void _TestSearchesRun(CuTest *tc, Analyzer* analyzer, Searcher* search, const TCHAR* qry){
Query* q = NULL;
Hits* h = NULL;
try{
q = QueryParser::parse(qry , _T("contents"), analyzer);
if ( q != NULL ){
h = search->search( q );
if ( h->length() > 0 ){
//check for explanation memory leaks...
CL_NS(search)::Explanation expl1;
search->explain(q, h->id(0), &expl1);
TCHAR* tmp = expl1.toString();
_CLDELETE_CARRAY(tmp);
if ( h->length() > 1 ){ //do a second one just in case
CL_NS(search)::Explanation expl2;
search->explain(q, h->id(1), &expl2);
tmp = expl2.toString();
_CLDELETE_CARRAY(tmp);
}
}
}
}catch(CLuceneError& err){
CuFail(tc,_T("Error: %s\n"), err.twhat());
}catch(...){
CuFail(tc,_T("Error: unknown\n"));
}
_CLDELETE(h);
_CLDELETE(q);
}
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);
}
/**
Returns a Entry.
\id The key of the entry.
*/
Response * ZefaniaLex::getEntry(const QString &key)
{
try {
if(!hasIndex()) {
if(buildIndex() != 0) {
return new StringResponse(QObject::tr("Cannot build index."));
}
}
const QString index = indexPath();
const QString queryText = "key:" + key;
const TCHAR* stop_words[] = { nullptr };
standard::StandardAnalyzer analyzer(stop_words);
IndexReader* reader = IndexReader::open(index.toStdString().c_str());
IndexSearcher s(reader);
Query* q = QueryParser::parse(SearchTools::toTCHAR(queryText), _T("content"), &analyzer);
Hits* h = s.search(q);
QString ret = "";
for(size_t i = 0; i < h->length(); i++) {
Document* doc = &h->doc(i);
if(!ret.isEmpty())
ret.append("<hr /> ");
ret.append(SearchTools::toQString(doc->get(_T("content"))));
}
return ret.isEmpty() ? new StringResponse(QObject::tr("Nothing found for %1").arg(key)) : new StringResponse(ret);
} catch(...) {
return new StringResponse(QString());
}
}
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 SearchFilesC(const char* index, const char* fobizzle){
standard::StandardAnalyzer analyzer;
char line[80];
TCHAR tline[80];
TCHAR* buf;
IndexReader* reader = IndexReader::open(index);
//printf("Enter query string: ");
strncpy(line,fobizzle,80);
//line[strlen(line)-1]=0;
IndexReader* newreader = reader->reopen();
if ( newreader != reader ){
_CLLDELETE(reader);
reader = newreader;
}
IndexSearcher s(reader);
STRCPY_AtoT(tline,line,80);
Query* q = QueryParser::parse(tline,_T("contents"),&analyzer);
buf = q->toString(_T("contents"));
_tprintf(_T("Searching for: %S\n\n"), buf);
_CLDELETE_LCARRAY(buf);
uint64_t str = Misc::currentTimeMillis();
Hits* h = s.search(q);
uint32_t srch = (int32_t)(Misc::currentTimeMillis() - str);
str = Misc::currentTimeMillis();
//SearchData search[h->length()];
for ( size_t i=0; i < h->length(); i++ ){
Document* doc = &h->doc(i);
//const TCHAR* buf = doc.get(_T("contents"));
_tprintf(_T("%d. %S - %f\n"), i, doc->get(_T("path")), h->score(i));
//search[i].set_path(doc->get(_T("path")));
}
printf("\n\nSearch took: %d ms.\n", srch);
printf("Screen dump took: %d ms.\n\n", (int32_t)(Misc::currentTimeMillis() - str));
_CLLDELETE(h);
_CLLDELETE(q);
s.close();
reader->close();
_CLLDELETE(reader);
};
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 SearchFiles(const char* index){
standard::StandardAnalyzer analyzer;
char line[80];
TCHAR tline[80];
TCHAR* buf;
IndexReader* reader = IndexReader::open(index);
while (true) {
printf("Enter query string: ");
char* tmp = fgets(line,80,stdin);
if ( tmp == NULL ) continue;
line[strlen(line)-1]=0;
IndexReader* newreader = reader->reopen();
if ( newreader != reader ){
_CLLDELETE(reader);
reader = newreader;
}
IndexSearcher s(reader);
if ( strlen(line) == 0 )
break;
STRCPY_AtoT(tline,line,80);
Query* q = QueryParser::parse(tline,_T("contents"),&analyzer);
buf = q->toString(_T("contents"));
_tprintf(_T("Searching for: %s\n\n"), buf);
_CLDELETE_LCARRAY(buf);
uint64_t str = Misc::currentTimeMillis();
Hits* h = s.search(q);
uint32_t srch = (int32_t)(Misc::currentTimeMillis() - str);
str = Misc::currentTimeMillis();
for ( size_t i=0;i<h->length();i++ ){
Document* doc = &h->doc(i);
//const TCHAR* buf = doc.get(_T("contents"));
_tprintf(_T("%d. %s - %f\n"), i, doc->get(_T("path")), h->score(i));
//print result to web interface:
LINH_PRINT_WEB(i,doc->get(_T("path")),h->score(i));
}
printf("\n\nSearch took: %d ms.\n", srch);
printf("Screen dump took: %d ms.\n\n", (int32_t)(Misc::currentTimeMillis() - str));
_CLLDELETE(h);
_CLLDELETE(q);
s.close();
}
reader->close();
_CLLDELETE(reader);
}
QMap< int, float >
FuzzyIndex::searchAlbum( const Tomahawk::query_ptr& query )
{
Q_ASSERT( query->isFullTextQuery() );
QMutexLocker lock( &m_mutex );
QMap< int, float > resultsmap;
try
{
if ( !m_luceneReader )
{
if ( !IndexReader::indexExists( TomahawkUtils::appDataDir().absoluteFilePath( "tomahawk.lucene" ).toStdString().c_str() ) )
{
qDebug() << Q_FUNC_INFO << "index didn't exist.";
return resultsmap;
}
m_luceneReader = IndexReader::open( m_luceneDir );
m_luceneSearcher = _CLNEW IndexSearcher( m_luceneReader );
}
QueryParser parser( _T( "album" ), m_analyzer );
QString escapedName = QString::fromWCharArray( parser.escape( DatabaseImpl::sortname( query->fullTextQuery() ).toStdWString().c_str() ) );
Query* qry = _CLNEW FuzzyQuery( _CLNEW Term( _T( "album" ), escapedName.toStdWString().c_str() ) );
Hits* hits = m_luceneSearcher->search( qry );
for ( uint i = 0; i < hits->length(); i++ )
{
Document* d = &hits->doc( i );
float score = hits->score( i );
int id = QString::fromWCharArray( d->get( _T( "albumid" ) ) ).toInt();
if ( score > 0.30 )
{
resultsmap.insert( id, score );
// tDebug() << "Index hit:" << id << score;
}
}
delete hits;
delete qry;
}
catch( CLuceneError& error )
{
tDebug() << "Caught CLucene error:" << error.what();
Q_ASSERT( false );
}
return resultsmap;
}
static int EIO_Search(eio_req* req)
{
search_baton_t* baton = static_cast<search_baton_t*>(req->data);
standard::StandardAnalyzer analyzer;
IndexReader* reader = 0;
try {
reader = IndexReader::open(*(*baton->index));
} catch (CLuceneError& E) {
baton->error.assign(E.what());
return 0;
} catch(...) {
baton->error = "Got an unknown exception";
return 0;
}
IndexReader* newreader = reader->reopen();
if ( newreader != reader ) {
delete reader;
reader = newreader;
}
IndexSearcher s(reader);
try {
TCHAR* searchString = STRDUP_AtoT(*(*baton->search));
Query* q = QueryParser::parse(searchString, _T(""), &analyzer);
Hits* hits = s.search(q);
HandleScope scope;
//_CLDELETE(q);
free(searchString);
// Build the result array
Local<v8::Array> resultArray = v8::Array::New();
for (size_t i=0; i < hits->length(); i++) {
Document& doc(hits->doc(i));
// {"id":"ab34", "score":1.0}
Local<Object> resultObject = Object::New();
// TODO: This dup might be a leak
resultObject->Set(String::New("id"), String::New(STRDUP_TtoA(doc.get(_T("_id")))));
resultObject->Set(String::New("score"), Number::New(hits->score(i)));
resultArray->Set(i, resultObject);
}
baton->results = Persistent<v8::Array>::New(resultArray);
} catch (CLuceneError& E) {
baton->error.assign(E.what());
} catch(...) {
baton->error = "Got an unknown exception";
}
return 0;
}
vector<pair<string,uint32_t> >
CLuceneIndexReader::histogram(const string& query,
const string& fieldname, const string& labeltype) {
vector<pair<string,uint32_t> > h;
if (!checkReader()) {
return h;
}
Strigi::QueryParser parser;
Strigi::Query q = parser.buildQuery(query);
Query* bq = p->createQuery(q);
IndexSearcher searcher(reader);
Hits* hits = 0;
int s = 0;
try {
hits = searcher.search(bq);
s = hits->length();
} catch (CLuceneError& err) {
fprintf(stderr, "could not query: %s\n", err.what());
}
wstring field = utf8toucs2(fieldname);
int32_t max = INT_MIN;
int32_t min = INT_MAX;
vector<int32_t> values;
values.reserve(s);
char* end;
for (int i = 0; i < s; ++i) {
Document *d = &hits->doc(i);
const TCHAR* v = d->get(field.c_str());
if (v) {
int val = (int)strtol(wchartoutf8( v ).c_str(), &end, 10);
if ( *end != 0) {
_CLDELETE(hits);
return h;
}
values.push_back(val);
max = (max>val) ?max :val;
min = (min<val) ?min :val;
}
}
if (hits) {
_CLDELETE(hits);
}
searcher.close();
_CLDELETE(bq);
if (fieldname == FieldRegister::mtimeFieldName || labeltype == "time") {
return makeTimeHistogram(values);
} else {
return makeHistogram(values, min, max);
}
}
int32_t
CLuceneIndexReader::countHits(const Strigi::Query& q) {
if (!checkReader()) return -1;
// if the query is empty, we return the number of files in the index
if (q.term().string().size() == 0 && q.subQueries().size() == 0) {
return countDocuments();
}
Query* bq = p->createQuery(q);
if (reader == 0) {
return 0;
}
IndexSearcher searcher(reader);
vector<IndexedDocument> results;
Hits* hits = 0;
int s = 0;
try {
hits = searcher.search(bq);
s = hits->length();
} catch (CLuceneError& err) {
/* HitCounter counter;
QueryFilter* filter = _CLNEW QueryFilter(&bq);
try {
BitSet* bits = filter->bits(reader);
int32_t n = bits->size();
for (int32_t i=0; i<n; ++i) {
if (bits->get(i)) s++;
}
} catch (CLuceneError& err2) {
printf("ccould not query: %s\n", err.what());
}
try {
searcher._search(0, filter, &counter);
} catch (CLuceneError& err2) {
printf("ccould not query: %s\n", err.what());
}
s = counter.count();
printf("counted %i hits\n", count);
// try to do a constant score query
//QueryFilter* filter = _CLNEW QueryFilter(&bq);
ConstantScoreQuery csq(filter);*/
fprintf(stderr, "could not query: %s\n", err.what());
}
delete hits;
searcher.close();
_CLDELETE(bq);
return s;
}
vector<IndexedDocument>
CLuceneIndexReader::query(const Strigi::Query& q, int off, int max) {
vector<IndexedDocument> results;
if (!checkReader()) {
return results;
}
// handle special commands
if (q.fields().size() == 1 && q.fields()[0].empty()
&& q.term().string().substr(0, 14) == "strigispecial:") {
return p->strigiSpecial(q.term().string());
}
Query* bq = p->createQuery(q);
IndexSearcher searcher(reader);
Hits* hits = 0;
int s = 0;
try {
hits = searcher.search(bq);
s = hits->length();
} catch (CLuceneError& err) {
fprintf(stderr, "could not query: %s\n", err.what());
}
if (off < 0) off = 0;
max += off;
if (max < 0) max = s;
if (max > s) max = s;
if (max > off) {
results.reserve(max-off);
}
for (int i = off; i < max; ++i) {
Document *d = &hits->doc(i);
IndexedDocument doc;
doc.score = hits->score(i);
DocumentFieldEnumeration* e = d->fields();
while (e->hasMoreElements()) {
Field* f = e->nextElement();
Private::addField(f, doc);
}
results.push_back(doc);
_CLDELETE(e);
}
if (hits) {
_CLDELETE(hits);
}
searcher.close();
_CLDELETE(bq);
return results;
}
void
CLuceneIndexReader::getChildren(const std::string& parent,
std::map<std::string, time_t>& children) {
children.clear();
// force a fresh reader. This is important because the function
// getChildren is essential for updating the index
if ( !checkReader(true) ) {
return;
}
// build a query
Term* t = Private::createKeywordTerm(Private::parentlocation(),
parent);
Query* q = _CLNEW TermQuery(t);
_CLDECDELETE(t);
IndexSearcher searcher(reader);
Hits* hits = 0;
int nhits = 0;
try {
hits = searcher.search(q);
nhits = hits->length();
} catch (CLuceneError& err) {
fprintf(stderr, "could not query: %s\n", err.what());
}
const TCHAR* mtime = mapId(Private::mtime());
for (int i = 0; i < nhits; ++i) {
Document* d = &hits->doc(i);
const TCHAR* v = d->get(mtime);
// check that mtime is defined for this document
if (v) {
time_t mtime = atoi(wchartoutf8( v ).c_str());
v = d->get(Private::systemlocation());
if (v) {
children[wchartoutf8( v )] = mtime;
}
}
}
if (hits) {
_CLDELETE(hits);
}
searcher.close();
_CLDELETE(q);
}
/// 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());
}
}
/**
Returns a Entry.
\id The key of the entry.
*/
QString ZefaniaLex::getEntry(const QString &key)
{
try {
if(!hasIndex()) {
if(buildIndex() != 0) {
return QObject::tr("Cannot build index.");
}
}
const QString index = indexPath();
const QString queryText = "key:" + key;
const TCHAR* stop_words[] = { NULL };
standard::StandardAnalyzer analyzer(stop_words);
IndexReader* reader = IndexReader::open(index.toStdString().c_str());
IndexSearcher s(reader);
#ifdef OBV_USE_WSTRING
Query* q = QueryParser::parse(queryText.toStdWString().c_str(), _T("content"), &analyzer);
#else
Query* q = QueryParser::parse(reinterpret_cast<const wchar_t *>(queryText.utf16()), _T("content"), &analyzer);
#endif
Hits* h = s.search(q);
QString ret = "";
for(size_t i = 0; i < h->length(); i++) {
Document* doc = &h->doc(i);
if(!ret.isEmpty())
ret.append("<hr /> ");
#ifdef OBV_USE_WSTRING
ret.append(QString::fromWCharArray(doc->get(_T("content"))));
#else
ret.append(QString::fromUtf16((const ushort*)doc->get(_T("content"))));
#endif
}
return ret.isEmpty() ? QObject::tr("Nothing found for %1").arg(key) : ret;
} catch(...) {
return QString();
}
}
QMap< int, float >
FuzzyIndex::search( const Tomahawk::query_ptr& query )
{
QMutexLocker lock( &m_mutex );
QMap< int, float > resultsmap;
try
{
if ( !m_luceneReader )
{
if ( !IndexReader::indexExists( TomahawkUtils::appDataDir().absoluteFilePath( "tomahawk.lucene" ).toStdString().c_str() ) )
{
qDebug() << Q_FUNC_INFO << "index didn't exist.";
return resultsmap;
}
m_luceneReader = IndexReader::open( m_luceneDir );
m_luceneSearcher = _CLNEW IndexSearcher( m_luceneReader );
}
float minScore;
const TCHAR** fields = 0;
MultiFieldQueryParser parser( fields, m_analyzer );
BooleanQuery* qry = _CLNEW BooleanQuery();
if ( query->isFullTextQuery() )
{
QString escapedQuery = QString::fromWCharArray( parser.escape( DatabaseImpl::sortname( query->fullTextQuery() ).toStdWString().c_str() ) );
Term* term = _CLNEW Term( _T( "track" ), escapedQuery.toStdWString().c_str() );
Query* fqry = _CLNEW FuzzyQuery( term );
qry->add( fqry, true, BooleanClause::SHOULD );
term = _CLNEW Term( _T( "artist" ), escapedQuery.toStdWString().c_str() );
fqry = _CLNEW FuzzyQuery( term );
qry->add( fqry, true, BooleanClause::SHOULD );
term = _CLNEW Term( _T( "fulltext" ), escapedQuery.toStdWString().c_str() );
fqry = _CLNEW FuzzyQuery( term );
qry->add( fqry, true, BooleanClause::SHOULD );
minScore = 0.00;
}
else
{
QString track = QString::fromWCharArray( parser.escape( DatabaseImpl::sortname( query->track() ).toStdWString().c_str() ) );
QString artist = QString::fromWCharArray( parser.escape( DatabaseImpl::sortname( query->artist() ).toStdWString().c_str() ) );
// QString album = QString::fromWCharArray( parser.escape( query->album().toStdWString().c_str() ) );
Term* term = _CLNEW Term( _T( "track" ), track.toStdWString().c_str() );
Query* fqry = _CLNEW FuzzyQuery( term );
qry->add( fqry, true, BooleanClause::MUST );
term = _CLNEW Term( _T( "artist" ), artist.toStdWString().c_str() );
fqry = _CLNEW FuzzyQuery( term );
qry->add( fqry, true, BooleanClause::MUST );
minScore = 0.00;
}
Hits* hits = m_luceneSearcher->search( qry );
for ( uint i = 0; i < hits->length(); i++ )
{
Document* d = &hits->doc( i );
float score = hits->score( i );
int id = QString::fromWCharArray( d->get( _T( "trackid" ) ) ).toInt();
if ( score > minScore )
{
resultsmap.insert( id, score );
// tDebug() << "Index hit:" << id << score << QString::fromWCharArray( ((Query*)qry)->toString() );
}
}
delete hits;
delete qry;
}
catch( CLuceneError& error )
{
tDebug() << "Caught CLucene error:" << error.what();
Q_ASSERT( false );
}
return resultsmap;
}
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);
}
SearchData *SearchFilesC(const char* index, const char* fobizzle){
standard::StandardAnalyzer analyzer;
char line[80];
TCHAR tline[80];
TCHAR* buf;
IndexReader* reader = IndexReader::open(index);
//printf("Enter query string: ");
strncpy(line,fobizzle,80);
//line[strlen(line)-1]=0;
IndexReader* newreader = reader->reopen();
if ( newreader != reader ){
_CLLDELETE(reader);
reader = newreader;
}
IndexSearcher s(reader);
STRCPY_AtoT(tline,line,80);
Query* q = QueryParser::parse(tline,_T("contents"),&analyzer);
buf = q->toString(_T("contents"));
_tprintf(_T("Searching for: %S\n\n"), buf);
_CLDELETE_LCARRAY(buf);
uint64_t str = Misc::currentTimeMillis();
Hits* h = s.search(q);
uint32_t srch = (int32_t)(Misc::currentTimeMillis() - str);
str = Misc::currentTimeMillis();
arrsize = h->length();
SearchData *search = new SearchData[h->length()];
for ( size_t i=0; i < h->length(); i++ ){
Document* doc = &h->doc(i);
//const TCHAR* buf = doc.get(_T("contents"));
_tprintf(_T("%d. %S - %f\n"), i, doc->get(_T("path")), h->score(i));
//const TCHAR* wtfbatman;
//wtfbatman = doc->get(_T("path"));
//search[(int)i].score = h->score(i);
//printf("Adding %S %d\n", search[i].path, i);
char *wtfbbq;
wtfbbq = new char[100];
sprintf(wtfbbq,"%S %f", doc->get(_T("path")), h->score(i));
search[(int)i].path = wtfbbq;
//sprintf(str,"%S", String::New((char*)doc->get(_T("path")),5));
//printf("PIZZA %s\n", wtfbbq);
//sprintf(search[i].path,"%S",(const char*)doc->get(_T("path")));
//printf("segfault");
//strcpy(search[i].path,(const char*)doc->get(_T("path")));
}
printf("\n\nSearch took: %d ms.\n", srch);
printf("Screen dump took: %d ms.\n\n", (int32_t)(Misc::currentTimeMillis() - str));
//_CLLDELETE(h);
//_CLLDELETE(q);
//s.close();
//reader->close();
//_CLLDELETE(reader);
//printf("Testing %S\n\n", search[0].path);
return search;
};
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 );
}
QMap< int, float >
FuzzyIndex::search( const QString& table, const QString& name, bool fulltext )
{
QMutexLocker lock( &m_mutex );
QMap< int, float > resultsmap;
try
{
if ( !m_luceneReader )
{
if ( !IndexReader::indexExists( TomahawkUtils::appDataDir().absoluteFilePath( "tomahawk.lucene" ).toStdString().c_str() ) )
{
qDebug() << Q_FUNC_INFO << "index didn't exist.";
return resultsmap;
}
m_luceneReader = IndexReader::open( m_luceneDir );
m_luceneSearcher = _CLNEW IndexSearcher( m_luceneReader );
}
if ( name.isEmpty() )
return resultsmap;
Hits* hits = 0;
Query* qry = 0;
QueryParser parser( table.toStdWString().c_str(), m_analyzer );
if ( fulltext )
{
QString escapedName = QString::fromWCharArray( parser.escape( name.toStdWString().c_str() ) );
QStringList sl = DatabaseImpl::sortname( escapedName ).split( " ", QString::SkipEmptyParts );
qry = parser.parse( QString( "%1:%2~" ).arg( table ).arg( sl.join( "~ " ) ).toStdWString().c_str() );
}
else
{
// qry = _CLNEW FuzzyQuery( _CLNEW Term( table.toStdWString().c_str(), DatabaseImpl::sortname( name ).toStdWString().c_str() ) );
QString escapedName = QString::fromWCharArray( parser.escape( name.toStdWString().c_str() ) );
QStringList sl = DatabaseImpl::sortname( escapedName ).split( " ", QString::SkipEmptyParts );
qry = parser.parse( QString( "%1:\"%2\"~" ).arg( table ).arg( sl.join( " " ) ).toStdWString().c_str() );
}
hits = m_luceneSearcher->search( qry );
for ( uint i = 0; i < hits->length(); i++ )
{
Document* d = &hits->doc( i );
float score = hits->score( i );
int id = QString::fromWCharArray( d->get( _T( "id" ) ) ).toInt();
QString result = QString::fromWCharArray( d->get( table.toStdWString().c_str() ) );
if ( DatabaseImpl::sortname( result ) == DatabaseImpl::sortname( name ) )
score = 1.0;
else
score = qMin( score, (float)0.99 );
if ( score > 0.20 )
{
resultsmap.insert( id, score );
// qDebug() << "Hitres:" << result << id << score << table << name;
}
}
delete hits;
delete qry;
}
catch( CLuceneError& error )
{
tDebug() << "Caught CLucene error:" << error.what();
Q_ASSERT( false );
}
return resultsmap;
}
void
CLuceneIndexReader::getHits(const Strigi::Query& q,
const std::vector<std::string>& fields,
const std::vector<Strigi::Variant::Type>& types,
std::vector<std::vector<Strigi::Variant> >& result, int off, int max) {
result.clear();
if (!checkReader() || types.size() < fields.size()) {
return;
}
vector<string> fullFields;
fullFields.resize(fields.size());
for (size_t i = 0; i < fields.size(); i++) {
if (fields[i].compare(0, 6, "xesam:") == 0) {
fullFields[i].assign(
"http://freedesktop.org/standards/xesam/1.0/core#"
+ fields[i].substr(6));
} else if (fields[i].compare(0, 4, "nie:") == 0) {
fullFields[i].assign(
"http://www.semanticdesktop.org/ontologies/2007/01/19/nie#"
+ fields[i].substr(4));
} else {
fullFields[i].assign(fields[i]);
}
}
// if the query is empty, we return the number of files in the index
if (q.term().string().size() == 0 && q.subQueries().size() == 0) {
getDocuments(fullFields, types, result, off, max);
return;
}
Query* bq = p->createQuery(q);
IndexSearcher searcher(reader);
Hits* hits = 0;
int s = 0;
try {
hits = searcher.search(bq);
s = hits->length();
} catch (CLuceneError& err) {
fprintf(stderr, "could not query: %s\n", err.what());
}
if (off < 0) off = 0;
max += off;
if (max < 0) max = s;
if (max > s) max = s;
if (max > off) {
result.reserve(max-off);
}
result.resize(max-off);
for (int i = off; i < max; ++i) {
Document *d = &hits->doc(i);
vector<Variant>& doc = result[i-off];
doc.clear();
doc.resize(fields.size());
DocumentFieldEnumeration* e = d->fields();
while (e->hasMoreElements()) {
Field* field = e->nextElement();
string name(wchartoutf8(field->name()));
for (uint j = 0; j < fullFields.size(); ++j) {
if (fullFields[j] == name) {
doc[j] = p->getFieldValue(field, types[j]);
}
}
}
_CLDELETE(e);
}
if (hits) {
_CLDELETE(hits);
}
searcher.close();
_CLDELETE(bq);
}