Examples with SynonymQuery org.apache.lucene.search.SynonymQuery used on opensource projects

Example 11 with SynonymQuery

use of org.apache.lucene.search.SynonymQuery in project elasticsearch by elastic.

the class SimpleQueryParser method newPossiblyAnalyzedQuery.

/**
     * Analyze the given string using its analyzer, constructing either a
     * {@code PrefixQuery} or a {@code BooleanQuery} made up
     * of {@code TermQuery}s and {@code PrefixQuery}s
     */
private Query newPossiblyAnalyzedQuery(String field, String termStr) {
    List<List<BytesRef>> tlist = new ArrayList<>();
    // get Analyzer from superclass and tokenize the term
    try (TokenStream source = getAnalyzer().tokenStream(field, termStr)) {
        source.reset();
        List<BytesRef> currentPos = new ArrayList<>();
        CharTermAttribute termAtt = source.addAttribute(CharTermAttribute.class);
        PositionIncrementAttribute posAtt = source.addAttribute(PositionIncrementAttribute.class);
        try {
            boolean hasMoreTokens = source.incrementToken();
            while (hasMoreTokens) {
                if (currentPos.isEmpty() == false && posAtt.getPositionIncrement() > 0) {
                    tlist.add(currentPos);
                    currentPos = new ArrayList<>();
                }
                final BytesRef term = getAnalyzer().normalize(field, termAtt.toString());
                currentPos.add(term);
                hasMoreTokens = source.incrementToken();
            }
            if (currentPos.isEmpty() == false) {
                tlist.add(currentPos);
            }
        } catch (IOException e) {
        // ignore
        // TODO: we should not ignore the exception and return a prefix query with the original term ?
        }
    } catch (IOException e) {
        // Bail on any exceptions, going with a regular prefix query
        return new PrefixQuery(new Term(field, termStr));
    }
    if (tlist.size() == 0) {
        return null;
    }
    if (tlist.size() == 1 && tlist.get(0).size() == 1) {
        return new PrefixQuery(new Term(field, tlist.get(0).get(0)));
    }
    // build a boolean query with prefix on the last position only.
    BooleanQuery.Builder builder = new BooleanQuery.Builder();
    for (int pos = 0; pos < tlist.size(); pos++) {
        List<BytesRef> plist = tlist.get(pos);
        boolean isLastPos = (pos == tlist.size() - 1);
        Query posQuery;
        if (plist.size() == 1) {
            if (isLastPos) {
                posQuery = new PrefixQuery(new Term(field, plist.get(0)));
            } else {
                posQuery = newTermQuery(new Term(field, plist.get(0)));
            }
        } else if (isLastPos == false) {
            // build a synonym query for terms in the same position.
            Term[] terms = new Term[plist.size()];
            for (int i = 0; i < plist.size(); i++) {
                terms[i] = new Term(field, plist.get(i));
            }
            posQuery = new SynonymQuery(terms);
        } else {
            BooleanQuery.Builder innerBuilder = new BooleanQuery.Builder();
            for (BytesRef token : plist) {
                innerBuilder.add(new BooleanClause(new PrefixQuery(new Term(field, token)), BooleanClause.Occur.SHOULD));
            }
            posQuery = innerBuilder.setDisableCoord(true).build();
        }
        builder.add(new BooleanClause(posQuery, getDefaultOperator()));
    }
    return builder.build();
}

Example 12 with SynonymQuery

use of org.apache.lucene.search.SynonymQuery in project elasticsearch by elastic.

the class CustomFieldQuery method flatten.

@Override
void flatten(Query sourceQuery, IndexReader reader, Collection<Query> flatQueries, float boost) throws IOException {
    if (sourceQuery instanceof BoostQuery) {
        BoostQuery bq = (BoostQuery) sourceQuery;
        sourceQuery = bq.getQuery();
        boost *= bq.getBoost();
        flatten(sourceQuery, reader, flatQueries, boost);
    } else if (sourceQuery instanceof SpanTermQuery) {
        super.flatten(new TermQuery(((SpanTermQuery) sourceQuery).getTerm()), reader, flatQueries, boost);
    } else if (sourceQuery instanceof ConstantScoreQuery) {
        flatten(((ConstantScoreQuery) sourceQuery).getQuery(), reader, flatQueries, boost);
    } else if (sourceQuery instanceof FunctionScoreQuery) {
        flatten(((FunctionScoreQuery) sourceQuery).getSubQuery(), reader, flatQueries, boost);
    } else if (sourceQuery instanceof MultiPhrasePrefixQuery) {
        flatten(sourceQuery.rewrite(reader), reader, flatQueries, boost);
    } else if (sourceQuery instanceof FiltersFunctionScoreQuery) {
        flatten(((FiltersFunctionScoreQuery) sourceQuery).getSubQuery(), reader, flatQueries, boost);
    } else if (sourceQuery instanceof MultiPhraseQuery) {
        MultiPhraseQuery q = ((MultiPhraseQuery) sourceQuery);
        convertMultiPhraseQuery(0, new int[q.getTermArrays().length], q, q.getTermArrays(), q.getPositions(), reader, flatQueries);
    } else if (sourceQuery instanceof BlendedTermQuery) {
        final BlendedTermQuery blendedTermQuery = (BlendedTermQuery) sourceQuery;
        flatten(blendedTermQuery.rewrite(reader), reader, flatQueries, boost);
    } else if (sourceQuery instanceof ESToParentBlockJoinQuery) {
        ESToParentBlockJoinQuery blockJoinQuery = (ESToParentBlockJoinQuery) sourceQuery;
        flatten(blockJoinQuery.getChildQuery(), reader, flatQueries, boost);
    } else if (sourceQuery instanceof BoostingQuery) {
        BoostingQuery boostingQuery = (BoostingQuery) sourceQuery;
        //flatten positive query with query boost
        flatten(boostingQuery.getMatch(), reader, flatQueries, boost);
        //flatten negative query with negative boost
        flatten(boostingQuery.getContext(), reader, flatQueries, boostingQuery.getBoost());
    } else if (sourceQuery instanceof SynonymQuery) {
        // SynonymQuery should be handled by the parent class directly.
        // This statement should be removed when https://issues.apache.org/jira/browse/LUCENE-7484 is merged.
        SynonymQuery synQuery = (SynonymQuery) sourceQuery;
        for (Term term : synQuery.getTerms()) {
            flatten(new TermQuery(term), reader, flatQueries, boost);
        }
    } else {
        super.flatten(sourceQuery, reader, flatQueries, boost);
    }
}

Example 13 with SynonymQuery

use of org.apache.lucene.search.SynonymQuery in project elasticsearch by elastic.

the class QueryAnalyzerTests method testSynonymQuery.

public void testSynonymQuery() {
    SynonymQuery query = new SynonymQuery();
    Result result = analyze(query);
    assertThat(result.verified, is(true));
    assertThat(result.terms.isEmpty(), is(true));
    query = new SynonymQuery(new Term("_field", "_value1"), new Term("_field", "_value2"));
    result = analyze(query);
    assertThat(result.verified, is(true));
    assertTermsEqual(result.terms, new Term("_field", "_value1"), new Term("_field", "_value2"));
}

Example 14 with SynonymQuery

use of org.apache.lucene.search.SynonymQuery in project lucene-solr by apache.

the class SynonymTokenizer method testHighlightingSynonymQuery.

public void testHighlightingSynonymQuery() throws Exception {
    searcher = newSearcher(reader);
    Query query = new SynonymQuery(new Term(FIELD_NAME, "jfk"), new Term(FIELD_NAME, "kennedy"));
    QueryScorer scorer = new QueryScorer(query, FIELD_NAME);
    Highlighter highlighter = new Highlighter(scorer);
    TokenStream stream = getAnyTokenStream(FIELD_NAME, 2);
    Fragmenter fragmenter = new SimpleSpanFragmenter(scorer);
    highlighter.setTextFragmenter(fragmenter);
    String storedField = searcher.doc(2).get(FIELD_NAME);
    String fragment = highlighter.getBestFragment(stream, storedField);
    assertEquals("<B>JFK</B> has been shot", fragment);
    stream = getAnyTokenStream(FIELD_NAME, 3);
    storedField = searcher.doc(3).get(FIELD_NAME);
    fragment = highlighter.getBestFragment(stream, storedField);
    assertEquals("John <B>Kennedy</B> has been shot", fragment);
}

Example 15 with SynonymQuery

use of org.apache.lucene.search.SynonymQuery in project lucene-solr by apache.

the class TestQueryBuilder method testCJKSynonymsAND.

/** synonyms with default AND operator */
public void testCJKSynonymsAND() throws Exception {
    BooleanQuery.Builder expected = new BooleanQuery.Builder();
    expected.add(new TermQuery(new Term("field", "中")), BooleanClause.Occur.MUST);
    SynonymQuery inner = new SynonymQuery(new Term("field", "国"), new Term("field", "國"));
    expected.add(inner, BooleanClause.Occur.MUST);
    QueryBuilder builder = new QueryBuilder(new MockCJKSynonymAnalyzer());
    assertEquals(expected.build(), builder.createBooleanQuery("field", "中国", BooleanClause.Occur.MUST));
}