void IndexContentTestCase::testIndexContent_DL()
{
Index* pIndex;
IndexReaderPtr pReader;
const Term* pTerm;
TermIteratorPtr pTermIter;
int docCount = 0;
int termCount = 0;
uint32_t i;
uint32_t indexTermId;
string fileName;
//Check posting list
Path indexPath = TestHelper::getTestDataPath();
indexPath.makeDirectory();
indexPath.pushDirectory(_T("test_dlindex"));
pIndex = new Index(indexPath.toString().c_str(), Index::READ, NULL);
auto_ptr<Index> indexPtr(pIndex);
pReader = pIndex->acquireReader();
TermReaderPtr pTermReader = pReader->termReader();
pTermIter = pTermReader->termIterator("BODY");
StoredFieldsReaderPtr pDocReader = pReader->createStoredFieldsReader();
//Iterator all terms
while(pTermIter->next())
{
pTerm = pTermIter->term();
CPPUNIT_ASSERT(pTermReader->seek(pTerm));
indexTermId = (pTerm->cast<int32_t>())->getValue();
docCount = 0;
TermPostingIteratorPtr pTermDocFreqs = pTermReader->termPostings();
while(pTermDocFreqs->nextDoc())
{
DocumentPtr pDoc = pDocReader->document(pTermDocFreqs->doc());
docCount++;
// 获取文件路径
fileName.assign(pDoc->getField("PATH")->getValue().c_str());
TermList* pTermIdList = m_pDocScanner->getTermListOfFile(fileName);
CPPUNIT_ASSERT(pTermIdList != NULL);
for(i = 0, termCount = 0; i < pTermIdList->getSize(); i++)
{
if(indexTermId == pTermIdList->getValue(i))
{
termCount++;
}
}
CPPUNIT_ASSERT_EQUAL((tf_t)termCount, pTermDocFreqs->freq());
}//end while nextDoc()
CPPUNIT_ASSERT_EQUAL((df_t)docCount, pTermDocFreqs->getDocFreq());
}
CPPUNIT_ASSERT(m_pDocScanner->getTotalTermCount() == pReader->getNumTerms());
}
void IndexTestCase::testKeywordIndex()
{
DocumentSchema schema;
schema.addUnIndexedField("PATH");
schema.addField("Keyword", "KEYWORD", false);
buildIndex(schema, "file1.txt, hello world.");
tstring str = getTestPath();
Index index;
index.open(str, Index::READ, NULL);
IndexReaderPtr pReader = index.acquireReader();
CPPUNIT_ASSERT(pReader != NULL);
TermReaderPtr pTermReader = pReader->termReader();
CPPUNIT_ASSERT(pTermReader);
TermIteratorPtr pTermIterator = pTermReader->termIterator("Keyword");
CPPUNIT_ASSERT(pTermIterator != NULL);
while (pTermIterator->hasNext())
{
TermIterator::TermEntry entry = pTermIterator->next();
const TermMeta& termMeta = entry.postingIterator->getTermMeta();
CPPUNIT_ASSERT_EQUAL((df_t)1, termMeta.getDocFreq());
CPPUNIT_ASSERT_EQUAL((ctf_t)1, termMeta.getCTF());
}
Term term("Keyword", "hello world.");
TermPostingIteratorPtr pPost = pTermReader->seek(&term);
CPPUNIT_ASSERT(pPost);
docid_t docId = pPost->skipTo(0);
CPPUNIT_ASSERT_EQUAL((docid_t)0, docId);
docId = pPost->skipTo(++docId);
CPPUNIT_ASSERT_EQUAL((docid_t)INVALID_DOCID, docId);
}
void DateTimeIndexTestCase::testCTF()
{
const size_t NUM_DOCS = 100;
stringstream ss;
for (size_t i = 0; i < NUM_DOCS; ++i)
{
ss << 2009 << "-" << i % 12 + 1 << "-" << i % 27 + 1 << " "
<< i % 24 << ":" << i % 60 << ":" << i % 60 << ";";
}
buildDateTimeIndex(ss.str());
Index index;
index.open(getIndexPath(), Index::READ, NULL);
IndexReaderPtr pReader = index.acquireReader();
CPPUNIT_ASSERT(pReader != NULL);
TermReaderPtr pTermReader = pReader->termReader();
CPPUNIT_ASSERT(pTermReader);
TermIteratorPtr pTermIterator = pTermReader->termIterator("DateTime1");
CPPUNIT_ASSERT(pTermIterator != NULL);
df_t ttf = 0;
while (pTermIterator->hasNext())
{
TermIterator::TermEntry entry = pTermIterator->next();
const TermMeta& termMeta = entry.postingIterator->getTermMeta();
CPPUNIT_ASSERT_EQUAL((df_t)1, termMeta.getDocFreq());
CPPUNIT_ASSERT_EQUAL((ctf_t)1, termMeta.getCTF());
ttf += termMeta.getDocFreq();
}
CPPUNIT_ASSERT_EQUAL((df_t)100, ttf);
}
void KeywordIndexTestCase::testCTF()
{
GLOBAL_CONF().Build.buildThreadCount = 2;
buildKeywordIndex("1, 2; 3, 4; 5, 6; 7, 8; 9, 10; "
"1, 2; 3, 4; 5, 6; 7, 8; 9, 10");
Index index;
index.open(getIndexPath(), Index::READ, NULL);
IndexReaderPtr pReader = index.acquireReader();
CPPUNIT_ASSERT(pReader != NULL);
TermReaderPtr pTermReader = pReader->termReader();
CPPUNIT_ASSERT(pTermReader);
TermIteratorPtr pTermIterator = pTermReader->termIterator("Keyword1");
CPPUNIT_ASSERT(pTermIterator != NULL);
try
{
while (pTermIterator->hasNext())
{
TermIterator::TermEntry entry = pTermIterator->next();
const TermMeta& termMeta = entry.postingIterator->getTermMeta();
CPPUNIT_ASSERT_EQUAL((df_t)2, termMeta.getDocFreq());
CPPUNIT_ASSERT_EQUAL((ctf_t)2, termMeta.getCTF());
}
}
catch(const FirteXException& e)
{
cout << "ERROR:" << e.what() << endl;
CPPUNIT_ASSERT(false);
}
}
void IndexTestCase::testInt32_IF()
{
DocumentSchema schema;
schema.addField("Int32", "INT32_IF", false);
const static size_t NUM_DOCS = 1000;
stringstream ss;
for (size_t i = 0; i < NUM_DOCS; ++i)
{
ss << (i % 100) << ";";
}
GLOBAL_CONF().Build.buildThreadCount = 1;
buildIndex(schema, ss.str());
tstring str = getTestPath();
Index index;
index.open(str, Index::READ, NULL);
IndexReaderPtr pReader = index.acquireReader();
CPPUNIT_ASSERT(pReader != NULL);
TermReaderPtr pTermReader = pReader->termReader();
CPPUNIT_ASSERT(pTermReader);
TermIteratorPtr pTermIterator = pTermReader->termIterator("Int32");
CPPUNIT_ASSERT(pTermIterator != NULL);
while (pTermIterator->hasNext())
{
TermIterator::TermEntry entry = pTermIterator->next();
const TermMeta& termMeta = entry.postingIterator->getTermMeta();
CPPUNIT_ASSERT_EQUAL((df_t)10, termMeta.getDocFreq());
CPPUNIT_ASSERT_EQUAL((ctf_t)10, termMeta.getCTF());
}
Term term("Int32", "0");
TermPostingIteratorPtr pPost = pTermReader->seek(&term);
CPPUNIT_ASSERT(pPost);
docid_t docId = pPost->skipTo(0);
CPPUNIT_ASSERT_EQUAL((docid_t)0, docId);
docId = pPost->skipTo(901);
CPPUNIT_ASSERT_EQUAL((docid_t)INVALID_DOCID, docId);
ForwardIndexIteratorPtr pForIndexIt = pReader->forwardIndexReader("Int32");
CPPUNIT_ASSERT(pForIndexIt != NULL);
Int32ForwardIndexIteratorPtr pInt32ForIndexIt =
pForIndexIt.cast<Int32ForwardIndexIterator>();
CPPUNIT_ASSERT(pInt32ForIndexIt != NULL);
int32_t value = 0;
docId = 0;
for (; docId < (docid_t)NUM_DOCS; ++docId)
{
CPPUNIT_ASSERT(pInt32ForIndexIt->seek(docId, value));
CPPUNIT_ASSERT_EQUAL((int32_t)(docId % 100), value);
}
CPPUNIT_ASSERT(!pInt32ForIndexIt->seek(docId, value));
}
void DateTimeIndexTestCase::testPosting()
{
const size_t NUM_DOCS = 100;
vector<TimeAndDoc> timeVec;
stringstream ss;
for (size_t i = 0; i < NUM_DOCS; ++i)
{
int y = 2011;
int m = (i % 12) + 1;
int d = (i % 27) + 1;
int h = i % 24;
int min = i % 60;
int s = i % 60;
ss << y << "-" << m << "-" << d << " "
<< h << ":" << min << ":" << s << ";";
TimeAndDoc td;
td.nTime = DateTimeAnalyzer::makeTime(y, m, d, h, min, s);
td.docId = (docid_t)i;
timeVec .push_back(td);
}
sort(timeVec.begin(), timeVec.end(), timeLess);
GLOBAL_CONF().Build.buildThreadCount = 1;
buildDateTimeIndex(ss.str());
Index index;
index.open(getIndexPath(), Index::READ, NULL);
IndexReaderPtr pReader = index.acquireReader();
TermReaderPtr pTermReader = pReader->termReader();
TermIteratorPtr pIterator = pTermReader->termIterator("DateTime1");
CPPUNIT_ASSERT(pIterator);
size_t j = 0;
while (pIterator->hasNext())
{
TermIterator::TermEntry entry = pIterator->next();
const Term* pTerm = entry.term;
const Int64Term* pTermX = pTerm->cast<int64_t>();
int64_t nCurTerm = pTermX->getValue();
int64_t nExpTerm = timeVec[j].nTime;
CPPUNIT_ASSERT_EQUAL(nExpTerm, nCurTerm);
TermPostingIteratorPtr pPostingIter = entry.postingIterator;
CPPUNIT_ASSERT(pPostingIter != NULL);
docid_t docId = pPostingIter->skipTo(0);
CPPUNIT_ASSERT_EQUAL(timeVec[j].docId, docId);
docId = pPostingIter->skipTo(docId + 1);
CPPUNIT_ASSERT_EQUAL((docid_t)INVALID_DOCID, docId);
j++;
}
}
void IndexTestCase::testPrimaryKeyIndex()
{
DocumentSchema schema;
schema.addField("PK", "PRIMARY_KEY", false);
const static size_t NUM_DOCS = 1000;
stringstream ss;
for (size_t i = 0; i < NUM_DOCS; ++i)
{
ss << i << ";";
}
GLOBAL_CONF().Build.buildThreadCount = 1;
buildIndex(schema, ss.str());
tstring str = getTestPath();
Index index;
index.open(str, Index::READ, NULL);
IndexReaderPtr pReader = index.acquireReader();
CPPUNIT_ASSERT(pReader != NULL);
TermReaderPtr pTermReader = pReader->termReader();
CPPUNIT_ASSERT(pTermReader);
TermIteratorPtr pTermIterator = pTermReader->termIterator("PK");
CPPUNIT_ASSERT(pTermIterator != NULL);
while (pTermIterator->hasNext())
{
TermIterator::TermEntry entry = pTermIterator->next();
const TermMeta& termMeta = entry.postingIterator->getTermMeta();
CPPUNIT_ASSERT_EQUAL((df_t)1, termMeta.getDocFreq());
CPPUNIT_ASSERT_EQUAL((ctf_t)1, termMeta.getCTF());
}
for (size_t i = 0; i < NUM_DOCS; ++i)
{
stringstream ss2;
ss2 << i;
Term term("PK", ss2.str());
TermPostingIteratorPtr pPost = pTermReader->seek(&term);
CPPUNIT_ASSERT(pPost);
docid_t docId = pPost->skipTo(0);
CPPUNIT_ASSERT_EQUAL((docid_t)i, docId);
docId = pPost->skipTo(++docId);
CPPUNIT_ASSERT_EQUAL((docid_t)INVALID_DOCID, docId);
}
}
void TextIndexMergerTestCase::checkMergedResult(TermIteratorPtr& pTermIt, TermMap& answer)
{
CPPUNIT_ASSERT(pTermIt);
CPPUNIT_ASSERT_EQUAL((int64_t)answer.size(), pTermIt->size());
size_t termCount = 0;
while (pTermIt->hasNext())
{
TermIterator::TermEntry entry = pTermIt->next();
const TextTermIterator::TermType* pTerm =
dynamic_cast<const TextTermIterator::TermType*>(entry.term);
CPPUNIT_ASSERT(pTerm != NULL);
uint64_t hash = pTerm->getValue();
FX_TRACE("Term hash: %llu", hash);
TermMap::const_iterator it = answer.find(hash);
CPPUNIT_ASSERT(it != answer.end());
CPPUNIT_ASSERT_EQUAL(it->first, hash);
const TermMeta& termMeta = entry.postingIterator->getTermMeta();
size_t df = it->second.size();
CPPUNIT_ASSERT_EQUAL((df_t)df, termMeta.getDocFreq());
ctf_t ctf = 0;
for (size_t i = 0; i < df; ++i)
{
docid_t docId = entry.postingIterator->skipTo(it->second[i].first);
CPPUNIT_ASSERT_EQUAL(it->second[i].first, docId);
tf_t tf = it->second[i].second.size();
CPPUNIT_ASSERT_EQUAL(tf, entry.postingIterator->freq());
for (size_t j = 0; j < (size_t)tf; ++j)
{
loc_t posExp = it->second[i].second[j];
loc_t pos = entry.postingIterator->skipToPosition(posExp);
CPPUNIT_ASSERT_EQUAL(posExp, pos);
}
ctf += it->second[i].second.size();
}
CPPUNIT_ASSERT_EQUAL(ctf, termMeta.getCTF());
termCount++;
}
CPPUNIT_ASSERT_EQUAL(answer.size(), termCount);
}
void IndexTestCase::testTextIndex()
{
DocumentSchema schema;
schema.addUnIndexedField("PATH");
schema.addTextField("CONTENT");
buildIndex(schema, "file1.txt, hello world.");
tstring str = getTestPath();
Index index;
index.open(str, Index::READ, NULL);
IndexReaderPtr pReader = index.acquireReader();
CPPUNIT_ASSERT(pReader != NULL);
TermReaderPtr pTermReader = pReader->termReader();
CPPUNIT_ASSERT(pTermReader);
TermIteratorPtr pTermIterator = pTermReader->termIterator("CONTENT");
CPPUNIT_ASSERT(pTermIterator != NULL);
while (pTermIterator->hasNext())
{
TermIterator::TermEntry entry = pTermIterator->next();
const TermMeta& termMeta = entry.postingIterator->getTermMeta();
CPPUNIT_ASSERT_EQUAL((df_t)1, termMeta.getDocFreq());
CPPUNIT_ASSERT_EQUAL((ctf_t)1, termMeta.getCTF());
}
Term term("CONTENT", "hello");
TermPostingIteratorPtr pPost = pTermReader->seek(&term);
CPPUNIT_ASSERT(pPost);
docid_t docId = pPost->skipTo(0);
CPPUNIT_ASSERT_EQUAL((docid_t)0, docId);
docId = pPost->skipTo(++docId);
CPPUNIT_ASSERT_EQUAL((docid_t)INVALID_DOCID, docId);
StoredFieldsReaderPtr pDocReader = pReader->createStoredFieldsReader();
CPPUNIT_ASSERT(pDocReader);
FieldSelector selector(pReader->getDocSchema(), true, false);
ResultDoc resultDoc(0);
bool ret = pDocReader->getDocument(selector, resultDoc);
CPPUNIT_ASSERT(ret);
CPPUNIT_ASSERT(resultDoc.size() > 0);
}
void IndexContentTestCase::testIndexContent_WL()
{
Index* pIndex;
IndexReaderPtr pReader;
const Term* pTerm;
TermIteratorPtr pTermIter;
int docCount = 0;
int termCount = 0;
int pos = -1;
uint32_t indexTermId;
string fileName;
//Check posting list
Path indexPath = TestHelper::getTestDataPath();
indexPath.makeDirectory();
indexPath.pushDirectory(_T("test_wlindex"));
pIndex = new Index(indexPath.toString().c_str(), Index::READ, NULL);
auto_ptr<Index> indexPtr(pIndex);
CPPUNIT_ASSERT(pIndex != NULL);
pReader = pIndex->acquireReader();
TermReaderPtr pTermReader = pReader->termReader();
pTermIter = pTermReader->termIterator("BODY");
StoredFieldsReaderPtr pDocReader = pReader->createStoredFieldsReader();
//Iterator all terms
while(pTermIter->next())
{
pTerm = pTermIter->term();
CPPUNIT_ASSERT(pTermReader->seek(pTerm));
indexTermId = (pTerm->cast<int32_t>())->getValue();
TermPositionIteratorPtr pPositions = pTermReader->termPositions();
docCount = 0;
while(pPositions->nextDoc())
{
DocumentPtr pDoc = pDocReader->document(pPositions->doc());
docCount++;
fileName.assign(pDoc->getField("PATH")->getValue().c_str());
TermList* pTermIdList = m_pDocScanner->getTermListOfFile(fileName);
CPPUNIT_ASSERT(pTermIdList != NULL);
pos = pPositions->nextPosition();
termCount = 0;
while(pos != -1)
{
termCount++;
CPPUNIT_ASSERT(indexTermId == pTermIdList->getValue(pos));
pos = pPositions->nextPosition();
}
CPPUNIT_ASSERT(termCount == pPositions->freq());
}//end while nextDoc()
CPPUNIT_ASSERT(docCount == pPositions->getDocFreq());
}
CPPUNIT_ASSERT_EQUAL((int64_t)m_pDocScanner->getTotalTermCount(),
(int64_t)pReader->getNumTerms());
}
void TextIndexMergerTestCase::printPositng(TermIteratorPtr& pTermIt)
{
while (pTermIt->hasNext())
{
TermIterator::TermEntry entry = pTermIt->next();
const TextTermIterator::TermType* pTerm =
dynamic_cast<const TextTermIterator::TermType*>(entry.term);
CPPUNIT_ASSERT(pTerm != NULL);
uint64_t hash = pTerm->getValue();
std::cout << "Term hash: " << hash << std::endl;
const TermMeta& termMeta = entry.postingIterator->getTermMeta();
std::cout << "Term meta: df: " << termMeta.getDocFreq()
<< ", ctf: " << termMeta.getCTF() << std::endl;
PostingDecoderPtr pDecoder = entry.postingIterator->getPostingDecoder();
docid_t docBuffer[RECORD_SIZE];
tf_t tfBuffer[RECORD_SIZE];
uint32_t nDecoded = 0;
docid_t lastDocId = 0;
while ((nDecoded = pDecoder->decodeDocRecord(docBuffer, lastDocId)) > 0)
{
if (pDecoder->decodeTfRecord(tfBuffer) != nDecoded)
{
FIRTEX_THROW(IndexCollapseException, "Doc and Tf record is inconsistant.");
}
std::cout << "doc list: " << std::endl;
for (size_t i = 0;i < nDecoded; ++i)
{
std::cout << docBuffer[i] << ", ";
}
std::cout << "end doc list: " << std::endl;
std::cout << "tf list: " << std::endl;
for (size_t i = 0;i < nDecoded; ++i)
{
std::cout << tfBuffer[i] << ", ";
}
std::cout << "end tf list: " << std::endl;
lastDocId = docBuffer[nDecoded - 1] + 1;
}
loc_t posBuffer[RECORD_SIZE];
uint32_t startOff = 0;
uint32_t nDecodedPos = 0;
uint32_t nTotalDecodedPos = 0;
while ((nDecodedPos = pDecoder->decodePosRecord(posBuffer,
nTotalDecodedPos, startOff)) > 0)
{
std::cout << "pos list: " << std::endl;
for (size_t i = 0;i < nDecodedPos; ++i)
{
std::cout << posBuffer[i] << ", ";
}
std::cout << "end pos list: " << std::endl;
nTotalDecodedPos += nDecodedPos;
}
}
}
void KeywordIndexTestCase::testPosting()
{
const size_t NUM_DOCS = 100;
vector<HashAndDoc> hashVec;
stringstream ss;
size_t i;
for (i = 0; i < NUM_DOCS; ++i)
{
stringstream ss1;
ss1 << "abc" << i;
ss << ss1.str() << ", " << "abc" << ";";
HashAndDoc hd;
hd.nHash = Hash::hashString64(ss1.str().c_str());
hd.docId = (docid_t)i;
hashVec.push_back(hd);
}
GLOBAL_CONF().Build.buildThreadCount = 1;
buildKeywordIndex(ss.str());
Index index;
index.open(getIndexPath(), Index::READ, NULL);
IndexReaderPtr pReader = index.acquireReader();
TermReaderPtr pTermReader = pReader->termReader();
TermIteratorPtr pIterator = pTermReader->termIterator("Keyword1");
CPPUNIT_ASSERT(!pIterator.isNull());
sort(hashVec.begin(), hashVec.end(), hashLess);
int32_t j = 0;
while (pIterator->hasNext())
{
TermIterator::TermEntry entry = pIterator->next();
const Term* pTerm = entry.term;
const UInt64Term* pTermX = pTerm->cast<uint64_t>();
uint64_t nCurTerm = pTermX->getValue();
uint64_t nExpTerm = hashVec[j].nHash;
CPPUNIT_ASSERT_EQUAL(nExpTerm, nCurTerm);
TermPostingIteratorPtr pPostingIter = entry.postingIterator;
CPPUNIT_ASSERT(pPostingIter != NULL);
docid_t docId = pPostingIter->skipTo(0);
CPPUNIT_ASSERT_EQUAL(hashVec[j].docId, docId);
docId = pPostingIter->skipTo(docId + 1);
CPPUNIT_ASSERT_EQUAL((docid_t)INVALID_DOCID, docId);
j++;
}
//test abc term
TermIteratorPtr pIterator2 = pTermReader->termIterator("Keyword2");
CPPUNIT_ASSERT(!pIterator2.isNull());
i = 0;
while (pIterator2->hasNext())
{
TermIterator::TermEntry entry = pIterator2->next();
const Term* pTerm = entry.term;
const UInt64Term* pTermX = pTerm->cast<uint64_t>();
uint64_t nCurTerm = pTermX->getValue();
uint64_t nExpTerm = Hash::hashString64("abc");
CPPUNIT_ASSERT_EQUAL(nExpTerm, nCurTerm);
TermPostingIteratorPtr pPostingIter = entry.postingIterator;
CPPUNIT_ASSERT(pPostingIter != NULL);
docid_t docId = 0;
docid_t docIdRet = 0;
while ((docIdRet = pPostingIter->skipTo(docId)) != INVALID_DOCID)
{
CPPUNIT_ASSERT_EQUAL(docId++, docIdRet);
}
CPPUNIT_ASSERT_EQUAL((docid_t)100, docId);
i++;
}
CPPUNIT_ASSERT_EQUAL((size_t)1, i);
}