org.apache.lucene.search.MultiPhraseQuery Java Examples

@Override
void flatten(Query sourceQuery, IndexReader reader, Collection<Query> flatQueries, float boost) throws IOException {
    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().size()], 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 {
        super.flatten(sourceQuery, reader, flatQueries, boost);
    }
}
 

private Query applySlop(Query q, int slop) {
    if (q instanceof PhraseQuery) {
        PhraseQuery pq = (PhraseQuery) q;
        PhraseQuery.Builder builder = new PhraseQuery.Builder();
        builder.setSlop(slop);
        final Term[] terms = pq.getTerms();
        final int[] positions = pq.getPositions();
        for (int i = 0; i < terms.length; ++i) {
            builder.add(terms[i], positions[i]);
        }
        pq = builder.build();
        pq.setBoost(q.getBoost());
        return pq;
    } else if (q instanceof MultiPhraseQuery) {
        ((MultiPhraseQuery) q).setSlop(slop);
        return q;
    } else {
        return q;
    }
}
 

public void testQueryScorerMultiPhraseQueryHighlightingWithGap() throws Exception {
  MultiPhraseQuery.Builder mpqb = new MultiPhraseQuery.Builder();

  /*
   * The toString of MultiPhraseQuery doesn't work so well with these
   * out-of-order additions, but the Query itself seems to match accurately.
   */

  mpqb.add(new Term[] { new Term(FIELD_NAME, "wordz") }, 2);
  mpqb.add(new Term[] { new Term(FIELD_NAME, "wordx") }, 0);

  doSearching(mpqb.build());

  final int maxNumFragmentsRequired = 1;
  final int expectedHighlights = 2;

  assertExpectedHighlightCount(maxNumFragmentsRequired, expectedHighlights);
}
 

private Query applySlop(Query q, int slop) {
	if (q instanceof PhraseQuery) {
		PhraseQuery.Builder builder = new PhraseQuery.Builder();
		builder.setSlop(slop);
		PhraseQuery pq = (PhraseQuery) q;
		org.apache.lucene.index.Term[] terms = pq.getTerms();
		int[] positions = pq.getPositions();
		for (int i = 0; i < terms.length; ++i) {
			builder.add(terms[i], positions[i]);
		}
		q = builder.build();
	}
	else if (q instanceof MultiPhraseQuery) {
		MultiPhraseQuery mpq = (MultiPhraseQuery) q;

		if (slop != mpq.getSlop()) {
			q = new MultiPhraseQuery.Builder(mpq).setSlop(slop).build();
		}
	}
	return q;
}
 

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);
  if (highlighter.getFlags("body").contains(HighlightFlag.WEIGHT_MATCHES)) {
    assertArrayEquals(new String[]{"<b>mother father</b> w <b>mom father</b> w dad"}, snippets);
  } else {
    assertArrayEquals(new String[]{"<b>mother</b> <b>father</b> w <b>mom</b> <b>father</b> w dad"}, snippets);
  }
}
 

private Query applySlop(Query q, int slop) {
  if (q instanceof PhraseQuery) {
    PhraseQuery.Builder builder = new PhraseQuery.Builder();
    builder.setSlop(slop);
    PhraseQuery pq = (PhraseQuery) q;
    org.apache.lucene.index.Term[] terms = pq.getTerms();
    int[] positions = pq.getPositions();
    for (int i = 0; i < terms.length; ++i) {
      builder.add(terms[i], positions[i]);
    }
    q = builder.build();
  } else if (q instanceof MultiPhraseQuery) {
    MultiPhraseQuery mpq = (MultiPhraseQuery)q;
    
    if (slop != mpq.getSlop()) {
      q = new MultiPhraseQuery.Builder(mpq).setSlop(slop).build();
    }
  }
  return q;
}
 

public void testMultiPhraseQuery() throws Exception {
  Directory dir = newDirectory();
  IndexWriter w = new IndexWriter(dir, newIndexWriterConfig(new CannedAnalyzer()));
  Document doc = new Document();
  doc.add(newTextField("field", "", Field.Store.NO));
  w.addDocument(doc);
  IndexReader r = DirectoryReader.open(w);
  IndexSearcher s = newSearcher(r);
  
  Query q = new StandardQueryParser(new CannedAnalyzer()).parse("\"a\"", "field");
  assertTrue(q instanceof MultiPhraseQuery);
  assertEquals(1, s.search(q, 10).totalHits.value);
  r.close();
  w.close();
  dir.close();
}
 

public void testMultiPhraseQueryParsing() throws Exception {
  TokenAndPos[] INCR_0_QUERY_TOKENS_AND = new TokenAndPos[]{
      new TokenAndPos("a", 0),
      new TokenAndPos("1", 0),
      new TokenAndPos("b", 1),
      new TokenAndPos("1", 1),
      new TokenAndPos("c", 2)
  };

  QueryParser qp = new QueryParser("field", new CannedAnalyzer(INCR_0_QUERY_TOKENS_AND));
  Query q = qp.parse("\"this text is acually ignored\"");
  assertTrue("wrong query type!", q instanceof MultiPhraseQuery);

  MultiPhraseQuery.Builder multiPhraseQueryBuilder = new MultiPhraseQuery.Builder();
  multiPhraseQueryBuilder.add(new Term[]{ new Term("field", "a"), new Term("field", "1") }, -1);
  multiPhraseQueryBuilder.add(new Term[]{ new Term("field", "b"), new Term("field", "1") }, 0);
  multiPhraseQueryBuilder.add(new Term[]{ new Term("field", "c") }, 1);

  assertEquals(multiPhraseQueryBuilder.build(), q);
}
 

@Override
public Query rewrite(IndexReader reader) throws IOException {
    Query rewritten = super.rewrite(reader);
    if (rewritten != this) {
        return rewritten;
    }
    if (termArrays.isEmpty()) {
        return new MatchNoDocsQuery();
    }
    MultiPhraseQuery.Builder query = new MultiPhraseQuery.Builder();
    query.setSlop(slop);
    int sizeMinus1 = termArrays.size() - 1;
    for (int i = 0; i < sizeMinus1; i++) {
        query.add(termArrays.get(i), positions.get(i));
    }
    Term[] suffixTerms = termArrays.get(sizeMinus1);
    int position = positions.get(sizeMinus1);
    ObjectHashSet<Term> terms = new ObjectHashSet<>();
    for (Term term : suffixTerms) {
        getPrefixTerms(terms, term, reader);
        if (terms.size() > maxExpansions) {
            break;
        }
    }
    if (terms.isEmpty()) {
        if (sizeMinus1 == 0) {
            // no prefix and the phrase query is empty
            return Queries.newMatchNoDocsQuery("No terms supplied for " + MultiPhrasePrefixQuery.class.getName());
        }

        // if the terms does not exist we could return a MatchNoDocsQuery but this would break the unified highlighter
        // which rewrites query with an empty reader.
        return new BooleanQuery.Builder()
            .add(query.build(), BooleanClause.Occur.MUST)
            .add(Queries.newMatchNoDocsQuery("No terms supplied for " + MultiPhrasePrefixQuery.class.getName()),
                BooleanClause.Occur.MUST).build();
    }
    query.add(terms.toArray(Term.class), position);
    return query.build();
}
 

@Test
public void testMultiPhraseQuery() throws Exception {
    // 多短语查询
    Term[] terms = new Term[] { new Term("title", "NeverEnding"), new Term("title", "Xinghua,") };
    Term term = new Term("title", "The");
    // add之间认为是OR操作,即"NeverEnding", "Xinghua,"和"The"之间的slop不大于3
    MultiPhraseQuery multiPhraseQuery = new MultiPhraseQuery.Builder().add(terms).add(term).setSlop(3).build();
    TopDocs search = searcher.search(multiPhraseQuery, 1000);
    Assert.assertEquals(2, search.totalHits.value);
}
 

@Override
public 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();
}
 

public static void assertEqualsQuery(Query expected, Query actual) {
  assertEquals(expected.getClass(), actual.getClass());
  if (expected instanceof BooleanQuery) {
    assertEqualsBooleanQuery((BooleanQuery) expected, (BooleanQuery) actual);
  } else if (expected instanceof SuperQuery) {
    assertEqualsSuperQuery((SuperQuery) expected, (SuperQuery) actual);
  } else if (expected instanceof TermQuery) {
    assertEqualsTermQuery((TermQuery) expected, (TermQuery) actual);
  } else if (expected instanceof PrefixQuery) {
    assertEqualsPrefixQuery((PrefixQuery) expected, (PrefixQuery) actual);
  } else if (expected instanceof WildcardQuery) {
    assertEqualsWildcardQuery((WildcardQuery) expected, (WildcardQuery) actual);
  } else if (expected instanceof FuzzyQuery) {
    assertEqualsFuzzyQuery((FuzzyQuery) expected, (FuzzyQuery) actual);
  } else if (expected instanceof RegexpQuery) {
    assertEqualsRegexpQuery((RegexpQuery) expected, (RegexpQuery) actual);
  } else if (expected instanceof TermRangeQuery) {
    assertEqualsTermRangeQuery((TermRangeQuery) expected, (TermRangeQuery) actual);
  } else if (expected instanceof MatchAllDocsQuery) {
    assertEqualsMatchAllDocsQuery((MatchAllDocsQuery) expected, (MatchAllDocsQuery) actual);
  } else if (expected instanceof MultiPhraseQuery) {
    assertEqualsMultiPhraseQuery((MultiPhraseQuery) expected, (MultiPhraseQuery) actual);
  } else if (expected instanceof PhraseQuery) {
    assertEqualsPhraseQuery((PhraseQuery) expected, (PhraseQuery) actual);
  } else if (expected instanceof NumericRangeQuery<?>) {
    assertEqualsNumericRangeQuery((NumericRangeQuery<?>) expected, (NumericRangeQuery<?>) actual);
  } else {
    fail("Type [" + expected.getClass() + "] not supported");
  }
}
 

@Override
protected MultiPhraseQuery newMultiPhraseQuery() {
  return new MultiPhraseQuery() {
    @Override
    public void add(Term[] terms, int position) {
      super.add(terms, position);
      for (Term term : terms) {
        String resolvedField = _fieldManager.resolveField(term.field());
        customQueryCheck(resolvedField);
        addField(this, resolvedField);
      }
    }
  };
}
 

/**
 * Base implementation delegates to {@link #getFieldQuery(String,String,boolean,boolean)}.
 * This method may be overridden, for example, to return
 * a SpanNearQuery instead of a PhraseQuery.
 *
 */
protected Query getFieldQuery(String field, String queryText, int slop)
      throws SyntaxError {
  Query query = getFieldQuery(field, queryText, true, false);

  // only set slop of the phrase query was a result of this parser
  // and not a sub-parser.
  if (subQParser == null) {
    if (query instanceof PhraseQuery) {
      PhraseQuery pq = (PhraseQuery) query;
      Term[] terms = pq.getTerms();
      int[] positions = pq.getPositions();
      PhraseQuery.Builder builder = new PhraseQuery.Builder();
      for (int i = 0; i < terms.length; ++i) {
        builder.add(terms[i], positions[i]);
      }
      builder.setSlop(slop);
      query = builder.build();
    } else if (query instanceof MultiPhraseQuery) {
      MultiPhraseQuery mpq = (MultiPhraseQuery)query;

      if (slop != mpq.getSlop()) {
        query = new MultiPhraseQuery.Builder(mpq).setSlop(slop).build();
      }
    }
  }

  return query;
}
 

/** forms multiphrase query */
public void testCJKSynonymsPhrase() throws Exception {
  MultiPhraseQuery.Builder expectedBuilder = new MultiPhraseQuery.Builder();
  expectedBuilder.add(new Term("field", "中"));
  expectedBuilder.add(new Term[] { new Term("field", "国"), new Term("field", "國")});
  QueryBuilder builder = new QueryBuilder(new MockCJKSynonymAnalyzer());
  assertEquals(expectedBuilder.build(), builder.createPhraseQuery("field", "中国"));
  expectedBuilder.setSlop(3);
  assertEquals(expectedBuilder.build(), builder.createPhraseQuery("field", "中国", 3));
}
 

/** forms multiphrase query */
public void testSynonymsPhrase() throws Exception {
  MultiPhraseQuery.Builder expectedBuilder = new MultiPhraseQuery.Builder();
  expectedBuilder.add(new Term("field", "old"));
  expectedBuilder.add(new Term[] { new Term("field", "dogs"), new Term("field", "dog") });
  QueryBuilder builder = new QueryBuilder(new MockSynonymAnalyzer());
  assertEquals(expectedBuilder.build(), builder.createPhraseQuery("field", "old dogs"));
}
 

/** 
 * Creates complex phrase query from the cached tokenstream contents 
 */
protected Query analyzeMultiPhrase(String field, TokenStream stream, int slop) throws IOException {
  MultiPhraseQuery.Builder mpqb = newMultiPhraseQueryBuilder();
  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();
}
 

@Override
public Query build(QueryNode queryNode) throws QueryNodeException {
  SlopQueryNode phraseSlopNode = (SlopQueryNode) queryNode;

  Query query = (Query) phraseSlopNode.getChild().getTag(
      QueryTreeBuilder.QUERY_TREE_BUILDER_TAGID);

  if (query instanceof PhraseQuery) {
    PhraseQuery.Builder builder = new PhraseQuery.Builder();
    builder.setSlop(phraseSlopNode.getValue());
    PhraseQuery pq = (PhraseQuery) query;
    org.apache.lucene.index.Term[] terms = pq.getTerms();
    int[] positions = pq.getPositions();
    for (int i = 0; i < terms.length; ++i) {
      builder.add(terms[i], positions[i]);
    }
    query = builder.build();

  } else {
    MultiPhraseQuery mpq = (MultiPhraseQuery)query;
    
    int slop = phraseSlopNode.getValue();
    
    if (slop != mpq.getSlop()) {
      query = new MultiPhraseQuery.Builder(mpq).setSlop(slop).build();
    }
  }

  return query;

}
 

public void testHighlighterWithMultiPhraseQuery() throws IOException, InvalidTokenOffsetsException {
  final String fieldName = "substring";

  final MultiPhraseQuery mpq = new MultiPhraseQuery.Builder()
      .add(new Term(fieldName, "uchu")).build();

  assertHighlighting(mpq, new SimpleHTMLFormatter("<b>", "</b>"), "Buchung", "B<b>uchu</b>ng", fieldName);
}
 

public void testQueryScorerMultiPhraseQueryHighlighting() throws Exception {
  MultiPhraseQuery.Builder mpqb = new MultiPhraseQuery.Builder();

  mpqb.add(new Term[] { new Term(FIELD_NAME, "wordx"), new Term(FIELD_NAME, "wordb") });
  mpqb.add(new Term(FIELD_NAME, "wordy"));

  doSearching(mpqb.build());

  final int maxNumFragmentsRequired = 2;
  assertExpectedHighlightCount(maxNumFragmentsRequired, 6);
}
 

@Override
public Query rewrite(IndexReader reader) throws IOException {
    if (getBoost() != 1.0F) {
        return super.rewrite(reader);
    }
    if (termArrays.isEmpty()) {
        return new MatchNoDocsQuery();
    }
    MultiPhraseQuery query = new MultiPhraseQuery();
    query.setSlop(slop);
    int sizeMinus1 = termArrays.size() - 1;
    for (int i = 0; i < sizeMinus1; i++) {
        query.add(termArrays.get(i), positions.get(i));
    }
    Term[] suffixTerms = termArrays.get(sizeMinus1);
    int position = positions.get(sizeMinus1);
    ObjectHashSet<Term> terms = new ObjectHashSet<>();
    for (Term term : suffixTerms) {
        getPrefixTerms(terms, term, reader);
        if (terms.size() > maxExpansions) {
            break;
        }
    }
    if (terms.isEmpty()) {
        return Queries.newMatchNoDocsQuery();
    }
    query.add(terms.toArray(Term.class), position);
    query.setBoost(getBoost());
    return query.rewrite(reader);
}
 

private Query getQuery() {
  try {
    
    switch (type) {
      case FIELD:  // fallthrough
      case PHRASE:
        Query query;
        if (val == null) {
          query = super.getFieldQuery(field, vals, false);
        } else {
          query = super.getFieldQuery(field, val, type == QType.PHRASE, false);
        }
        // Boolean query on a whitespace-separated string
        // If these were synonyms we would have a SynonymQuery
        if (query instanceof BooleanQuery) {
          if (type == QType.FIELD) { // Don't set mm for boolean query containing phrase queries
            BooleanQuery bq = (BooleanQuery) query;
            query = SolrPluginUtils.setMinShouldMatch(bq, minShouldMatch, false);
          }
        } else if (query instanceof PhraseQuery) {
          PhraseQuery pq = (PhraseQuery)query;
          if (minClauseSize > 1 && pq.getTerms().length < minClauseSize) return null;
          PhraseQuery.Builder builder = new PhraseQuery.Builder();
          Term[] terms = pq.getTerms();
          int[] positions = pq.getPositions();
          for (int i = 0; i < terms.length; ++i) {
            builder.add(terms[i], positions[i]);
          }
          builder.setSlop(slop);
          query = builder.build();
        } else if (query instanceof MultiPhraseQuery) {
          MultiPhraseQuery mpq = (MultiPhraseQuery)query;
          if (minClauseSize > 1 && mpq.getTermArrays().length < minClauseSize) return null;
          if (slop != mpq.getSlop()) {
            query = new MultiPhraseQuery.Builder(mpq).setSlop(slop).build();
          }
        } else if (query instanceof SpanQuery) {
          return query;
        } else if (minClauseSize > 1) {
          // if it's not a type of phrase query, it doesn't meet the minClauseSize requirements
          return null;
        }
        return query;
      case PREFIX: return super.getPrefixQuery(field, val);
      case WILDCARD: return super.getWildcardQuery(field, val);
      case FUZZY: return super.getFuzzyQuery(field, val, flt);
      case RANGE: return super.getRangeQuery(field, val, val2, bool, bool2);
    }
    return null;
    
  } catch (Exception e) {
    // an exception here is due to the field query not being compatible with the input text
    // for example, passing a string to a numeric field.
    return null;
  }
}
 

private static void assertEqualsMultiPhraseQuery(MultiPhraseQuery expected, MultiPhraseQuery actual) {
  throw new RuntimeException("Not Implemented");
}
 

/**
 * Builds a new MultiPhraseQuery instance.
 * <p>
 * This is intended for subclasses that wish to customize the generated queries.
 * @return new MultiPhraseQuery instance
 */
protected MultiPhraseQuery.Builder newMultiPhraseQueryBuilder() {
  return new MultiPhraseQuery.Builder();
}