void searchPhrase(IndexSearcher &is, const string &qStr, bool fuzzy) {
string queryString = qStr;
BoolAnalyzer analyzer;
const Query *pQuery1 = QueryParser::parsePhrase(
queryString,
"text",
analyzer, fuzzy);
Query *pQuery2 = QueryParser::parsePhrase(
queryString,
"title",
analyzer, fuzzy);
pQuery2->setBoost(10);
const Query *query = new OrQuery(*pQuery1, *pQuery2);
cout << "Inner Query: " << query->toString() << endl;
vector<ScoreDoc> scoreDocs = is.search(*query);
cout << "Hit Numbers: " << scoreDocs.size() << endl;
cout << "Hit Documents:" << endl;
for (size_t i = 0; i < scoreDocs.size(); i ++) {
fprintf(stdout, "%4ld. #%-4ld ", i + 1, scoreDocs[i].id());
cout << is.doc(scoreDocs[i].id()).toString() << endl;
}
delete query;
}
// BK> all test functions are the same except RAMDirectory constructor, so shared code moved here
void checkDir(CuTest *tc, MockRAMDirectory * ramDir) {
// Check size
CuAssertTrue(tc, ramDir->sizeInBytes == ramDir->getRecomputedSizeInBytes(), _T("RAMDir size"));
// open reader to test document count
IndexReader * reader = IndexReader::open(ramDir);
CuAssertEquals(tc, docsToAdd, reader->numDocs(), _T("document count"));
// open search to check if all doc's are there
IndexSearcher * searcher = _CLNEW IndexSearcher(reader);
// search for all documents
Document doc;
for (int i = 0; i < docsToAdd; i++) {
searcher->doc(i, doc);
CuAssertTrue(tc, doc.getField(_T("content")) != NULL, _T("content is NULL"));
}
// cleanup
reader->close();
searcher->close();
_CLLDELETE(reader);
_CLLDELETE(searcher);
}
void testEqualScores()
{
// NOTE: uses index build in *this* setUp
IndexReader * pReader = IndexReader::open( m_pSmall );
IndexSearcher * pSearch = _CLNEW IndexSearcher( pReader );
Hits * pResult;
// some hits match more terms then others, score should be the same
Query * q = csrq( _T( "data" ), _T( "1" ), _T( "6" ), true, true );
pResult = pSearch->search( q );
size_t numHits = pResult->length();
assertEqualsMsg( _T( "wrong number of results" ), 6, numHits );
float_t score = pResult->score( 0 );
for( size_t i = 1; i < numHits; i++ )
{
assertTrueMsg( _T( "score was not the same" ), score == pResult->score( i ));
}
_CLDELETE( pResult );
_CLDELETE( q );
pSearch->close();
_CLDELETE( pSearch );
pReader->close();
_CLDELETE( pReader );
}
void testBoost()
{
// NOTE: uses index build in *this* setUp
IndexReader * pReader = IndexReader::open( m_pSmall );
IndexSearcher * pSearch = _CLNEW IndexSearcher( pReader );
Hits * pResult;
// test for correct application of query normalization
// must use a non score normalizing method for this.
Query * q = csrq( _T( "data" ), _T( "1" ), _T( "6" ), true, true );
q->setBoost( 100 );
pResult = pSearch->search( q );
for( size_t i = 1; i < pResult->length(); i++ )
{
assertTrueMsg( _T( "score was not was not correct" ), 1.0f == pResult->score( i ));
}
_CLDELETE( pResult );
_CLDELETE( q );
//
// Ensure that boosting works to score one clause of a query higher
// than another.
//
Query * q1 = csrq( _T( "data" ), _T( "A" ), _T( "A" ), true, true ); // matches document #0
q1->setBoost( .1f );
Query * q2 = csrq( _T( "data" ), _T( "Z" ), _T( "Z" ), true, true ); // matches document #1
BooleanQuery * bq = _CLNEW BooleanQuery( true );
bq->add( q1, true, BooleanClause::SHOULD );
bq->add( q2, true, BooleanClause::SHOULD );
pResult = pSearch->search( bq );
assertEquals( 1, pResult->id( 0 ));
assertEquals( 0, pResult->id( 1 ));
assertTrue( pResult->score( 0 ) > pResult->score( 1 ));
_CLDELETE( pResult );
_CLDELETE( bq );
q1 = csrq( _T( "data" ), _T( "A" ), _T( "A" ), true, true ); // matches document #0
q1->setBoost( 10.0f );
q2 = csrq( _T( "data" ), _T( "Z" ), _T( "Z" ), true, true ); // matches document #1
bq = _CLNEW BooleanQuery( true );
bq->add( q1, true, BooleanClause::SHOULD );
bq->add( q2, true, BooleanClause::SHOULD );
pResult = pSearch->search( bq );
assertEquals( 0, pResult->id( 0 ));
assertEquals( 1, pResult->id( 1 ));
assertTrue( pResult->score( 0 ) > pResult->score( 1 ));
_CLDELETE( pResult );
_CLDELETE( bq );
pSearch->close();
_CLDELETE( pSearch );
pReader->close();
_CLDELETE( pReader );
}
void testEndThreadException(CuTest *tc) {
const int MAX_DOCS=1500;
RAMDirectory ram;
WhitespaceAnalyzer an;
IndexWriter* writer = _CLNEW IndexWriter(&ram, &an, true);
// add some documents
Document doc;
for (int i = 0; i < MAX_DOCS; i++) {
TCHAR * tmp = English::IntToEnglish(i);
doc.add(* new Field(_T("content"), tmp, Field::STORE_YES | Field::INDEX_UNTOKENIZED));
writer->addDocument(&doc);
doc.clear();
_CLDELETE_ARRAY( tmp );
}
CuAssertEquals(tc, MAX_DOCS, writer->docCount());
writer->close();
_CLLDELETE(writer);
// this sequence is OK: delete searcher after search thread finish
{
IndexSearcher * searcher = _CLNEW IndexSearcher(&ram);
_LUCENE_THREADID_TYPE thread = _LUCENE_THREAD_CREATE(&searchDocs, searcher);
SCOPED_LOCK_MUTEX(searchMutex);
CONDITION_WAIT(searchMutex, searchCondition);
// _LUCENE_SLEEP(9999); //make sure that deleteMutex is being waited on...
CONDITION_NOTIFYALL(deleteCondition);
_LUCENE_THREAD_JOIN(thread);
searcher->close();
_CLLDELETE(searcher);
}
// this produces memory exception: delete searcher after search finish but before thread finish
{
IndexSearcher * searcher = _CLNEW IndexSearcher(&ram);
_LUCENE_THREADID_TYPE thread = _LUCENE_THREAD_CREATE(&searchDocs, searcher);
SCOPED_LOCK_MUTEX(searchMutex);
CONDITION_WAIT(searchMutex, searchCondition);
searcher->close();
_CLLDELETE(searcher);
CONDITION_NOTIFYALL(deleteCondition);
_LUCENE_THREAD_JOIN(thread);
}
ram.close();
}
void testIncludeLowerTrue(CuTest* tc)
{
WhitespaceAnalyzer a;
RAMDirectory* index = _CLNEW RAMDirectory();
IndexWriter* writer = _CLNEW IndexWriter(index,
&a, true);
Document doc;
doc.add(*_CLNEW Field(_T("Category"), _T("a 1"), Field::STORE_YES | Field::INDEX_TOKENIZED));
writer->addDocument(&doc); doc.clear();
doc.add(*_CLNEW Field(_T("Category"), _T("a 2"), Field::STORE_YES | Field::INDEX_TOKENIZED));
writer->addDocument(&doc); doc.clear();
doc.add(*_CLNEW Field(_T("Category"), _T("a 3"), Field::STORE_YES | Field::INDEX_TOKENIZED));
writer->addDocument(&doc); doc.clear();
writer->close();
_CLLDELETE(writer);
IndexSearcher* s = _CLNEW IndexSearcher(index);
Filter* f = _CLNEW RangeFilter(_T("Category"), _T("3"), _T("3"), true, true);
Term* t = _CLNEW Term(_T("Category"), _T("a"));
Query* q1 = _CLNEW TermQuery(t);
_CLLDECDELETE(t);
t = _CLNEW Term(_T("Category"), _T("3"));
Query* q2 = _CLNEW TermQuery(t);
_CLLDECDELETE(t);
Hits* h = s->search(q1);
assertTrue(h->length() == 3);
_CLLDELETE(h);
h = s->search(q2);
assertTrue(h->length() == 1);
_CLLDELETE(h);
h = s->search(q1, f);
assertTrue(h->length() == 1);
_CLLDELETE(h);
s->close();
_CLLDELETE(s);
_CLLDELETE(q1);
_CLLDELETE(q2);
_CLLDELETE(f);
index->close();
_CLLDECDELETE(index);
}
_LUCENE_THREAD_FUNC(searchDocs, _searcher) {
WhitespaceAnalyzer an;
IndexSearcher * searcher = (IndexSearcher *)_searcher;
Query * query = QueryParser::parse(_T("one"), _T("content"), &an);
Hits * hits = searcher->search(query);
// _LUCENE_SLEEP(9999); //make sure that searchMutex is being waited on...
CONDITION_NOTIFYALL(searchCondition);
SCOPED_LOCK_MUTEX(deleteMutex);
_CLLDELETE(hits);
_CLLDELETE(query);
CONDITION_WAIT(deleteMutex, deleteCondition);
// _LUCENE_THREAD_FUNC_RETURN(0);
}
/// TestBooleanScorer.java, ported 5/9/2009
void testBooleanScorer(CuTest *tc) {
const TCHAR* FIELD = _T("category");
RAMDirectory directory;
TCHAR* values[] = { _T("1"), _T("2"), _T("3"), _T("4"), NULL};
try {
WhitespaceAnalyzer a;
IndexWriter* writer = _CLNEW IndexWriter(&directory, &a, true);
for (size_t i = 0; values[i]!=NULL; i++) {
Document* doc = _CLNEW Document();
doc->add(*_CLNEW Field(FIELD, values[i], Field::STORE_YES | Field::INDEX_TOKENIZED));
writer->addDocument(doc);
_CLLDELETE(doc);
}
writer->close();
_CLLDELETE(writer);
BooleanQuery* booleanQuery1 = _CLNEW BooleanQuery();
Term *t = _CLNEW Term(FIELD, _T("1"));
booleanQuery1->add(_CLNEW TermQuery(t), true, BooleanClause::SHOULD);
_CLDECDELETE(t);
t = _CLNEW Term(FIELD, _T("2"));
booleanQuery1->add(_CLNEW TermQuery(t), true, BooleanClause::SHOULD);
_CLDECDELETE(t);
BooleanQuery* query = _CLNEW BooleanQuery();
query->add(booleanQuery1, true, BooleanClause::MUST);
t = _CLNEW Term(FIELD, _T("9"));
query->add(_CLNEW TermQuery(t), true, BooleanClause::MUST_NOT);
_CLDECDELETE(t);
IndexSearcher *indexSearcher = _CLNEW IndexSearcher(&directory);
Hits *hits = indexSearcher->search(query);
CLUCENE_ASSERT(2 == hits->length()); // Number of matched documents
_CLLDELETE(hits);
_CLLDELETE(indexSearcher);
_CLLDELETE(query);
}
catch (CLuceneError& e) {
CuFail(tc, e.twhat());
}
}
void testBooleanOrderUnAffected()
{
// NOTE: uses index build in *this* setUp
IndexReader * pReader = IndexReader::open( m_pSmall );
IndexSearcher * pSearch = _CLNEW IndexSearcher( pReader );
// first do a regular RangeQuery which uses term expansion so
// docs with more terms in range get higher scores
Term * pLower = _CLNEW Term( _T( "data" ), _T( "1" ));
Term * pUpper = _CLNEW Term( _T( "data" ), _T( "4" ));
Query * rq = _CLNEW RangeQuery( pLower, pUpper, true );
_CLLDECDELETE( pUpper );
_CLLDECDELETE( pLower );
Hits * pExpected = pSearch->search( rq );
size_t numHits = pExpected->length();
// now do a boolean where which also contains a
// ConstantScoreRangeQuery and make sure the order is the same
BooleanQuery * q = _CLNEW BooleanQuery();
q->add( rq, true, BooleanClause::MUST );
q->add( csrq( _T( "data" ), _T( "1" ), _T( "6" ), true, true ), true, BooleanClause::MUST );
Hits * pActual = pSearch->search( q );
assertEqualsMsg( _T( "wrong number of hits" ), numHits, pActual->length() );
for( size_t i = 0; i < numHits; i++ )
{
assertEqualsMsg( _T( "mismatch in docid for a hit" ), pExpected->id( i ), pActual->id( i ));
}
_CLDELETE( pActual );
_CLDELETE( pExpected );
_CLDELETE( q );
pSearch->close();
_CLDELETE( pSearch );
pReader->close();
_CLDELETE( pReader );
}
void _TestSearchesRun( Analyzer& analyzer, IndexSearcher& search, char_t* qry){
try{
Query& q = QueryParser::Parse(qry , _T("contents"), analyzer);
Hits& h = search.search( q );
delete &h;
delete &q;
}catch(THROW_TYPE e){
_cout << _T("Failed: ") << qry << " - " << e.what() << endl;
}catch(...){
_cout << _T("FAILED: ") << qry << endl;
}
}
/////////////////////////////////////////////////////////////////////////////
// CLucene specific
// Visual Studio 2005 shows memory leaks for this test, but some other
// tools do not detect any memory leaks. So what is right?
// IN VC80 shows memory leaks ONLY if both sub-queries are added as
// MUST BooleanClauses.
void testBooleanMemLeaks()
{
IndexReader * pReader = IndexReader::open( m_pSmall );
IndexSearcher * pSearch = _CLNEW IndexSearcher( pReader );
Query * q1 = csrq( _T( "data" ), _T( "A" ), _T( "A" ), true, true ); // matches document #0
Query * q2 = csrq( _T( "data" ), _T( "Z" ), _T( "Z" ), true, true ); // matches document #1
BooleanQuery * bq = _CLNEW BooleanQuery( true );
bq->add( q1, true, BooleanClause::MUST );
bq->add( q2, true, BooleanClause::MUST );
Hits * pResult = pSearch->search( bq );
_CLDELETE( pResult );
_CLDELETE( bq );
pSearch->close();
_CLDELETE( pSearch );
pReader->close();
_CLDELETE( pReader );
}
void testRangeFilterId() {
IndexReader* reader = IndexReader::open(index);
IndexSearcher* search = new IndexSearcher(reader);
int medId = ((maxId - minId) / 2);
std::tstring minIPstr = pad(minId);
const TCHAR* minIP = minIPstr.c_str();
std::tstring maxIPstr = pad(maxId);
const TCHAR* maxIP = maxIPstr.c_str();
std::tstring medIPstr = pad(medId);
const TCHAR* medIP = medIPstr.c_str();
size_t numDocs = static_cast<size_t>(reader->numDocs());
assertEqualsMsg(_T("num of docs"), numDocs, static_cast<size_t>(1+ maxId - minId));
Hits* result;
Term* term = _CLNEW Term(_T("body"),_T("body"));
Query* q = _CLNEW TermQuery(term);
_CLDECDELETE(term);
// test id, bounded on both ends
Filter* f = _CLNEW RangeFilter(_T("id"),minIP,maxIP,T,T);
result = search->search(q, f);
assertEqualsMsg(_T("find all"), numDocs, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),minIP,maxIP,T,F);
result = search->search(q,f);
assertEqualsMsg(_T("all but last"), numDocs-1, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f =_CLNEW RangeFilter(_T("id"),minIP,maxIP,F,T);
result = search->search(q,f);
assertEqualsMsg(_T("all but first"), numDocs-1, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),minIP,maxIP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("all but ends"), numDocs-2, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),medIP,maxIP,T,T);
result = search->search(q,f);
assertEqualsMsg(_T("med and up"), 1+ maxId-medId, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),minIP,medIP,T,T);
result = search->search(q,f);
assertEqualsMsg(_T("up to med"), 1+ medId-minId, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
// unbounded id
f=_CLNEW RangeFilter(_T("id"),minIP,NULL,T,F);
result = search->search(q,f);
assertEqualsMsg(_T("min and up"), numDocs, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),NULL,maxIP,F,T);
result = search->search(q,f);
assertEqualsMsg(_T("max and down"), numDocs, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),minIP,NULL,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("not min, but up"), numDocs-1, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),NULL,maxIP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("not max, but down"), numDocs-1, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),medIP,maxIP,T,F);
result = search->search(q,f);
assertEqualsMsg(_T("med and up, not max"), maxId-medId, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),minIP,medIP,F,T);
result = search->search(q,f);
assertEqualsMsg(_T("not min, up to med"), medId-minId, result->length());
_CLLDELETE(result);
_CLLDELETE(f);
// very small sets
f=_CLNEW RangeFilter(_T("id"),minIP,minIP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("min,min,F,F"), 0, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),medIP,medIP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("med,med,F,F"), 0, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),maxIP,maxIP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("max,max,F,F"), 0, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),minIP,minIP,T,T);
result = search->search(q,f);
assertEqualsMsg(_T("min,min,T,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),NULL,minIP,F,T);
result = search->search(q,f);
assertEqualsMsg(_T("nul,min,F,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),maxIP,maxIP,T,T);
result = search->search(q,f);
assertEqualsMsg(_T("max,max,T,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),maxIP,NULL,T,F);
result = search->search(q,f);
assertEqualsMsg(_T("max,nul,T,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("id"),medIP,medIP,T,T);
result = search->search(q,f);
assertEqualsMsg(_T("med,med,T,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
search->close();
_CLLDELETE(search);
reader->close();
_CLLDELETE(reader);
_CLLDELETE(q);
}
void testRangeFilterRand()
{
IndexReader* reader = IndexReader::open(index);
IndexSearcher* search = _CLNEW IndexSearcher(reader);
std::tstring minRPstr = pad(minR);
const TCHAR* minRP = minRPstr.c_str();
std::tstring maxRPstr = pad(maxR);
const TCHAR* maxRP = maxRPstr.c_str();
size_t numDocs = static_cast<size_t>(reader->numDocs());
assertEqualsMsg(_T("num of docs"), numDocs, 1+ maxId - minId);
Hits* result;
Term* term = _CLNEW Term(_T("body"),_T("body"));
Query* q = _CLNEW TermQuery(term);
_CLDECDELETE(term);
// test extremes, bounded on both ends
Filter* f = _CLNEW RangeFilter(_T("rand"),minRP,maxRP,T,T);
result = search->search(q,f);
assertEqualsMsg(_T("find all"), numDocs, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),minRP,maxRP,T,F);
result = search->search(q,f);
assertEqualsMsg(_T("all but biggest"), numDocs-1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),minRP,maxRP,F,T);
result = search->search(q,f);
assertEqualsMsg(_T("all but smallest"), numDocs-1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),minRP,maxRP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("all but extremes"), numDocs-2, result->length());
_CLLDELETE(result); _CLLDELETE(f);
// unbounded
f=_CLNEW RangeFilter(_T("rand"),minRP,NULL,T,F);
result = search->search(q,f);
assertEqualsMsg(_T("smallest and up"), numDocs, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),NULL,maxRP,F,T);
result = search->search(q,f);
assertEqualsMsg(_T("biggest and down"), numDocs, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),minRP,NULL,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("not smallest, but up"), numDocs-1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),NULL,maxRP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("not biggest, but down"), numDocs-1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
// very small sets
f=_CLNEW RangeFilter(_T("rand"),minRP,minRP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("min,min,F,F"), 0, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),maxRP,maxRP,F,F);
result = search->search(q,f);
assertEqualsMsg(_T("max,max,F,F"), 0, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),minRP,minRP,T,T);
result = search->search(q,f);
assertEqualsMsg(_T("min,min,T,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),NULL,minRP,F,T);
result = search->search(q,f);
assertEqualsMsg(_T("nul,min,F,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),maxRP,maxRP,T,T);
result = search->search(q,f);
assertEqualsMsg(_T("max,max,T,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
f=_CLNEW RangeFilter(_T("rand"),maxRP,NULL,T,F);
result = search->search(q,f);
assertEqualsMsg(_T("max,nul,T,T"), 1, result->length());
_CLLDELETE(result); _CLLDELETE(f);
search->close();
_CLLDELETE(search);
reader->close();
_CLLDELETE(reader);
_CLLDELETE(q);
}
void testRangeQueryRand()
{
// NOTE: uses index build in *super* setUp
IndexReader * pReader = IndexReader::open( index );
IndexSearcher * pSearch = _CLNEW IndexSearcher( pReader );
std::tstring sMinRP = pad(minR);
std::tstring sMaxRP = pad(maxR);
const TCHAR* minRP = sMinRP.c_str();
const TCHAR* maxRP = sMaxRP.c_str();
size_t numDocs = static_cast<size_t>( pReader->numDocs() );
assertEqualsMsg( _T("num of docs"), numDocs, static_cast<size_t>(1+ maxId - minId));
Hits * pResult;
Query * q;
// test extremes, bounded on both ends
q = csrq( _T( "rand" ), minRP, maxRP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "find all" ), numDocs, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), minRP, maxRP, true, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "all but biggest" ), numDocs-1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), minRP, maxRP, false, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "all but smallest" ), numDocs-1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), minRP, maxRP, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "all but extremes" ), numDocs-2, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
// unbounded
q = csrq( _T( "rand" ), minRP, NULL, true, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "smallest and up" ), numDocs, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), NULL, maxRP, false, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "biggest and down" ), numDocs, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), minRP, NULL, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "not smallest, but up" ), numDocs-1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), NULL, maxRP, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "not biggest, but down" ), numDocs-1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
// very small sets
q = csrq( _T( "rand" ), minRP, minRP, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "min,min,F,F" ), 0, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), maxRP, maxRP, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "max,max,F,F" ), 0, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), minRP, minRP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "min,min,T,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), NULL, minRP, false, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "nul,min,F,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), maxRP, maxRP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "max,max,T,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "rand" ), maxRP, NULL, true, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "max,nul,T,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
pSearch->close();
_CLDELETE( pSearch );
pReader->close();
_CLDELETE( pReader );
}
void testRangeQueryId()
{
// NOTE: uses index build in *super* setUp
IndexReader * pReader = IndexReader::open( index );
IndexSearcher * pSearch = _CLNEW IndexSearcher( pReader );
int32_t medId = ((maxId - minId) / 2);
std::tstring sMinIP = pad(minId);
std::tstring sMaxIP = pad(maxId);
std::tstring sMedIP = pad(medId);
const TCHAR* minIP = sMinIP.c_str();
const TCHAR* maxIP = sMaxIP.c_str();
const TCHAR* medIP = sMedIP.c_str();
size_t numDocs = static_cast<size_t>( pReader->numDocs() );
assertEqualsMsg( _T("num of docs"), numDocs, static_cast<size_t>(1+ maxId - minId));
Hits * pResult;
Query * q;
// test id, bounded on both ends
q = csrq( _T( "id" ), minIP, maxIP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "find all" ), numDocs, pResult->length() );
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), minIP, maxIP, true, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "all but last" ), numDocs-1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), minIP, maxIP, false, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "all but first" ), numDocs-1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), minIP, maxIP, false,false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "all but ends" ), numDocs-2, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), medIP, maxIP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "med and up" ), 1+maxId-medId, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), minIP, medIP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "up to med" ), 1+medId-minId, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
// unbounded id
q = csrq( _T( "id" ), minIP, NULL, true, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "min and up" ), numDocs, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), NULL, maxIP, false, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "max and down" ), numDocs, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), minIP, NULL, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "not min, but up" ), numDocs-1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), NULL, maxIP, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "not max, but down" ), numDocs-1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), medIP, maxIP, true, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "med and up, not max" ), maxId-medId, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), minIP, medIP, false,true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "not min, up to med" ), medId-minId, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
// very small sets
q = csrq( _T( "id" ), minIP, minIP, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "min,min,F,F" ), 0, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), medIP, medIP, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "med,med,F,F" ), 0, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id") , maxIP, maxIP, false, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "max,max,F,F" ), 0, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), minIP, minIP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "min,min,T,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), NULL, minIP, false, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "nul,min,F,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), maxIP, maxIP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "max,max,T,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), maxIP, NULL, true, false );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "max,nul,T,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
q = csrq( _T( "id" ), medIP, medIP, true, true );
pResult = pSearch->search( q );
assertEqualsMsg( _T( "med,med,T,T" ), 1, pResult->length());
_CLDELETE( pResult ); _CLDELETE( q );
pSearch->close();
_CLDELETE( pSearch );
pReader->close();
_CLDELETE( pReader );
}
