Examples with ByteBuffersDirectory org.apache.lucene.store.ByteBuffersDirectory used on opensource projects

Example 11 with ByteBuffersDirectory

use of org.apache.lucene.store.ByteBuffersDirectory in project jena by apache.

the class AbstractTestDatasetWithLuceneGraphTextIndex method init.

@Before
public void init() {
    Dataset ds1 = TDBFactory.createDataset();
    Directory dir = new ByteBuffersDirectory();
    EntityDefinition eDef = new EntityDefinition("iri", "text");
    eDef.setGraphField("graph");
    eDef.setPrimaryPredicate(RDFS.label);
    // some tests require indexing rdfs:comment
    eDef.set("comment", RDFS.comment.asNode());
    TextIndex tidx = new TextIndexLucene(dir, new TextIndexConfig(eDef));
    dataset = TextDatasetFactory.create(ds1, tidx);
}

Example 12 with ByteBuffersDirectory

use of org.apache.lucene.store.ByteBuffersDirectory in project crate by crate.

the class LuceneBatchIteratorBenchmark method createLuceneBatchIterator.

@Setup
public void createLuceneBatchIterator() throws Exception {
    IndexWriter iw = new IndexWriter(new ByteBuffersDirectory(), new IndexWriterConfig(new StandardAnalyzer()));
    String columnName = "x";
    for (int i = 0; i < 10_000_000; i++) {
        Document doc = new Document();
        doc.add(new NumericDocValuesField(columnName, i));
        iw.addDocument(doc);
    }
    iw.commit();
    iw.forceMerge(1, true);
    indexSearcher = new IndexSearcher(DirectoryReader.open(iw));
    IntegerColumnReference columnReference = new IntegerColumnReference(columnName);
    columnRefs = Collections.singletonList(columnReference);
    collectorContext = new CollectorContext();
}

Example 13 with ByteBuffersDirectory

use of org.apache.lucene.store.ByteBuffersDirectory in project crate by crate.

the class LuceneOrderedDocCollectorTest method testSearchWithScores.

@Test
public void testSearchWithScores() throws Exception {
    IndexWriter w = new IndexWriter(new ByteBuffersDirectory(), new IndexWriterConfig(new KeywordAnalyzer()));
    FieldType fieldType = KeywordFieldMapper.Defaults.FIELD_TYPE;
    for (int i = 0; i < 3; i++) {
        addDoc(w, "x", fieldType, "Arthur");
    }
    // not "Arthur" to lower score
    addDoc(w, "x", fieldType, "Arthur");
    w.commit();
    IndexSearcher searcher = new IndexSearcher(DirectoryReader.open(w, true, true));
    List<LuceneCollectorExpression<?>> columnReferences = Collections.singletonList(new ScoreCollectorExpression());
    Query query = new ConstantScoreQuery(new TermQuery(new Term("x", new BytesRef("Arthur"))));
    LuceneOrderedDocCollector collector = collector(searcher, columnReferences, query, null, true);
    KeyIterable<ShardId, Row> result = collector.collect();
    assertThat(StreamSupport.stream(result.spliterator(), false).count(), is(2L));
    Iterator<Row> values = result.iterator();
    assertThat(values.next().get(0), Matchers.is(1.0F));
    assertThat(values.next().get(0), Matchers.is(1.0F));
}

Example 14 with ByteBuffersDirectory

use of org.apache.lucene.store.ByteBuffersDirectory in project crate by crate.

the class LuceneOrderedDocCollectorTest method testSearchMoreAppliesMinScoreFilter.

@Test
public void testSearchMoreAppliesMinScoreFilter() throws Exception {
    IndexWriter w = new IndexWriter(new ByteBuffersDirectory(), new IndexWriterConfig(new KeywordAnalyzer()));
    var keywordFieldType = new KeywordFieldMapper.KeywordFieldType("x");
    var fieldType = KeywordFieldMapper.Defaults.FIELD_TYPE;
    for (int i = 0; i < 3; i++) {
        addDoc(w, "x", fieldType, "Arthur");
    }
    // not "Arthur" to lower score
    addDoc(w, "x", fieldType, "Arthurr");
    w.commit();
    IndexSearcher searcher = new IndexSearcher(DirectoryReader.open(w, true, true));
    List<LuceneCollectorExpression<?>> columnReferences = Collections.singletonList(new ScoreCollectorExpression());
    Query query = new FuzzyQuery(new Term("x", "Arthur"), Fuzziness.AUTO.asDistance("Arthur"), 2, 3, true);
    LuceneOrderedDocCollector collector;
    // without minScore filter we get 2 and 2 docs - this is not necessary for the test but is here
    // to make sure the "FuzzyQuery" matches the right documents
    collector = collector(searcher, columnReferences, query, null, true);
    assertThat(StreamSupport.stream(collector.collect().spliterator(), false).count(), is(2L));
    assertThat(StreamSupport.stream(collector.collect().spliterator(), false).count(), is(2L));
    collector = collector(searcher, columnReferences, query, 0.15f, true);
    int count = 0;
    // initialSearch -> 2 rows
    for (Row row : collector.collect()) {
        assertThat((float) row.get(0), Matchers.greaterThanOrEqualTo(0.15f));
        count++;
    }
    assertThat(count, is(2));
    count = 0;
    // searchMore -> 1 row is below minScore
    for (Row row : collector.collect()) {
        assertThat((float) row.get(0), Matchers.greaterThanOrEqualTo(0.15f));
        count++;
    }
    assertThat(count, is(1));
}

Example 15 with ByteBuffersDirectory

use of org.apache.lucene.store.ByteBuffersDirectory in project crate by crate.

the class LuceneOrderedDocCollectorTest method testSearchNoScores.

@Test
public void testSearchNoScores() throws Exception {
    IndexWriter w = new IndexWriter(new ByteBuffersDirectory(), new IndexWriterConfig(new KeywordAnalyzer()));
    String name = "x";
    var keywordFieldType = new KeywordFieldMapper.KeywordFieldType(name);
    var fieldType = KeywordFieldMapper.Defaults.FIELD_TYPE;
    for (int i = 0; i < 3; i++) {
        addDoc(w, name, fieldType, "Arthur");
    }
    // not "Arthur" to lower score
    addDoc(w, name, fieldType, "Arthur");
    w.commit();
    IndexSearcher searcher = new IndexSearcher(DirectoryReader.open(w, true, true));
    List<LuceneCollectorExpression<?>> columnReferences = Collections.singletonList(new ScoreCollectorExpression());
    Query query = new TermQuery(new Term(name, new BytesRef("Arthur")));
    LuceneOrderedDocCollector collector = collector(searcher, columnReferences, query, null, false);
    KeyIterable<ShardId, Row> result = collector.collect();
    assertThat(StreamSupport.stream(result.spliterator(), false).count(), is(2L));
    Iterator<Row> values = result.iterator();
    assertThat(values.next().get(0), Matchers.is(Float.NaN));
    assertThat(values.next().get(0), Matchers.is(Float.NaN));
}