Java Code Examples for org.apache.lucene.index.LogMergePolicy#setMergeFactor()

private static IndexWriter initWriter(int id, Random random, Path indexPath, boolean doCheckIndexOnClose) throws IOException {
  Directory dir = SimpleReplicaNode.getDirectory(random, id, indexPath, doCheckIndexOnClose);

  MockAnalyzer analyzer = new MockAnalyzer(random);
  analyzer.setMaxTokenLength(TestUtil.nextInt(random, 1, IndexWriter.MAX_TERM_LENGTH));
  IndexWriterConfig iwc = LuceneTestCase.newIndexWriterConfig(random, analyzer);

  MergePolicy mp = iwc.getMergePolicy();
  //iwc.setInfoStream(new PrintStreamInfoStream(System.out));

  // Force more frequent merging so we stress merge warming:
  if (mp instanceof TieredMergePolicy) {
    TieredMergePolicy tmp = (TieredMergePolicy) mp;
    tmp.setSegmentsPerTier(3);
    tmp.setMaxMergeAtOnce(3);
  } else if (mp instanceof LogMergePolicy) {
    LogMergePolicy lmp = (LogMergePolicy) mp;
    lmp.setMergeFactor(3);
  }

  IndexWriter writer = new IndexWriter(dir, iwc);

  TestUtil.reduceOpenFiles(writer);
  return writer;
}
 

/** just tries to configure things to keep the open file
 * count lowish */
public static void reduceOpenFiles(IndexWriter w) {
  // keep number of open files lowish
  MergePolicy mp = w.getConfig().getMergePolicy();
  mp.setNoCFSRatio(1.0);
  if (mp instanceof LogMergePolicy) {
    LogMergePolicy lmp = (LogMergePolicy) mp;
    lmp.setMergeFactor(Math.min(5, lmp.getMergeFactor()));
  } else if (mp instanceof TieredMergePolicy) {
    TieredMergePolicy tmp = (TieredMergePolicy) mp;
    tmp.setMaxMergeAtOnce(Math.min(5, tmp.getMaxMergeAtOnce()));
    tmp.setSegmentsPerTier(Math.min(5, tmp.getSegmentsPerTier()));
  }
  MergeScheduler ms = w.getConfig().getMergeScheduler();
  if (ms instanceof ConcurrentMergeScheduler) {
    // wtf... shouldnt it be even lower since it's 1 by default?!?!
    ((ConcurrentMergeScheduler) ms).setMaxMergesAndThreads(3, 2);
  }
}
 

public static IndexWriter getIndexWriter(String indexPath, boolean create) throws IOException {
    Directory dir = FSDirectory.open(Paths.get(indexPath));
    Analyzer analyzer = new SmartChineseAnalyzer();
    IndexWriterConfig iwc = new IndexWriterConfig(analyzer);
    LogMergePolicy mergePolicy = new LogByteSizeMergePolicy();
    mergePolicy.setMergeFactor(50);
    mergePolicy.setMaxMergeDocs(5000);
    if (create){
        iwc.setOpenMode(IndexWriterConfig.OpenMode.CREATE);
    } else {
        iwc.setOpenMode(IndexWriterConfig.OpenMode.CREATE_OR_APPEND);
    }
    return new IndexWriter(dir, iwc);
}
 

@Test
public void testOpenIfChangedManySegments() throws Exception {
  // test openIfChanged() when the taxonomy contains many segments
  Directory dir = newDirectory();
  
  DirectoryTaxonomyWriter writer = new DirectoryTaxonomyWriter(dir) {
    @Override
    protected IndexWriterConfig createIndexWriterConfig(OpenMode openMode) {
      IndexWriterConfig conf = super.createIndexWriterConfig(openMode);
      LogMergePolicy lmp = (LogMergePolicy) conf.getMergePolicy();
      lmp.setMergeFactor(2);
      return conf;
    }
  };
  TaxonomyReader reader = new DirectoryTaxonomyReader(writer);
  
  int numRounds = random().nextInt(10) + 10;
  int numCategories = 1; // one for root
  for (int i = 0; i < numRounds; i++) {
    int numCats = random().nextInt(4) + 1;
    for (int j = 0; j < numCats; j++) {
      writer.addCategory(new FacetLabel(Integer.toString(i), Integer.toString(j)));
    }
    numCategories += numCats + 1 /* one for round-parent */;
    TaxonomyReader newtr = TaxonomyReader.openIfChanged(reader);
    assertNotNull(newtr);
    reader.close();
    reader = newtr;
    
    // assert categories
    assertEquals(numCategories, reader.getSize());
    int roundOrdinal = reader.getOrdinal(new FacetLabel(Integer.toString(i)));
    int[] parents = reader.getParallelTaxonomyArrays().parents();
    assertEquals(0, parents[roundOrdinal]); // round's parent is root
    for (int j = 0; j < numCats; j++) {
      int ord = reader.getOrdinal(new FacetLabel(Integer.toString(i), Integer.toString(j)));
      assertEquals(roundOrdinal, parents[ord]); // round's parent is root
    }
  }
  
  reader.close();
  writer.close();
  dir.close();
}
 

public void testSubclassConcurrentMergeScheduler() throws IOException {
  MockDirectoryWrapper dir = newMockDirectory();
  dir.failOn(new FailOnlyOnMerge());

  Document doc = new Document();
  Field idField = newStringField("id", "", Field.Store.YES);
  doc.add(idField);

  IndexWriterConfig iwc = newIndexWriterConfig(new MockAnalyzer(random()))
    .setMergeScheduler(new MyMergeScheduler())
    .setMaxBufferedDocs(2).setRAMBufferSizeMB(IndexWriterConfig.DISABLE_AUTO_FLUSH)
    .setMergePolicy(newLogMergePolicy());

  ByteArrayOutputStream baos = new ByteArrayOutputStream();
  infoStream = new PrintStreamInfoStream(new PrintStream(baos, true, IOUtils.UTF_8));
  iwc.setInfoStream(infoStream);

  IndexWriter writer = new IndexWriter(dir, iwc);
  LogMergePolicy logMP = (LogMergePolicy) writer.getConfig().getMergePolicy();
  logMP.setMergeFactor(10);
  for(int i=0;i<20;i++) {
    writer.addDocument(doc);
  }

  try {
    ((MyMergeScheduler) writer.getConfig().getMergeScheduler()).sync();
  } catch (IllegalStateException ise) {
    // OK
  }
  writer.rollback();

  try {
    assertTrue(mergeThreadCreated);
    assertTrue(mergeCalled);
    assertTrue(excCalled);
  } catch (AssertionError ae) {
    System.out.println("TEST FAILED; IW infoStream output:");
    System.out.println(baos.toString(IOUtils.UTF_8));
    throw ae;
  }
  dir.close();
}
 

private void loadExternalFileDictionary(SolrCore core, SolrIndexSearcher searcher) {
  try {
    IndexSchema schema = null == searcher ? core.getLatestSchema() : searcher.getSchema();
    // Get the field's analyzer
    if (fieldTypeName != null && schema.getFieldTypeNoEx(fieldTypeName) != null) {
      FieldType fieldType = schema.getFieldTypes().get(fieldTypeName);
      // Do index-time analysis using the given fieldType's analyzer
      Directory ramDir = new ByteBuffersDirectory();

      LogMergePolicy mp = new LogByteSizeMergePolicy();
      mp.setMergeFactor(300);

      IndexWriter writer = new IndexWriter(
          ramDir,
          new IndexWriterConfig(fieldType.getIndexAnalyzer()).
              setMaxBufferedDocs(150).
              setMergePolicy(mp).
              setOpenMode(IndexWriterConfig.OpenMode.CREATE)
              // TODO: if we enable this, codec gets angry since field won't exist in the schema
              // .setCodec(core.getCodec())
      );

      List<String> lines = core.getResourceLoader().getLines(sourceLocation, characterEncoding);

      for (String s : lines) {
        Document d = new Document();
        d.add(new TextField(WORD_FIELD_NAME, s, Field.Store.NO));
        writer.addDocument(d);
      }
      writer.forceMerge(1);
      writer.close();

      dictionary = new HighFrequencyDictionary(DirectoryReader.open(ramDir),
              WORD_FIELD_NAME, 0.0f);
    } else {
      // check if character encoding is defined
      if (characterEncoding == null) {
        dictionary = new PlainTextDictionary(core.getResourceLoader().openResource(sourceLocation));
      } else {
        dictionary = new PlainTextDictionary(new InputStreamReader(core.getResourceLoader().openResource(sourceLocation), characterEncoding));
      }
    }


  } catch (IOException e) {
    log.error( "Unable to load spellings", e);
  }
}