Examples with CoreMap edu.stanford.nlp.util.CoreMap used on opensource projects

Example 76 with CoreMap

use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.

the class TokenSequenceMatcherITest method testCompile.

//just to test if a pattern is compiling or not
public void testCompile() {
    String s = "(?$se \"matching\" \"this\"|\"don't\")";
    CoreMap doc = createDocument("does this do matching this");
    TokenSequencePattern p = TokenSequencePattern.compile(s);
    TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    boolean match = m.find();
    assertTrue(match);
//assertEquals(m.group(), "matching this");
}

Example 77 with CoreMap

use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.

the class TokenSequenceMatcherITest method testTokenSequenceMatcher1.

public void testTokenSequenceMatcher1() throws IOException {
    CoreMap doc = createDocument(testText1);
    // Test simple sequence
    TokenSequencePattern p = TokenSequencePattern.compile(getSequencePatternExpr("Archbishop", "of", "Canterbury"));
    TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    boolean match = m.find();
    assertTrue(match);
    assertEquals("Archbishop of Canterbury", m.group());
    match = m.find();
    assertFalse(match);
    m.reset();
    match = m.find();
    assertTrue(match);
    assertEquals("Archbishop of Canterbury", m.group());
    m.reset();
    match = m.matches();
    assertFalse(match);
    // Test sequence with or
    p = TokenSequencePattern.compile(new SequencePattern.OrPatternExpr(getSequencePatternExpr("Archbishop", "of", "Canterbury"), getSequencePatternExpr("Bishop", "of", "London")));
    m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("Bishop of London", m.group());
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("Archbishop of Canterbury", m.group());
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("Bishop of London", m.group());
    match = m.find();
    assertFalse(match);
    p = TokenSequencePattern.compile(new SequencePattern.SequencePatternExpr(SequencePattern.SEQ_BEGIN_PATTERN_EXPR, getSequencePatternExpr("Archbishop", "of", "Canterbury")));
    m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    match = m.find();
    assertFalse(match);
    p = TokenSequencePattern.compile(new SequencePattern.SequencePatternExpr(SequencePattern.SEQ_BEGIN_PATTERN_EXPR, getSequencePatternExpr("Mellitus", "was", "the")));
    m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("Mellitus was the", m.group());
    match = m.find();
    assertFalse(match);
    p = TokenSequencePattern.compile(new SequencePattern.SequencePatternExpr(getSequencePatternExpr("Archbishop", "of", "Canterbury"), SequencePattern.SEQ_END_PATTERN_EXPR));
    m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    match = m.find();
    assertFalse(match);
    p = TokenSequencePattern.compile(new SequencePattern.SequencePatternExpr(getSequencePatternExpr("London", "in", "604", "."), SequencePattern.SEQ_END_PATTERN_EXPR));
    m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("London in 604.", m.group());
    match = m.find();
    assertFalse(match);
}

Example 78 with CoreMap

use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.

the class TokenSequenceMatcherITest method testTokenSequenceMatcher2.

public void testTokenSequenceMatcher2() throws IOException {
    CoreMap doc = createDocument(testText1);
    TokenSequencePattern p = TokenSequencePattern.compile(getSequencePatternExpr(".*", ".*", "of", ".*"));
    TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    boolean match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("first Bishop of London", m.group());
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("third Archbishop of Canterbury", m.group());
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("a member of the", m.group());
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("as Bishop of London", m.group());
    match = m.find();
    assertFalse(match);
    // Test sequence with groups
    p = TokenSequencePattern.compile(new SequencePattern.SequencePatternExpr(new SequencePattern.GroupPatternExpr(getSequencePatternExpr(".*", ".*")), getNodePatternExpr("of"), new SequencePattern.GroupPatternExpr(getSequencePatternExpr(".*"))));
    m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("first Bishop of London", m.group());
    assertEquals("first Bishop", m.group(1));
    assertEquals("London", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("third Archbishop of Canterbury", m.group());
    assertEquals("third Archbishop", m.group(1));
    assertEquals("Canterbury", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("a member of the", m.group());
    assertEquals("a member", m.group(1));
    assertEquals("the", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("as Bishop of London", m.group());
    assertEquals("as Bishop", m.group(1));
    assertEquals("London", m.group(2));
    match = m.find();
    assertFalse(match);
}

Example 79 with CoreMap

use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.

the class TokenSequenceMatcherITest method testTokenSequenceMatcher6.

public void testTokenSequenceMatcher6() throws IOException {
    CoreMap doc = createDocument(testText1);
    TokenSequencePattern p = TokenSequencePattern.compile("[ /.*/ ] [ /.*/ ] [/of/] [/.*/]");
    TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    boolean match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("first Bishop of London", m.group());
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("third Archbishop of Canterbury", m.group());
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("a member of the", m.group());
    match = m.find();
    assertTrue(match);
    assertEquals(0, m.groupCount());
    assertEquals("as Bishop of London", m.group());
    match = m.find();
    assertFalse(match);
    p = TokenSequencePattern.compile("([ /.*/ ] [ /.*/ ]) [/of/] ([/.*/])");
    m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("first Bishop of London", m.group());
    assertEquals("first Bishop", m.group(1));
    assertEquals("London", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("third Archbishop of Canterbury", m.group());
    assertEquals("third Archbishop", m.group(1));
    assertEquals("Canterbury", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("a member of the", m.group());
    assertEquals("a member", m.group(1));
    assertEquals("the", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("as Bishop of London", m.group());
    assertEquals("as Bishop", m.group(1));
    assertEquals("London", m.group(2));
    match = m.find();
    assertFalse(match);
}

Example 80 with CoreMap

use of edu.stanford.nlp.util.CoreMap in project CoreNLP by stanfordnlp.

the class TokenSequenceMatcherITest method testTokenSequenceMatcherAll.

public void testTokenSequenceMatcherAll() throws IOException {
    CoreMap doc = createDocument(testText1);
    TokenSequencePattern p = TokenSequencePattern.compile("(/[A-Za-z]+/{1,2}) /of/ (/[A-Za-z]+/{1,3}?) ");
    TokenSequenceMatcher m = p.getMatcher(doc.get(CoreAnnotations.TokensAnnotation.class));
    m.setFindType(SequenceMatcher.FindType.FIND_ALL);
    // Test finding of ALL matching sequences
    // NOTE: when using FIND_ALL greedy/reluctant modifiers are not enforced
    //       perhaps should add syntax where some of them are enforced...
    boolean match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("first Bishop of London", m.group());
    assertEquals("first Bishop", m.group(1));
    assertEquals("London", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("Bishop of London", m.group());
    assertEquals("Bishop", m.group(1));
    assertEquals("London", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("third Archbishop of Canterbury", m.group());
    assertEquals("third Archbishop", m.group(1));
    assertEquals("Canterbury", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("Archbishop of Canterbury", m.group());
    assertEquals("Archbishop", m.group(1));
    assertEquals("Canterbury", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("a member of the", m.group());
    assertEquals("a member", m.group(1));
    assertEquals("the", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("a member of the Gregorian", m.group());
    assertEquals("a member", m.group(1));
    assertEquals("the Gregorian", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("a member of the Gregorian mission", m.group());
    assertEquals("a member", m.group(1));
    assertEquals("the Gregorian mission", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("member of the", m.group());
    assertEquals("member", m.group(1));
    assertEquals("the", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("member of the Gregorian", m.group());
    assertEquals("member", m.group(1));
    assertEquals("the Gregorian", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("member of the Gregorian mission", m.group());
    assertEquals("member", m.group(1));
    assertEquals("the Gregorian mission", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("as Bishop of London", m.group());
    assertEquals("as Bishop", m.group(1));
    assertEquals("London", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("as Bishop of London in", m.group());
    assertEquals("as Bishop", m.group(1));
    assertEquals("London in", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("Bishop of London", m.group());
    assertEquals("Bishop", m.group(1));
    assertEquals("London", m.group(2));
    match = m.find();
    assertTrue(match);
    assertEquals(2, m.groupCount());
    assertEquals("Bishop of London in", m.group());
    assertEquals("Bishop", m.group(1));
    assertEquals("London in", m.group(2));
    match = m.find();
    assertFalse(match);
}