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

Example 6 with MultiPhraseQuery

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

the class TestUnifiedHighlighterStrictPhrases method testSynonyms.

public void testSynonyms() throws IOException {
    indexWriter.addDocument(newDoc("mother father w mom father w dad"));
    initReaderSearcherHighlighter();
    MultiPhraseQuery query = new MultiPhraseQuery.Builder().add(new Term[] { new Term("body", "mom"), new Term("body", "mother") }).add(new Term[] { new Term("body", "dad"), new Term("body", "father") }).build();
    TopDocs topDocs = searcher.search(query, 10, Sort.INDEXORDER);
    String[] snippets = highlighter.highlight("body", query, topDocs);
    assertArrayEquals(new String[] { "<b>mother</b> <b>father</b> w <b>mom</b> <b>father</b> w dad" }, snippets);
}

Example 7 with MultiPhraseQuery

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

the class WeightedSpanTermExtractor method extract.

/**
   * Fills a <code>Map</code> with {@link WeightedSpanTerm}s using the terms from the supplied <code>Query</code>.
   * 
   * @param query
   *          Query to extract Terms from
   * @param terms
   *          Map to place created WeightedSpanTerms in
   * @throws IOException If there is a low-level I/O error
   */
protected void extract(Query query, float boost, Map<String, WeightedSpanTerm> terms) throws IOException {
    if (query instanceof BoostQuery) {
        BoostQuery boostQuery = (BoostQuery) query;
        extract(boostQuery.getQuery(), boost * boostQuery.getBoost(), terms);
    } else if (query instanceof BooleanQuery) {
        for (BooleanClause clause : (BooleanQuery) query) {
            if (!clause.isProhibited()) {
                extract(clause.getQuery(), boost, terms);
            }
        }
    } else if (query instanceof PhraseQuery) {
        PhraseQuery phraseQuery = ((PhraseQuery) query);
        Term[] phraseQueryTerms = phraseQuery.getTerms();
        if (phraseQueryTerms.length == 1) {
            extractWeightedSpanTerms(terms, new SpanTermQuery(phraseQueryTerms[0]), boost);
        } else {
            SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.length];
            for (int i = 0; i < phraseQueryTerms.length; i++) {
                clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
            }
            // sum position increments beyond 1
            int positionGaps = 0;
            int[] positions = phraseQuery.getPositions();
            if (positions.length >= 2) {
                // positions are in increasing order.   max(0,...) is just a safeguard.
                positionGaps = Math.max(0, positions[positions.length - 1] - positions[0] - positions.length + 1);
            }
            //if original slop is 0 then require inOrder
            boolean inorder = (phraseQuery.getSlop() == 0);
            SpanNearQuery sp = new SpanNearQuery(clauses, phraseQuery.getSlop() + positionGaps, inorder);
            extractWeightedSpanTerms(terms, sp, boost);
        }
    } else if (query instanceof TermQuery || query instanceof SynonymQuery) {
        extractWeightedTerms(terms, query, boost);
    } else if (query instanceof SpanQuery) {
        extractWeightedSpanTerms(terms, (SpanQuery) query, boost);
    } else if (query instanceof ConstantScoreQuery) {
        final Query q = ((ConstantScoreQuery) query).getQuery();
        if (q != null) {
            extract(q, boost, terms);
        }
    } else if (query instanceof CommonTermsQuery) {
        // specialized since rewriting would change the result query 
        // this query is TermContext sensitive.
        extractWeightedTerms(terms, query, boost);
    } else if (query instanceof DisjunctionMaxQuery) {
        for (Query clause : ((DisjunctionMaxQuery) query)) {
            extract(clause, boost, terms);
        }
    } else if (query instanceof ToParentBlockJoinQuery) {
        extract(((ToParentBlockJoinQuery) query).getChildQuery(), boost, terms);
    } else if (query instanceof ToChildBlockJoinQuery) {
        extract(((ToChildBlockJoinQuery) query).getParentQuery(), boost, terms);
    } else if (query instanceof MultiPhraseQuery) {
        final MultiPhraseQuery mpq = (MultiPhraseQuery) query;
        final Term[][] termArrays = mpq.getTermArrays();
        final int[] positions = mpq.getPositions();
        if (positions.length > 0) {
            int maxPosition = positions[positions.length - 1];
            for (int i = 0; i < positions.length - 1; ++i) {
                if (positions[i] > maxPosition) {
                    maxPosition = positions[i];
                }
            }
            @SuppressWarnings({ "unchecked", "rawtypes" }) final List<SpanQuery>[] disjunctLists = new List[maxPosition + 1];
            int distinctPositions = 0;
            for (int i = 0; i < termArrays.length; ++i) {
                final Term[] termArray = termArrays[i];
                List<SpanQuery> disjuncts = disjunctLists[positions[i]];
                if (disjuncts == null) {
                    disjuncts = (disjunctLists[positions[i]] = new ArrayList<>(termArray.length));
                    ++distinctPositions;
                }
                for (Term aTermArray : termArray) {
                    disjuncts.add(new SpanTermQuery(aTermArray));
                }
            }
            int positionGaps = 0;
            int position = 0;
            final SpanQuery[] clauses = new SpanQuery[distinctPositions];
            for (List<SpanQuery> disjuncts : disjunctLists) {
                if (disjuncts != null) {
                    clauses[position++] = new SpanOrQuery(disjuncts.toArray(new SpanQuery[disjuncts.size()]));
                } else {
                    ++positionGaps;
                }
            }
            if (clauses.length == 1) {
                extractWeightedSpanTerms(terms, clauses[0], boost);
            } else {
                final int slop = mpq.getSlop();
                final boolean inorder = (slop == 0);
                SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
                extractWeightedSpanTerms(terms, sp, boost);
            }
        }
    } else if (query instanceof MatchAllDocsQuery) {
    //nothing
    } else if (query instanceof CustomScoreQuery) {
        extract(((CustomScoreQuery) query).getSubQuery(), boost, terms);
    } else if (isQueryUnsupported(query.getClass())) {
    // nothing
    } else {
        if (query instanceof MultiTermQuery && (!expandMultiTermQuery || !fieldNameComparator(((MultiTermQuery) query).getField()))) {
            return;
        }
        Query origQuery = query;
        final IndexReader reader = getLeafContext().reader();
        Query rewritten;
        if (query instanceof MultiTermQuery) {
            rewritten = MultiTermQuery.SCORING_BOOLEAN_REWRITE.rewrite(reader, (MultiTermQuery) query);
        } else {
            rewritten = origQuery.rewrite(reader);
        }
        if (rewritten != origQuery) {
            // only rewrite once and then flatten again - the rewritten query could have a special treatment
            // if this method is overwritten in a subclass or above in the next recursion
            extract(rewritten, boost, terms);
        } else {
            extractUnknownQuery(query, terms);
        }
    }
}

Example 8 with MultiPhraseQuery

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

the class AbstractIndexStoreTest method multiPhraseQuery.

@Test
public void multiPhraseQuery() throws Exception {
    ZimbraLog.test.debug("--->TEST multiPhraseQuery");
    Mailbox mbox = MailboxManager.getInstance().getMailboxByAccountId(MockProvisioning.DEFAULT_ACCOUNT_ID);
    createContact(mbox, "Non", "Match", "nOn.MaTchiNg@zimbra.com");
    Contact contact1 = createContact(mbox, "Paul", "AA", "aa@example.net", "Software Development Engineer");
    createContact(mbox, "Jane", "BB", "bb@example.net", "Software Planning Engineer");
    Contact contact2 = createContact(mbox, "Peter", "CC", "cc@example.net", "Software Dev Engineer");
    createContact(mbox, "Avril", "DD", "dd@example.net", "Software Architectural Engineer");
    Contact contact3 = createContact(mbox, "Leo", "EE", "ee@example.net", "Software Developer Engineer");
    Contact contact4 = createContact(mbox, "Wow", "DD", "dd@example.net", "Softly Development Engineer");
    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();
    MultiPhraseQuery pquery = new MultiPhraseQuery();
    // Lower case required for each term for Lucene
    Term[] firstWords = { new Term(LuceneFields.L_CONTENT, "softly"), new Term(LuceneFields.L_CONTENT, "software") };
    pquery.add(firstWords);
    Term[] secondWords = { new Term(LuceneFields.L_CONTENT, "dev"), new Term(LuceneFields.L_CONTENT, "development"), new Term(LuceneFields.L_CONTENT, "developer") };
    pquery.add(secondWords);
    pquery.add(new Term(LuceneFields.L_CONTENT, "engineer"));
    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", 4, result.getTotalHits());
    List<String> expecteds = Lists.newArrayList();
    List<String> matches = Lists.newArrayList();
    matches.add(getBlobIdForResultDoc(searcher, result, 0));
    matches.add(getBlobIdForResultDoc(searcher, result, 1));
    matches.add(getBlobIdForResultDoc(searcher, result, 2));
    matches.add(getBlobIdForResultDoc(searcher, result, 3));
    expecteds.add(String.valueOf(contact1.getId()));
    expecteds.add(String.valueOf(contact2.getId()));
    expecteds.add(String.valueOf(contact3.getId()));
    expecteds.add(String.valueOf(contact4.getId()));
    Collections.sort(matches);
    Collections.sort(expecteds);
    for (int ndx = 0; ndx < 4; ndx++) {
        Assert.assertEquals("Match Blob ID", expecteds.get(0), matches.get(0));
    }
}

Example 9 with MultiPhraseQuery

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

the class MatchQuery method multiPhraseQuery.

private static Query multiPhraseQuery(String field, TokenStream stream, int slop, boolean enablePositionIncrements) throws IOException {
    MultiPhraseQuery.Builder mpqb = new MultiPhraseQuery.Builder();
    mpqb.setSlop(slop);
    TermToBytesRefAttribute termAtt = stream.getAttribute(TermToBytesRefAttribute.class);
    PositionIncrementAttribute posIncrAtt = stream.getAttribute(PositionIncrementAttribute.class);
    int position = -1;
    List<Term> multiTerms = new ArrayList<>();
    stream.reset();
    while (stream.incrementToken()) {
        int positionIncrement = posIncrAtt.getPositionIncrement();
        if (positionIncrement > 0 && multiTerms.size() > 0) {
            if (enablePositionIncrements) {
                mpqb.add(multiTerms.toArray(new Term[0]), position);
            } else {
                mpqb.add(multiTerms.toArray(new Term[0]));
            }
            multiTerms.clear();
        }
        position += positionIncrement;
        multiTerms.add(new Term(field, termAtt.getBytesRef()));
    }
    if (enablePositionIncrements) {
        mpqb.add(multiTerms.toArray(new Term[0]), position);
    } else {
        mpqb.add(multiTerms.toArray(new Term[0]));
    }
    return mpqb.build();
}

Example 10 with MultiPhraseQuery

use of org.apache.lucene.search.MultiPhraseQuery 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);
    }
}