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);
}
WeightedTermsPtr QueryTermExtractor::getIdfWeightedTerms(const Query* pQuery,
IndexReader* pReader, const string& fieldName)
{
WeightedTermsPtr pWTs = getTerms(pQuery, false, fieldName);
int32_t totalNumDocs = pReader->getDocCount();
WeightedTermPtr pWT;
for (size_t i = 0; i < pWTs->size(); i++)
{
try
{
pWT = (*pWTs)[i];
TermReaderPtr pTermReader = pReader->termReader();
TermPostingIteratorPtr pPostIter = pTermReader->seek(pWT->getTerm());
df_t docFreq = pPostIter->getTermMeta().getDocFreq();
//IDF algorithm taken from DefaultSimilarity class
float idf = (float)(std::log((float)totalNumDocs/(double)(docFreq + 1)) + 1.0);
pWT->m_fWeight *= idf;
}
catch (FX_NS(utility)::FirteXException&)
{
//ignore
}
}
return pWTs;
}
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::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++;
}
}
QueryExecutorPtr TermQuery::createExecutor(IndexReaderPtr& pIndexReader,
FeatureProviderPtr& pProvider, PoolPtr& pPool) const
{
TermReaderPtr pTermReader = pIndexReader->termReader();
TermPostingIteratorPtr pPostIter = pTermReader->seek(m_pTerm.get());
if (pPostIter.isNull())
{
return QueryExecutorPtr();
}
QueryExecutorPtr pExe = new TermQueryExecutor(this, pPostIter,
pProvider, pPool);
return pExe;
}
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 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 IndexTestCase::checkDocFreq(IndexReaderPtr& pIndexReader,
const tstring& sField,
const tstring& sTerm,
df_t expDf)
{
TermReaderPtr pTermReader = pIndexReader->termReader();
CPPUNIT_ASSERT(pTermReader);
StandardAnalyzer sa;
sa.init();
TokenViewPtr pTokens = sa.tokenize(sTerm.c_str(), sTerm.length());
CPPUNIT_ASSERT(pTokens);
CPPUNIT_ASSERT(pTokens->getNumTokens() == 1);
TokenView::Iterator it = pTokens->iterator();
TermPtr pTerm(new Term(sField, it.next().getTextValue()));
TermPostingIteratorPtr pPost = pTermReader->seek(pTerm.get());
CPPUNIT_ASSERT(pPost);
const TermMeta& termMeta = pPost->getTermMeta();
CPPUNIT_ASSERT_EQUAL(expDf, termMeta.getDocFreq());
}
void IndexTestCase::testDocumentDeletion()
{
DocumentSchema schema;
schema.addField("URL", "PRIMARY_KEY", true);
schema.addTextField("BODY");
schema.addField("MODIFIED", "INT64", true);
stringstream ss1;
const size_t NUM_DOCS = 1000;
size_t i = 0;
for (; i < NUM_DOCS; ++i)
{
ss1 << "url" << i << ", body" << i << " hot,"
<< (i * 100) % 1000 << ";";
}
buildIndex(schema, ss1.str());
stringstream ss2;
for (; i < 2 * NUM_DOCS; ++i)
{
ss2 << "url" << i << ", body" << i << " hot,"
<< (i * 100) % 1000 << ";";
}
buildIndex(schema, ss2.str(), true);
StandardAnalyzerPtr sa(new StandardAnalyzer());
sa->init();
TokenViewPtr pTokens = sa->tokenize("hot", 3);
CPPUNIT_ASSERT(pTokens);
CPPUNIT_ASSERT(pTokens->getNumTokens() == 1);
TokenView::Iterator it = pTokens->iterator();
TermPtr pTerm(new Term("BODY", it.next().getTextValue()));
tstring str = getTestPath();
std::set<docid_t> answer;
{
Index index;
index.open(str, Index::RDWR, NULL);
IndexWriterPtr pIndexWriter = index.acquireWriter();
CPPUNIT_ASSERT(pIndexWriter != NULL);
IndexReaderPtr pIndexReader = index.acquireReader();
CPPUNIT_ASSERT(pIndexReader != NULL);
for (size_t i = 0; i < 2 * NUM_DOCS; ++i)
{
stringstream ss;
ss << "url" << i;
if (i == 1000 || i == 1500 || i == 1505 || i == 1999)
{
pIndexWriter->deleteDocument(ss.str());
}
else
{
TermReaderPtr pTermReader = pIndexReader->termReader();
TermPtr pTerm(new Term("URL", ss.str()));
TermPostingIteratorPtr pIt = pTermReader->seek(pTerm.get());
docid_t docId = pIt->skipTo(0);
answer.insert(docId);
}
}
TermReaderPtr pTermReader = pIndexReader->termReader();
TermPostingIteratorPtr pDocFreqs = pTermReader->seek(pTerm.get());
CPPUNIT_ASSERT(pDocFreqs);
CPPUNIT_ASSERT_EQUAL((df_t)NUM_DOCS * 2, pDocFreqs->getTermMeta().getDocFreq());
std::set<docid_t>::const_iterator it = answer.begin();
for (docid_t i = 0; i < (docid_t)(2 * NUM_DOCS); )
{
docid_t docId = pDocFreqs->skipTo((docid_t)i);
i = docId + 1;
if (docId == INVALID_DOCID)
{
break;
}
CPPUNIT_ASSERT_EQUAL(*it, docId);
++it;
}
CPPUNIT_ASSERT(it == answer.end());
}
{
Index index;
index.open(str, Index::READ, NULL);
IndexReaderPtr pIndexReader = index.acquireReader();
CPPUNIT_ASSERT(pIndexReader != NULL);
TermReaderPtr pTermReader = pIndexReader->termReader();
TermPostingIteratorPtr pDocFreqs = pTermReader->seek(pTerm.get());
CPPUNIT_ASSERT(pDocFreqs);
CPPUNIT_ASSERT_EQUAL((df_t)(2 * NUM_DOCS), pDocFreqs->getTermMeta().getDocFreq());
std::set<docid_t>::const_iterator it = answer.begin();
for (docid_t i = 0; i < (docid_t)(2 * NUM_DOCS); )
{
docid_t docId = pDocFreqs->skipTo((docid_t)i);
i = docId + 1;
if (docId == INVALID_DOCID)
{
break;
}
CPPUNIT_ASSERT_EQUAL(*it, docId);
++it;
}
CPPUNIT_ASSERT(it == answer.end());
// for (std::set<docid_t>::const_iterator it = answer.begin();
// it != answer.end(); ++it)
// {
// docid_t docId = pDocFreqs->skipTo(*it);
// CPPUNIT_ASSERT_EQUAL(*it, docId);
// }
// docid_t docId = pDocFreqs->skipTo(NUM_DOCS + 0);
// CPPUNIT_ASSERT_EQUAL((docid_t)NUM_DOCS + 1, docId);
// docId = pDocFreqs->skipTo(NUM_DOCS + 500);
// CPPUNIT_ASSERT_EQUAL((docid_t)NUM_DOCS + 501, docId);
// docId = pDocFreqs->skipTo(NUM_DOCS + 505);
// CPPUNIT_ASSERT_EQUAL((docid_t)NUM_DOCS + 506, docId);
// docId = pDocFreqs->skipTo(2 * NUM_DOCS - 1);
// CPPUNIT_ASSERT_EQUAL((docid_t)INVALID_DOCID, docId);
}
}
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);
}
