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

Example 51 with PhraseQuery

use of org.apache.lucene.search.PhraseQuery in project greplin-lucene-utils by Cue.

the class PhraseFilterBenchmark method main.

public static void main(String[] argv) {
    Directory directory = new RAMDirectory();
    try {
        IndexWriter writer = new IndexWriter(directory, new IndexWriterConfig(Version.LUCENE_32, new WhitespaceAnalyzer(Version.LUCENE_32)));
        int done = 0;
        for (int i = 0; i < NUMBER_OF_SEGMENTS; i++) {
            int remaining = NUMBER_OF_SEGMENTS - i;
            int numberOfDocs;
            if (remaining == 1) {
                numberOfDocs = TOTAL_DOCS - done;
            } else {
                numberOfDocs = RANDOM.nextInt(TOTAL_DOCS - done - remaining) + 1;
            }
            done += numberOfDocs;
            System.out.println("Segment #" + i + " has " + numberOfDocs + " docs");
            for (int d = 0; d < numberOfDocs; d++) {
                int wordCount = RANDOM.nextInt(WORDS_PER_DOC_DEVIATION * 2) + AVERAGE_WORDS_PER_DOC - WORDS_PER_DOC_DEVIATION;
                Document doc = new Document();
                doc.add(new Field("f", Joiner.on(' ').join(words(wordCount)), Field.Store.YES, Field.Index.ANALYZED));
                doc.add(new Field("second", RANDOM.nextInt(100) < SECOND_FIELD_MATCH_PERCENTAGE ? "yes" : "no", Field.Store.NO, Field.Index.ANALYZED));
                writer.addDocument(doc);
            }
            writer.commit();
        }
        writer.close();
        IndexReader reader = IndexReader.open(directory);
        IndexSearcher searcher = new IndexSearcher(reader);
        String[][] queries = new String[TOTAL_QUERIES][];
        Term[][] terms = new Term[TOTAL_QUERIES][];
        for (int q = 0; q < TOTAL_QUERIES; q++) {
            queries[q] = words(WORDS_PER_QUERY[RANDOM.nextInt(WORDS_PER_QUERY.length)]);
            terms[q] = new Term[queries[q].length];
            for (int qw = 0; qw < queries[q].length; qw++) {
                terms[q][qw] = new Term(FIELD, queries[q][qw]);
            }
        }
        // Warm up.
        new PhraseFilter(FIELD, queries[0]).getDocIdSet(reader);
        for (int round = 0; round < ROUNDS; round++) {
            System.out.println();
            String name1 = "filter";
            String name2 = "query";
            long ms1 = 0, ms2 = 0;
            for (int step = 0; step < 2; step++) {
                System.gc();
                System.gc();
                System.gc();
                if (step == (round & 1)) {
                    long millis = System.currentTimeMillis();
                    long hits = 0;
                    for (String[] queryWords : queries) {
                        PhraseFilter pf = new PhraseFilter(new FilterIntersectionProvider(TermsFilter.from(new Term("second", "yes"))), FIELD, queryWords);
                        hits += searcher.search(new FilteredQuery(new MatchAllDocsQuery(), pf), 1).totalHits;
                    }
                    ms1 = System.currentTimeMillis() - millis;
                    System.out.println("Finished " + name1 + " in " + ms1 + "ms with " + hits + " hits");
                } else {
                    long millis = System.currentTimeMillis();
                    long hits = 0;
                    for (Term[] queryTerms : terms) {
                        PhraseQuery pq = new PhraseQuery();
                        for (Term term : queryTerms) {
                            pq.add(term);
                        }
                        Query query = BooleanQueryBuilder.builder().must(new TermQuery(new Term("second", "yes"))).must(pq).build();
                        hits += searcher.search(query, 1).totalHits;
                    }
                    ms2 = System.currentTimeMillis() - millis;
                    System.out.println("Finished " + name2 + " in " + ms2 + "ms with " + hits + " hits");
                }
            }
            System.out.println(name1 + " took " + (int) ((100.0 * ms1) / ms2) + "% as much time as " + name2);
        }
    } catch (IOException e) {
        e.printStackTrace();
    }
}

Example 52 with PhraseQuery

use of org.apache.lucene.search.PhraseQuery in project crate by crate.

the class MultiMatchQuery method blendPhrase.

/**
 * Expand a {@link PhraseQuery} to multiple fields that share the same analyzer.
 * Returns a {@link DisjunctionMaxQuery} with a disjunction for each expanded field.
 */
static Query blendPhrase(PhraseQuery query, float tiebreaker, FieldAndFieldType... fields) {
    List<Query> disjunctions = new ArrayList<>();
    for (FieldAndFieldType field : fields) {
        int[] positions = query.getPositions();
        Term[] terms = query.getTerms();
        PhraseQuery.Builder builder = new PhraseQuery.Builder();
        for (int i = 0; i < terms.length; i++) {
            builder.add(new Term(field.fieldType.name(), terms[i].bytes()), positions[i]);
        }
        Query q = builder.build();
        if (field.boost != MultiMatchQuery.DEFAULT_BOOST) {
            q = new BoostQuery(q, field.boost);
        }
        disjunctions.add(q);
    }
    return new DisjunctionMaxQuery(disjunctions, tiebreaker);
}

Example 53 with PhraseQuery

use of org.apache.lucene.search.PhraseQuery in project zm-mailbox by Zimbra.

the class AbstractIndexStoreTest method phraseQueryWithStopWord.

@Test
public void phraseQueryWithStopWord() throws Exception {
    ZimbraLog.test.debug("--->TEST phraseQueryWithStopWord");
    Mailbox mbox = MailboxManager.getInstance().getMailboxByAccountId(MockProvisioning.DEFAULT_ACCOUNT_ID);
    createContact(mbox, "Non", "Match", "nOn.MaTchiNg@zimbra.com");
    Contact contact2 = createContact(mbox, "First", "Last", "f.last@zimbra.com", "1066 and all that with William the conqueror and others");
    createContact(mbox, "Given", "Surname", "GiV.SurN@zimbra.com");
    // Make sure all indexing has been done
    mbox.index.indexDeferredItems();
    IndexStore index = mbox.index.getIndexStore();
    ZimbraIndexSearcher searcher = index.openSearcher();
    PhraseQuery pquery = new PhraseQuery();
    // Lower case required for each term for Lucene
    pquery.add(new Term(LuceneFields.L_CONTENT, "william"));
    // pquery.add(new Term(LuceneFields.L_CONTENT, "the")); - excluded because it is a stop word
    pquery.add(new Term(LuceneFields.L_CONTENT, "conqueror"));
    ZimbraTopDocs result = searcher.search(pquery, 100);
    Assert.assertNotNull("searcher.search result object", result);
    ZimbraLog.test.debug("Result for search [hits=%d]:%s", result.getTotalHits(), result.toString());
    Assert.assertEquals("Number of hits", 1, result.getTotalHits());
    String expected1Id = String.valueOf(contact2.getId());
    String match1Id = getBlobIdForResultDoc(searcher, result, 0);
    Assert.assertEquals("Mailbox Blob ID of match", expected1Id, match1Id);
}

Example 54 with PhraseQuery

use of org.apache.lucene.search.PhraseQuery in project textdb by TextDB.

the class KeywordMatcherSourceOperator method buildPhraseQuery.

private Query buildPhraseQuery() throws DataflowException {
    BooleanQuery.Builder booleanQueryBuilder = new BooleanQuery.Builder();
    for (String attributeName : this.predicate.getAttributeNames()) {
        AttributeType attributeType = this.inputSchema.getAttribute(attributeName).getType();
        // types other than TEXT and STRING: throw Exception for now
        if (attributeType != AttributeType.STRING && attributeType != AttributeType.TEXT) {
            throw new DataflowException("KeywordPredicate: Fields other than STRING and TEXT are not supported yet");
        }
        if (attributeType == AttributeType.STRING) {
            Query termQuery = new TermQuery(new Term(attributeName, predicate.getQuery()));
            booleanQueryBuilder.add(termQuery, BooleanClause.Occur.SHOULD);
        }
        if (attributeType == AttributeType.TEXT) {
            if (queryTokenList.size() == 1) {
                Query termQuery = new TermQuery(new Term(attributeName, predicate.getQuery().toLowerCase()));
                booleanQueryBuilder.add(termQuery, BooleanClause.Occur.SHOULD);
            } else {
                PhraseQuery.Builder phraseQueryBuilder = new PhraseQuery.Builder();
                for (int i = 0; i < queryTokensWithStopwords.size(); i++) {
                    if (!StandardAnalyzer.STOP_WORDS_SET.contains(queryTokensWithStopwords.get(i))) {
                        phraseQueryBuilder.add(new Term(attributeName, queryTokensWithStopwords.get(i).toLowerCase()), i);
                    }
                }
                PhraseQuery phraseQuery = phraseQueryBuilder.build();
                booleanQueryBuilder.add(phraseQuery, BooleanClause.Occur.SHOULD);
            }
        }
    }
    return booleanQueryBuilder.build();
}

Example 55 with PhraseQuery

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

the class QueryStringQueryBuilderTests method testToQueryPhraseQuery.

public void testToQueryPhraseQuery() throws IOException {
    assumeTrue("test runs only when at least a type is registered", getCurrentTypes().length > 0);
    Query query = queryStringQuery("\"term1 term2\"").defaultField(STRING_FIELD_NAME).phraseSlop(3).toQuery(createShardContext());
    assertThat(query, instanceOf(DisjunctionMaxQuery.class));
    DisjunctionMaxQuery disjunctionMaxQuery = (DisjunctionMaxQuery) query;
    assertThat(disjunctionMaxQuery.getDisjuncts().size(), equalTo(1));
    assertThat(disjunctionMaxQuery.getDisjuncts().get(0), instanceOf(PhraseQuery.class));
    PhraseQuery phraseQuery = (PhraseQuery) disjunctionMaxQuery.getDisjuncts().get(0);
    assertThat(phraseQuery.getTerms().length, equalTo(2));
    assertThat(phraseQuery.getTerms()[0], equalTo(new Term(STRING_FIELD_NAME, "term1")));
    assertThat(phraseQuery.getTerms()[1], equalTo(new Term(STRING_FIELD_NAME, "term2")));
    assertThat(phraseQuery.getSlop(), equalTo(3));
}