virtual void doWork()
{
IndexReaderPtr r = IndexReader::open(directory, true);
if (r->numDocs() != 100)
BOOST_FAIL("num docs failure");
r->close();
}
TEST_F(IndexWriterReaderTest, testAddIndexesAndDoDeletesThreads) {
int32_t numIter = 5;
int32_t numDirs = 3;
DirectoryPtr mainDir = newLucene<MockRAMDirectory>();
IndexWriterPtr mainWriter = newLucene<IndexWriter>(mainDir, newLucene<WhitespaceAnalyzer>(), IndexWriter::MaxFieldLengthLIMITED);
AddDirectoriesThreadsPtr addDirThreads = newLucene<AddDirectoriesThreads>(numIter, mainWriter);
addDirThreads->launchThreads(numDirs);
addDirThreads->joinThreads();
EXPECT_EQ(addDirThreads->count->intValue(), addDirThreads->mainWriter->numDocs());
addDirThreads->close(true);
EXPECT_TRUE(addDirThreads->failures.empty());
checkIndex(mainDir);
IndexReaderPtr reader = IndexReader::open(mainDir, true);
EXPECT_EQ(addDirThreads->count->intValue(), reader->numDocs());
reader->close();
addDirThreads->closeDir();
mainDir->close();
}
static void verifyNumDocs(DirectoryPtr dir, int32_t numDocs)
{
IndexReaderPtr reader = IndexReader::open(dir, true);
BOOST_CHECK_EQUAL(reader->maxDoc(), numDocs);
BOOST_CHECK_EQUAL(reader->numDocs(), numDocs);
reader->close();
}
// Run one indexer and 2 searchers against single index as stress test.
static void runTest(DirectoryPtr directory)
{
Collection<TimedThreadPtr> threads(Collection<TimedThreadPtr>::newInstance(4));
AnalyzerPtr analyzer = newLucene<SimpleAnalyzer>();
IndexWriterPtr writer = newLucene<MockIndexWriter>(directory, analyzer, true, IndexWriter::MaxFieldLengthUNLIMITED);
writer->setMaxBufferedDocs(7);
writer->setMergeFactor(3);
// Establish a base index of 100 docs
for (int32_t i = 0; i < 100; ++i)
{
DocumentPtr d = newLucene<Document>();
d->add(newLucene<Field>(L"id", StringUtils::toString(i), Field::STORE_YES, Field::INDEX_NOT_ANALYZED));
d->add(newLucene<Field>(L"contents", intToEnglish(i), Field::STORE_NO, Field::INDEX_ANALYZED));
if ((i - 1) % 7 == 0)
writer->commit();
writer->addDocument(d);
}
writer->commit();
IndexReaderPtr r = IndexReader::open(directory, true);
BOOST_CHECK_EQUAL(100, r->numDocs());
r->close();
IndexerThreadPtr indexerThread1 = newLucene<IndexerThread>(writer);
threads[0] = indexerThread1;
indexerThread1->start();
IndexerThreadPtr indexerThread2 = newLucene<IndexerThread>(writer);
threads[1] = indexerThread2;
indexerThread2->start();
SearcherThreadPtr searcherThread1 = newLucene<SearcherThread>(directory);
threads[2] = searcherThread1;
searcherThread1->start();
SearcherThreadPtr searcherThread2 = newLucene<SearcherThread>(directory);
threads[3] = searcherThread2;
searcherThread2->start();
indexerThread1->join();
indexerThread2->join();
searcherThread1->join();
searcherThread2->join();
writer->close();
BOOST_CHECK(!indexerThread1->failed); // hit unexpected exception in indexer1
BOOST_CHECK(!indexerThread2->failed); // hit unexpected exception in indexer2
BOOST_CHECK(!searcherThread1->failed); // hit unexpected exception in search1
BOOST_CHECK(!searcherThread2->failed); // hit unexpected exception in search2
}
void runTest(DirectoryPtr directory, MergeSchedulerPtr merger)
{
IndexWriterPtr writer = newLucene<IndexWriter>(directory, analyzer, true, IndexWriter::MaxFieldLengthUNLIMITED);
writer->setMaxBufferedDocs(2);
if (merger)
writer->setMergeScheduler(merger);
for (int32_t iter = 0; iter < NUM_ITER; ++iter)
{
int32_t iterFinal = iter;
writer->setMergeFactor(1000);
for (int32_t i = 0; i < 200; ++i)
{
DocumentPtr d = newLucene<Document>();
d->add(newLucene<Field>(L"id", StringUtils::toString(i), Field::STORE_YES, Field::INDEX_NOT_ANALYZED));
d->add(newLucene<Field>(L"contents", intToEnglish(i), Field::STORE_NO, Field::INDEX_ANALYZED));
writer->addDocument(d);
}
writer->setMergeFactor(4);
Collection<LuceneThreadPtr> threads = Collection<LuceneThreadPtr>::newInstance(NUM_THREADS);
for (int32_t i = 0; i < NUM_THREADS; ++i)
{
int32_t iFinal = i;
IndexWriterPtr writerFinal = writer;
threads[i] = newLucene<OptimizeThread>(NUM_ITER2, iterFinal, iFinal, writer, writerFinal);
}
for (int32_t i = 0; i < NUM_THREADS; ++i)
threads[i]->start();
for (int32_t i = 0; i < NUM_THREADS; ++i)
threads[i]->join();
int32_t expectedDocCount = (int32_t)((1 + iter) * (200 + 8 * NUM_ITER2 * (int32_t)(((double)NUM_THREADS / 2.0) * (double)(1 + NUM_THREADS))));
BOOST_CHECK_EQUAL(expectedDocCount, writer->maxDoc());
writer->close();
writer = newLucene<IndexWriter>(directory, analyzer, false, IndexWriter::MaxFieldLengthUNLIMITED);
writer->setMaxBufferedDocs(2);
IndexReaderPtr reader = IndexReader::open(directory, true);
BOOST_CHECK(reader->isOptimized());
BOOST_CHECK_EQUAL(expectedDocCount, reader->numDocs());
reader->close();
}
writer->close();
}
void testTermVectors()
{
int32_t numDocs = reader->numDocs();
int64_t start = 0;
for (int32_t docId = 0; docId < numDocs; ++docId)
{
start = MiscUtils::currentTimeMillis();
Collection<TermFreqVectorPtr> vectors = reader->getTermFreqVectors(docId);
timeElapsed += MiscUtils::currentTimeMillis() - start;
// verify vectors result
verifyVectors(vectors, docId);
start = MiscUtils::currentTimeMillis();
TermFreqVectorPtr vector = reader->getTermFreqVector(docId, L"field");
timeElapsed += MiscUtils::currentTimeMillis() - start;
vectors = newCollection<TermFreqVectorPtr>(vector);
verifyVectors(vectors, docId);
}
}
void doTestUndeleteAll()
{
sis->read(dir);
IndexReaderPtr reader = openReader();
BOOST_CHECK(reader);
BOOST_CHECK_EQUAL(2, reader->numDocs());
reader->deleteDocument(0);
BOOST_CHECK_EQUAL(1, reader->numDocs());
reader->undeleteAll();
BOOST_CHECK_EQUAL(2, reader->numDocs());
// Ensure undeleteAll survives commit/close/reopen
reader->commit(MapStringString());
reader->close();
if (boost::dynamic_pointer_cast<MultiReader>(reader))
{
// MultiReader does not "own" the directory so it does not write the changes to sis on commit
sis->commit(dir);
}
sis->read(dir);
reader = openReader();
BOOST_CHECK_EQUAL(2, reader->numDocs());
reader->deleteDocument(0);
BOOST_CHECK_EQUAL(1, reader->numDocs());
reader->commit(MapStringString());
reader->close();
if (boost::dynamic_pointer_cast<MultiReader>(reader))
{
// MultiReader does not "own" the directory so it does not write the changes to sis on commit
sis->commit(dir);
}
sis->read(dir);
reader = openReader();
BOOST_CHECK_EQUAL(1, reader->numDocs());
}