] Parse query into separate words so that when we expand we can avoid dups
    TokenStream ts = a.reusableTokenStream( field, new StringReader( query));
    CharTermAttribute termAtt = ts.addAttribute(CharTermAttribute.class);
    ts.reset();
    while (ts.incrementToken()) {
      String word = termAtt.toString();
      if ( already.add( word))
        top.add( word);
    }
    ts.end();
    ts.close();
View Full Code Here
123
124
125
126
127
128
129
130
131
132
133
134
135
136
    final Set<String> already = new HashSet<String>(); // avoid dups   
    List<String> top = new LinkedList<String>(); // needs to be separately listed..

    // [1] Parse query into separate words so that when we expand we can avoid dups
    TokenStream ts = a.reusableTokenStream( field, new StringReader( query));
    CharTermAttribute termAtt = ts.addAttribute(CharTermAttribute.class);
   
    while (ts.incrementToken()) {
      String word = termAtt.toString();
      if ( already.add( word))
        top.add( word);
    }
    final BooleanQuery tmp = new BooleanQuery();
   
View Full Code Here
TOP

Related Classes of org.apache.lucene.analysis.tokenattributes.CharTermAttribute

  • com.chenlb.mmseg4j.analysis.TokenUtils
  • com.code972.elasticsearch.rest.action.RestHebrewAnalyzerCheckWordAction
  • com.digitalpebble.behemoth.mahout.LuceneTokenizerMapper
  • com.gentics.cr.lucene.analysis.CustomPatternAnalyzerTest
  • com.ikanow.infinit.e.api.knowledge.SearchHandler
  • com.livingsocial.hive.udf.Tokenize
  • com.mozilla.grouperfish.lucene.analysis.en.NGramEnglishAnalyzer
  • com.mozilla.grouperfish.pig.eval.text.NGramTokenize
  • com.mozilla.grouperfish.pig.eval.text.Tokenize
  • ivory.core.tokenize.LuceneArabicAnalyzer
    • 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.