Examples with MemoryIndex org.apache.lucene.index.memory.MemoryIndex used on opensource projects

Example 1 with MemoryIndex

use of org.apache.lucene.index.memory.MemoryIndex in project elasticsearch by elastic.

the class PercolateQueryBuilder method doToQuery.

@Override
protected Query doToQuery(QueryShardContext context) throws IOException {
    // Call nowInMillis() so that this query becomes un-cacheable since we
    // can't be sure that it doesn't use now or scripts
    context.nowInMillis();
    if (indexedDocumentIndex != null || indexedDocumentType != null || indexedDocumentId != null) {
        throw new IllegalStateException("query builder must be rewritten first");
    }
    if (document == null) {
        throw new IllegalStateException("no document to percolate");
    }
    MapperService mapperService = context.getMapperService();
    DocumentMapperForType docMapperForType = mapperService.documentMapperWithAutoCreate(documentType);
    DocumentMapper docMapper = docMapperForType.getDocumentMapper();
    ParsedDocument doc = docMapper.parse(source(context.index().getName(), documentType, "_temp_id", document, documentXContentType));
    FieldNameAnalyzer fieldNameAnalyzer = (FieldNameAnalyzer) docMapper.mappers().indexAnalyzer();
    // Need to this custom impl because FieldNameAnalyzer is strict and the percolator sometimes isn't when
    // 'index.percolator.map_unmapped_fields_as_string' is enabled:
    Analyzer analyzer = new DelegatingAnalyzerWrapper(Analyzer.PER_FIELD_REUSE_STRATEGY) {

        @Override
        protected Analyzer getWrappedAnalyzer(String fieldName) {
            Analyzer analyzer = fieldNameAnalyzer.analyzers().get(fieldName);
            if (analyzer != null) {
                return analyzer;
            } else {
                return context.getIndexAnalyzers().getDefaultIndexAnalyzer();
            }
        }
    };
    final IndexSearcher docSearcher;
    if (doc.docs().size() > 1) {
        assert docMapper.hasNestedObjects();
        docSearcher = createMultiDocumentSearcher(analyzer, doc);
    } else {
        MemoryIndex memoryIndex = MemoryIndex.fromDocument(doc.rootDoc(), analyzer, true, false);
        docSearcher = memoryIndex.createSearcher();
        docSearcher.setQueryCache(null);
    }
    Version indexVersionCreated = context.getIndexSettings().getIndexVersionCreated();
    boolean mapUnmappedFieldsAsString = context.getIndexSettings().getValue(PercolatorFieldMapper.INDEX_MAP_UNMAPPED_FIELDS_AS_STRING_SETTING);
    // We have to make a copy of the QueryShardContext here so we can have a unfrozen version for parsing the legacy
    // percolator queries
    QueryShardContext percolateShardContext = new QueryShardContext(context);
    MappedFieldType fieldType = context.fieldMapper(field);
    if (fieldType == null) {
        throw new QueryShardException(context, "field [" + field + "] does not exist");
    }
    if (!(fieldType instanceof PercolatorFieldMapper.FieldType)) {
        throw new QueryShardException(context, "expected field [" + field + "] to be of type [percolator], but is of type [" + fieldType.typeName() + "]");
    }
    PercolatorFieldMapper.FieldType pft = (PercolatorFieldMapper.FieldType) fieldType;
    PercolateQuery.QueryStore queryStore = createStore(pft, percolateShardContext, mapUnmappedFieldsAsString);
    return pft.percolateQuery(documentType, queryStore, document, docSearcher);
}

Example 2 with MemoryIndex

use of org.apache.lucene.index.memory.MemoryIndex in project elasticsearch by elastic.

the class CandidateQueryTests method testDuel.

public void testDuel() throws Exception {
    List<Function<String, Query>> queryFunctions = new ArrayList<>();
    queryFunctions.add((id) -> new PrefixQuery(new Term("field", id)));
    queryFunctions.add((id) -> new WildcardQuery(new Term("field", id + "*")));
    queryFunctions.add((id) -> new CustomQuery(new Term("field", id)));
    queryFunctions.add((id) -> new SpanTermQuery(new Term("field", id)));
    queryFunctions.add((id) -> new TermQuery(new Term("field", id)));
    queryFunctions.add((id) -> {
        BooleanQuery.Builder builder = new BooleanQuery.Builder();
        return builder.build();
    });
    queryFunctions.add((id) -> {
        BooleanQuery.Builder builder = new BooleanQuery.Builder();
        builder.add(new TermQuery(new Term("field", id)), BooleanClause.Occur.MUST);
        if (randomBoolean()) {
            builder.add(new MatchNoDocsQuery("no reason"), BooleanClause.Occur.MUST_NOT);
        }
        if (randomBoolean()) {
            builder.add(new CustomQuery(new Term("field", id)), BooleanClause.Occur.MUST);
        }
        return builder.build();
    });
    queryFunctions.add((id) -> {
        BooleanQuery.Builder builder = new BooleanQuery.Builder();
        builder.add(new TermQuery(new Term("field", id)), BooleanClause.Occur.SHOULD);
        if (randomBoolean()) {
            builder.add(new MatchNoDocsQuery("no reason"), BooleanClause.Occur.MUST_NOT);
        }
        if (randomBoolean()) {
            builder.add(new CustomQuery(new Term("field", id)), BooleanClause.Occur.SHOULD);
        }
        return builder.build();
    });
    queryFunctions.add((id) -> {
        BooleanQuery.Builder builder = new BooleanQuery.Builder();
        builder.add(new MatchAllDocsQuery(), BooleanClause.Occur.MUST);
        builder.add(new MatchAllDocsQuery(), BooleanClause.Occur.MUST);
        if (randomBoolean()) {
            builder.add(new MatchNoDocsQuery("no reason"), BooleanClause.Occur.MUST_NOT);
        }
        return builder.build();
    });
    queryFunctions.add((id) -> {
        BooleanQuery.Builder builder = new BooleanQuery.Builder();
        builder.add(new MatchAllDocsQuery(), BooleanClause.Occur.SHOULD);
        builder.add(new MatchAllDocsQuery(), BooleanClause.Occur.SHOULD);
        if (randomBoolean()) {
            builder.add(new MatchNoDocsQuery("no reason"), BooleanClause.Occur.MUST_NOT);
        }
        return builder.build();
    });
    queryFunctions.add((id) -> {
        BooleanQuery.Builder builder = new BooleanQuery.Builder();
        builder.setMinimumNumberShouldMatch(randomIntBetween(0, 4));
        builder.add(new TermQuery(new Term("field", id)), BooleanClause.Occur.SHOULD);
        builder.add(new CustomQuery(new Term("field", id)), BooleanClause.Occur.SHOULD);
        return builder.build();
    });
    queryFunctions.add((id) -> new MatchAllDocsQuery());
    queryFunctions.add((id) -> new MatchNoDocsQuery("no reason at all"));
    int numDocs = randomIntBetween(queryFunctions.size(), queryFunctions.size() * 3);
    List<ParseContext.Document> documents = new ArrayList<>();
    for (int i = 0; i < numDocs; i++) {
        String id = Integer.toString(i);
        Query query = queryFunctions.get(i % queryFunctions.size()).apply(id);
        addQuery(query, documents);
    }
    indexWriter.addDocuments(documents);
    indexWriter.close();
    directoryReader = DirectoryReader.open(directory);
    IndexSearcher shardSearcher = newSearcher(directoryReader);
    // Disable query cache, because ControlQuery cannot be cached...
    shardSearcher.setQueryCache(null);
    for (int i = 0; i < numDocs; i++) {
        String id = Integer.toString(i);
        Iterable<? extends IndexableField> doc = Collections.singleton(new StringField("field", id, Field.Store.NO));
        MemoryIndex memoryIndex = MemoryIndex.fromDocument(doc, new WhitespaceAnalyzer());
        duelRun(queryStore, memoryIndex, shardSearcher);
    }
    Iterable<? extends IndexableField> doc = Collections.singleton(new StringField("field", "value", Field.Store.NO));
    MemoryIndex memoryIndex = MemoryIndex.fromDocument(doc, new WhitespaceAnalyzer());
    duelRun(queryStore, memoryIndex, shardSearcher);
    // Empty percolator doc:
    memoryIndex = new MemoryIndex();
    duelRun(queryStore, memoryIndex, shardSearcher);
}

Example 3 with MemoryIndex

use of org.apache.lucene.index.memory.MemoryIndex in project elasticsearch by elastic.

the class PercolateQueryTests method testPercolateQuery.

public void testPercolateQuery() throws Exception {
    List<Iterable<? extends IndexableField>> docs = new ArrayList<>();
    List<Query> queries = new ArrayList<>();
    PercolateQuery.QueryStore queryStore = ctx -> queries::get;
    queries.add(new TermQuery(new Term("field", "fox")));
    docs.add(Collections.singleton(new StringField("select", "a", Field.Store.NO)));
    SpanNearQuery.Builder snp = new SpanNearQuery.Builder("field", true);
    snp.addClause(new SpanTermQuery(new Term("field", "jumps")));
    snp.addClause(new SpanTermQuery(new Term("field", "lazy")));
    snp.addClause(new SpanTermQuery(new Term("field", "dog")));
    snp.setSlop(2);
    queries.add(snp.build());
    docs.add(Collections.singleton(new StringField("select", "b", Field.Store.NO)));
    PhraseQuery.Builder pq1 = new PhraseQuery.Builder();
    pq1.add(new Term("field", "quick"));
    pq1.add(new Term("field", "brown"));
    pq1.add(new Term("field", "jumps"));
    pq1.setSlop(1);
    queries.add(pq1.build());
    docs.add(Collections.singleton(new StringField("select", "b", Field.Store.NO)));
    BooleanQuery.Builder bq1 = new BooleanQuery.Builder();
    bq1.add(new TermQuery(new Term("field", "quick")), BooleanClause.Occur.MUST);
    bq1.add(new TermQuery(new Term("field", "brown")), BooleanClause.Occur.MUST);
    bq1.add(new TermQuery(new Term("field", "fox")), BooleanClause.Occur.MUST);
    queries.add(bq1.build());
    docs.add(Collections.singleton(new StringField("select", "b", Field.Store.NO)));
    indexWriter.addDocuments(docs);
    indexWriter.close();
    directoryReader = DirectoryReader.open(directory);
    IndexSearcher shardSearcher = newSearcher(directoryReader);
    MemoryIndex memoryIndex = new MemoryIndex();
    memoryIndex.addField("field", "the quick brown fox jumps over the lazy dog", new WhitespaceAnalyzer());
    IndexSearcher percolateSearcher = memoryIndex.createSearcher();
    // no scoring, wrapping it in a constant score query:
    Query query = new ConstantScoreQuery(new PercolateQuery("type", queryStore, new BytesArray("a"), new TermQuery(new Term("select", "a")), percolateSearcher, new MatchNoDocsQuery("")));
    TopDocs topDocs = shardSearcher.search(query, 10);
    assertThat(topDocs.totalHits, equalTo(1));
    assertThat(topDocs.scoreDocs.length, equalTo(1));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(0));
    Explanation explanation = shardSearcher.explain(query, 0);
    assertThat(explanation.isMatch(), is(true));
    assertThat(explanation.getValue(), equalTo(topDocs.scoreDocs[0].score));
    query = new ConstantScoreQuery(new PercolateQuery("type", queryStore, new BytesArray("b"), new TermQuery(new Term("select", "b")), percolateSearcher, new MatchNoDocsQuery("")));
    topDocs = shardSearcher.search(query, 10);
    assertThat(topDocs.totalHits, equalTo(3));
    assertThat(topDocs.scoreDocs.length, equalTo(3));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(1));
    explanation = shardSearcher.explain(query, 1);
    assertThat(explanation.isMatch(), is(true));
    assertThat(explanation.getValue(), equalTo(topDocs.scoreDocs[0].score));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(2));
    explanation = shardSearcher.explain(query, 2);
    assertThat(explanation.isMatch(), is(true));
    assertThat(explanation.getValue(), equalTo(topDocs.scoreDocs[1].score));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(3));
    explanation = shardSearcher.explain(query, 2);
    assertThat(explanation.isMatch(), is(true));
    assertThat(explanation.getValue(), equalTo(topDocs.scoreDocs[2].score));
    query = new ConstantScoreQuery(new PercolateQuery("type", queryStore, new BytesArray("c"), new MatchAllDocsQuery(), percolateSearcher, new MatchAllDocsQuery()));
    topDocs = shardSearcher.search(query, 10);
    assertThat(topDocs.totalHits, equalTo(4));
    query = new PercolateQuery("type", queryStore, new BytesArray("{}"), new TermQuery(new Term("select", "b")), percolateSearcher, new MatchNoDocsQuery(""));
    topDocs = shardSearcher.search(query, 10);
    assertThat(topDocs.totalHits, equalTo(3));
    assertThat(topDocs.scoreDocs.length, equalTo(3));
    assertThat(topDocs.scoreDocs[0].doc, equalTo(3));
    explanation = shardSearcher.explain(query, 3);
    assertThat(explanation.isMatch(), is(true));
    assertThat(explanation.getValue(), equalTo(topDocs.scoreDocs[0].score));
    assertThat(explanation.getDetails(), arrayWithSize(1));
    assertThat(topDocs.scoreDocs[1].doc, equalTo(2));
    explanation = shardSearcher.explain(query, 2);
    assertThat(explanation.isMatch(), is(true));
    assertThat(explanation.getValue(), equalTo(topDocs.scoreDocs[1].score));
    assertThat(explanation.getDetails(), arrayWithSize(1));
    assertThat(topDocs.scoreDocs[2].doc, equalTo(1));
    explanation = shardSearcher.explain(query, 1);
    assertThat(explanation.isMatch(), is(true));
    assertThat(explanation.getValue(), equalTo(topDocs.scoreDocs[2].score));
    assertThat(explanation.getDetails(), arrayWithSize(1));
}

Example 4 with MemoryIndex

use of org.apache.lucene.index.memory.MemoryIndex in project elasticsearch by elastic.

the class PercolatorFieldMapperTests method testCreateCandidateQuery.

public void testCreateCandidateQuery() throws Exception {
    addQueryMapping();
    MemoryIndex memoryIndex = new MemoryIndex(false);
    memoryIndex.addField("field1", "the quick brown fox jumps over the lazy dog", new WhitespaceAnalyzer());
    memoryIndex.addField("field2", "some more text", new WhitespaceAnalyzer());
    memoryIndex.addField("_field3", "unhide me", new WhitespaceAnalyzer());
    memoryIndex.addField("field4", "123", new WhitespaceAnalyzer());
    memoryIndex.addField(new LongPoint("number_field", 10L), new WhitespaceAnalyzer());
    IndexReader indexReader = memoryIndex.createSearcher().getIndexReader();
    BooleanQuery candidateQuery = (BooleanQuery) fieldType.createCandidateQuery(indexReader);
    assertEquals(2, candidateQuery.clauses().size());
    assertEquals(Occur.SHOULD, candidateQuery.clauses().get(0).getOccur());
    TermInSetQuery termsQuery = (TermInSetQuery) candidateQuery.clauses().get(0).getQuery();
    PrefixCodedTerms terms = termsQuery.getTermData();
    assertThat(terms.size(), equalTo(14L));
    PrefixCodedTerms.TermIterator termIterator = terms.iterator();
    assertTermIterator(termIterator, "_field3me", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "_field3unhide", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field1brown", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field1dog", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field1fox", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field1jumps", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field1lazy", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field1over", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field1quick", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field1the", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field2more", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field2some", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field2text", fieldType.queryTermsField.name());
    assertTermIterator(termIterator, "field4123", fieldType.queryTermsField.name());
    assertEquals(Occur.SHOULD, candidateQuery.clauses().get(1).getOccur());
    assertEquals(new TermQuery(new Term(fieldType.extractionResultField.name(), EXTRACTION_FAILED)), candidateQuery.clauses().get(1).getQuery());
}

Example 5 with MemoryIndex

use of org.apache.lucene.index.memory.MemoryIndex in project orientdb by orientechnologies.

the class OLuceneTextOperator method evaluateRecord.

@Override
public Object evaluateRecord(OIdentifiable iRecord, ODocument iCurrentResult, OSQLFilterCondition iCondition, Object iLeft, Object iRight, OCommandContext iContext) {
    OLuceneFullTextIndex index = involvedIndex(iRecord, iCurrentResult, iCondition, iLeft, iRight);
    if (index == null) {
        throw new OCommandExecutionException("Cannot evaluate lucene condition without index configuration.");
    }
    MemoryIndex memoryIndex = (MemoryIndex) iContext.getVariable(MEMORY_INDEX);
    if (memoryIndex == null) {
        memoryIndex = new MemoryIndex();
        iContext.setVariable(MEMORY_INDEX, memoryIndex);
    }
    memoryIndex.reset();
    try {
        for (IndexableField field : index.buildDocument(iLeft).getFields()) {
            memoryIndex.addField(field.name(), field.tokenStream(index.indexAnalyzer(), null));
        }
        return memoryIndex.search(index.buildQuery(iRight)) > 0.0f;
    } catch (ParseException e) {
        OLogManager.instance().error(this, "error occurred while building query", e);
    } catch (IOException e) {
        OLogManager.instance().error(this, "error occurred while building memory index", e);
    }
    return null;
}