Examples with RandomIndexWriter org.apache.lucene.index.RandomIndexWriter used on opensource projects

Example 26 with RandomIndexWriter

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

the class RecoverySourceHandlerTests method testHandleExceptinoOnSendSendFiles.

public void testHandleExceptinoOnSendSendFiles() throws Throwable {
    Settings settings = Settings.builder().put("indices.recovery.concurrent_streams", 1).put("indices.recovery.concurrent_small_file_streams", 1).build();
    final RecoverySettings recoverySettings = new RecoverySettings(settings, service);
    final StartRecoveryRequest request = new StartRecoveryRequest(shardId, new DiscoveryNode("b", buildNewFakeTransportAddress(), emptyMap(), emptySet(), Version.CURRENT), new DiscoveryNode("b", buildNewFakeTransportAddress(), emptyMap(), emptySet(), Version.CURRENT), null, randomBoolean(), randomNonNegativeLong(), randomBoolean() ? SequenceNumbersService.UNASSIGNED_SEQ_NO : 0L);
    Path tempDir = createTempDir();
    Store store = newStore(tempDir, false);
    AtomicBoolean failedEngine = new AtomicBoolean(false);
    RecoverySourceHandler handler = new RecoverySourceHandler(null, null, request, () -> 0L, e -> () -> {
    }, recoverySettings.getChunkSize().bytesAsInt(), Settings.EMPTY) {

        @Override
        protected void failEngine(IOException cause) {
            assertFalse(failedEngine.get());
            failedEngine.set(true);
        }
    };
    Directory dir = store.directory();
    RandomIndexWriter writer = new RandomIndexWriter(random(), dir, newIndexWriterConfig());
    int numDocs = randomIntBetween(10, 100);
    for (int i = 0; i < numDocs; i++) {
        Document document = new Document();
        document.add(new StringField("id", Integer.toString(i), Field.Store.YES));
        document.add(newField("field", randomUnicodeOfCodepointLengthBetween(1, 10), TextField.TYPE_STORED));
        writer.addDocument(document);
    }
    writer.commit();
    writer.close();
    Store.MetadataSnapshot metadata = store.getMetadata(null);
    List<StoreFileMetaData> metas = new ArrayList<>();
    for (StoreFileMetaData md : metadata) {
        metas.add(md);
    }
    final boolean throwCorruptedIndexException = randomBoolean();
    Store targetStore = newStore(createTempDir(), false);
    try {
        handler.sendFiles(store, metas.toArray(new StoreFileMetaData[0]), (md) -> {
            if (throwCorruptedIndexException) {
                throw new RuntimeException(new CorruptIndexException("foo", "bar"));
            } else {
                throw new RuntimeException("boom");
            }
        });
        fail("exception index");
    } catch (RuntimeException ex) {
        assertNull(ExceptionsHelper.unwrapCorruption(ex));
        if (throwCorruptedIndexException) {
            assertEquals(ex.getMessage(), "[File corruption occurred on recovery but checksums are ok]");
        } else {
            assertEquals(ex.getMessage(), "boom");
        }
    } catch (CorruptIndexException ex) {
        fail("not expected here");
    }
    assertFalse(failedEngine.get());
    IOUtils.close(store, targetStore);
}

Example 27 with RandomIndexWriter

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

the class CollapsingTopDocsCollectorTests method assertSearchCollapse.

private <T extends Comparable> void assertSearchCollapse(CollapsingDocValuesProducer<T> dvProducers, boolean numeric, boolean multivalued) throws IOException {
    final int numDocs = randomIntBetween(1000, 2000);
    int maxGroup = randomIntBetween(2, 500);
    final Directory dir = newDirectory();
    final RandomIndexWriter w = new RandomIndexWriter(random(), dir);
    Set<T> values = new HashSet<>();
    int totalHits = 0;
    for (int i = 0; i < numDocs; i++) {
        final T value = dvProducers.randomGroup(maxGroup);
        values.add(value);
        Document doc = new Document();
        dvProducers.add(doc, value, multivalued);
        doc.add(new NumericDocValuesField("sort1", randomIntBetween(0, 10)));
        doc.add(new NumericDocValuesField("sort2", randomLong()));
        w.addDocument(doc);
        totalHits++;
    }
    List<T> valueList = new ArrayList<>(values);
    Collections.sort(valueList);
    final IndexReader reader = w.getReader();
    final IndexSearcher searcher = newSearcher(reader);
    final SortField collapseField = dvProducers.sortField(multivalued);
    final SortField sort1 = new SortField("sort1", SortField.Type.INT);
    final SortField sort2 = new SortField("sort2", SortField.Type.LONG);
    Sort sort = new Sort(sort1, sort2, collapseField);
    int expectedNumGroups = values.size();
    final CollapsingTopDocsCollector collapsingCollector;
    if (numeric) {
        collapsingCollector = CollapsingTopDocsCollector.createNumeric(collapseField.getField(), sort, expectedNumGroups, false);
    } else {
        collapsingCollector = CollapsingTopDocsCollector.createKeyword(collapseField.getField(), sort, expectedNumGroups, false);
    }
    TopFieldCollector topFieldCollector = TopFieldCollector.create(sort, totalHits, true, false, false);
    searcher.search(new MatchAllDocsQuery(), collapsingCollector);
    searcher.search(new MatchAllDocsQuery(), topFieldCollector);
    CollapseTopFieldDocs collapseTopFieldDocs = collapsingCollector.getTopDocs();
    TopFieldDocs topDocs = topFieldCollector.topDocs();
    assertEquals(collapseField.getField(), collapseTopFieldDocs.field);
    assertEquals(expectedNumGroups, collapseTopFieldDocs.scoreDocs.length);
    assertEquals(totalHits, collapseTopFieldDocs.totalHits);
    assertEquals(totalHits, topDocs.scoreDocs.length);
    assertEquals(totalHits, topDocs.totalHits);
    Set<Object> seen = new HashSet<>();
    // collapse field is the last sort
    int collapseIndex = sort.getSort().length - 1;
    int topDocsIndex = 0;
    for (int i = 0; i < expectedNumGroups; i++) {
        FieldDoc fieldDoc = null;
        for (; topDocsIndex < totalHits; topDocsIndex++) {
            fieldDoc = (FieldDoc) topDocs.scoreDocs[topDocsIndex];
            if (seen.contains(fieldDoc.fields[collapseIndex]) == false) {
                break;
            }
        }
        FieldDoc collapseFieldDoc = (FieldDoc) collapseTopFieldDocs.scoreDocs[i];
        assertNotNull(fieldDoc);
        assertEquals(collapseFieldDoc.doc, fieldDoc.doc);
        assertArrayEquals(collapseFieldDoc.fields, fieldDoc.fields);
        seen.add(fieldDoc.fields[fieldDoc.fields.length - 1]);
    }
    for (; topDocsIndex < totalHits; topDocsIndex++) {
        FieldDoc fieldDoc = (FieldDoc) topDocs.scoreDocs[topDocsIndex];
        assertTrue(seen.contains(fieldDoc.fields[collapseIndex]));
    }
    // check merge
    final IndexReaderContext ctx = searcher.getTopReaderContext();
    final SegmentSearcher[] subSearchers;
    final int[] docStarts;
    if (ctx instanceof LeafReaderContext) {
        subSearchers = new SegmentSearcher[1];
        docStarts = new int[1];
        subSearchers[0] = new SegmentSearcher((LeafReaderContext) ctx, ctx);
        docStarts[0] = 0;
    } else {
        final CompositeReaderContext compCTX = (CompositeReaderContext) ctx;
        final int size = compCTX.leaves().size();
        subSearchers = new SegmentSearcher[size];
        docStarts = new int[size];
        int docBase = 0;
        for (int searcherIDX = 0; searcherIDX < subSearchers.length; searcherIDX++) {
            final LeafReaderContext leave = compCTX.leaves().get(searcherIDX);
            subSearchers[searcherIDX] = new SegmentSearcher(leave, compCTX);
            docStarts[searcherIDX] = docBase;
            docBase += leave.reader().maxDoc();
        }
    }
    final CollapseTopFieldDocs[] shardHits = new CollapseTopFieldDocs[subSearchers.length];
    final Weight weight = searcher.createNormalizedWeight(new MatchAllDocsQuery(), false);
    for (int shardIDX = 0; shardIDX < subSearchers.length; shardIDX++) {
        final SegmentSearcher subSearcher = subSearchers[shardIDX];
        final CollapsingTopDocsCollector c;
        if (numeric) {
            c = CollapsingTopDocsCollector.createNumeric(collapseField.getField(), sort, expectedNumGroups, false);
        } else {
            c = CollapsingTopDocsCollector.createKeyword(collapseField.getField(), sort, expectedNumGroups, false);
        }
        subSearcher.search(weight, c);
        shardHits[shardIDX] = c.getTopDocs();
    }
    CollapseTopFieldDocs mergedFieldDocs = CollapseTopFieldDocs.merge(sort, 0, expectedNumGroups, shardHits);
    assertTopDocsEquals(mergedFieldDocs, collapseTopFieldDocs);
    w.close();
    reader.close();
    dir.close();
}

Example 28 with RandomIndexWriter

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

the class CustomUnifiedHighlighterTests method assertHighlightOneDoc.

private void assertHighlightOneDoc(String fieldName, String[] inputs, Analyzer analyzer, Query query, Locale locale, BreakIterator breakIterator, int noMatchSize, String[] expectedPassages) throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig iwc = newIndexWriterConfig(analyzer);
    iwc.setMergePolicy(newTieredMergePolicy(random()));
    RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwc);
    FieldType ft = new FieldType(TextField.TYPE_STORED);
    ft.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS);
    ft.freeze();
    Document doc = new Document();
    for (String input : inputs) {
        Field field = new Field(fieldName, "", ft);
        field.setStringValue(input);
        doc.add(field);
    }
    iw.addDocument(doc);
    DirectoryReader reader = iw.getReader();
    IndexSearcher searcher = newSearcher(reader);
    iw.close();
    TopDocs topDocs = searcher.search(new MatchAllDocsQuery(), 1, Sort.INDEXORDER);
    assertThat(topDocs.totalHits, equalTo(1));
    String rawValue = Strings.arrayToDelimitedString(inputs, String.valueOf(MULTIVAL_SEP_CHAR));
    CustomUnifiedHighlighter highlighter = new CustomUnifiedHighlighter(searcher, analyzer, new CustomPassageFormatter("<b>", "</b>", new DefaultEncoder()), locale, breakIterator, rawValue, noMatchSize);
    highlighter.setFieldMatcher((name) -> "text".equals(name));
    final Snippet[] snippets = highlighter.highlightField("text", query, topDocs.scoreDocs[0].doc, expectedPassages.length);
    assertEquals(snippets.length, expectedPassages.length);
    for (int i = 0; i < snippets.length; i++) {
        assertEquals(snippets[i].getText(), expectedPassages[i]);
    }
    reader.close();
    dir.close();
}

Example 29 with RandomIndexWriter

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

the class MinDocQueryTests method testRandom.

public void testRandom() throws IOException {
    final int numDocs = randomIntBetween(10, 200);
    final Document doc = new Document();
    final Directory dir = newDirectory();
    final RandomIndexWriter w = new RandomIndexWriter(random(), dir);
    for (int i = 0; i < numDocs; ++i) {
        w.addDocument(doc);
    }
    final IndexReader reader = w.getReader();
    final IndexSearcher searcher = newSearcher(reader);
    for (int i = 0; i <= numDocs; ++i) {
        assertEquals(numDocs - i, searcher.count(new MinDocQuery(i)));
    }
    w.close();
    reader.close();
    dir.close();
}

Example 30 with RandomIndexWriter

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

the class CustomPostingsHighlighterTests method testCustomPostingsHighlighter.

public void testCustomPostingsHighlighter() throws Exception {
    Directory dir = newDirectory();
    IndexWriterConfig iwc = newIndexWriterConfig(new MockAnalyzer(random()));
    iwc.setMergePolicy(newLogMergePolicy());
    RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwc);
    FieldType offsetsType = new FieldType(TextField.TYPE_STORED);
    offsetsType.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS);
    //good position but only one match
    final String firstValue = "This is a test. Just a test1 highlighting from postings highlighter.";
    Field body = new Field("body", "", offsetsType);
    Document doc = new Document();
    doc.add(body);
    body.setStringValue(firstValue);
    //two matches, not the best snippet due to its length though
    final String secondValue = "This is the second highlighting value to perform highlighting on a longer text that gets scored lower.";
    Field body2 = new Field("body", "", offsetsType);
    doc.add(body2);
    body2.setStringValue(secondValue);
    //two matches and short, will be scored highest
    final String thirdValue = "This is highlighting the third short highlighting value.";
    Field body3 = new Field("body", "", offsetsType);
    doc.add(body3);
    body3.setStringValue(thirdValue);
    //one match, same as first but at the end, will be scored lower due to its position
    final String fourthValue = "Just a test4 highlighting from postings highlighter.";
    Field body4 = new Field("body", "", offsetsType);
    doc.add(body4);
    body4.setStringValue(fourthValue);
    iw.addDocument(doc);
    IndexReader ir = iw.getReader();
    iw.close();
    String firstHlValue = "Just a test1 <b>highlighting</b> from postings highlighter.";
    String secondHlValue = "This is the second <b>highlighting</b> value to perform <b>highlighting</b> on a longer text that gets scored lower.";
    String thirdHlValue = "This is <b>highlighting</b> the third short <b>highlighting</b> value.";
    String fourthHlValue = "Just a test4 <b>highlighting</b> from postings highlighter.";
    IndexSearcher searcher = newSearcher(ir);
    Query query = new TermQuery(new Term("body", "highlighting"));
    TopDocs topDocs = searcher.search(query, 10, Sort.INDEXORDER);
    assertThat(topDocs.totalHits, equalTo(1));
    int docId = topDocs.scoreDocs[0].doc;
    String fieldValue = firstValue + HighlightUtils.PARAGRAPH_SEPARATOR + secondValue + HighlightUtils.PARAGRAPH_SEPARATOR + thirdValue + HighlightUtils.PARAGRAPH_SEPARATOR + fourthValue;
    CustomPostingsHighlighter highlighter = new CustomPostingsHighlighter(null, new CustomPassageFormatter("<b>", "</b>", new DefaultEncoder()), fieldValue, false);
    Snippet[] snippets = highlighter.highlightField("body", query, searcher, docId, 5);
    assertThat(snippets.length, equalTo(4));
    assertThat(snippets[0].getText(), equalTo(firstHlValue));
    assertThat(snippets[1].getText(), equalTo(secondHlValue));
    assertThat(snippets[2].getText(), equalTo(thirdHlValue));
    assertThat(snippets[3].getText(), equalTo(fourthHlValue));
    ir.close();
    dir.close();
}