[Clucene-cvs] SF.net SVN: clucene: [2242] trunk/test/search

[<ust...@us...>]

Revision: 2242
Author:   ustramooner
Date:     2006-06-29 05:45:03 -0700 (Thu, 29 Jun 2006)
ViewCVS:  http://svn.sourceforge.net/clucene/?rev=2242&view=rev

Log Message:
-----------
added WildcardFilter and bitset fix

Modified Paths:
--------------
    trunk/src/CLucene/index/MultiReader.h
    trunk/src/CLucene/index/SegmentTermDocs.cpp
    trunk/src/CLucene/index/SegmentTermEnum.h
    trunk/src/CLucene/index/Terms.h
    trunk/src/CLucene/search/BooleanScorer.cpp
    trunk/src/CLucene/search/FuzzyQuery.h
    trunk/src/CLucene/search/WildcardQuery.cpp
    trunk/src/CLucene/search/WildcardQuery.h
    trunk/src/CLucene/search/WildcardTermEnum.h
    trunk/src/CLucene/util/BitSet.cpp
    trunk/test/search/TestWildcard.cpp
Modified: trunk/src/CLucene/index/MultiReader.h
===================================================================
--- trunk/src/CLucene/index/MultiReader.h	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/index/MultiReader.h	2006-06-29 12:45:03 UTC (rev 2242)
@@ -94,6 +94,10 @@
 
 		//Closes the set of enumerations in the queue
 		void close();
+
+		
+		const char* getObjectName(){ return MultiTermEnum::getClassName(); }
+		static const char* getClassName(){ return "MultiTermEnum"; }
 	};
 
 

Modified: trunk/src/CLucene/index/SegmentTermDocs.cpp
===================================================================
--- trunk/src/CLucene/index/SegmentTermDocs.cpp	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/index/SegmentTermDocs.cpp	2006-06-29 12:45:03 UTC (rev 2242)
@@ -63,15 +63,11 @@
     TermInfo* ti=NULL;
     
     // use comparison of fieldinfos to verify that termEnum belongs to the same segment as this SegmentTermDocs
-    try{
-      //todo: this is a really dodgy way of doing this
-      //we try and convert the Term Enum to a segment term enum... if its
-      //not a segment term enum, then we throw and assume it is not a segmenttermenum
-      //should have a way of knowing what type the TermEnum is...
+	if ( termEnum->getObjectName() == SegmentTermEnum::getClassName() ){
       SegmentTermEnum* te = (SegmentTermEnum*)termEnum;
       te->fieldInfos = parent->fieldInfos;
       ti = te->getTermInfo();
-    }catch(...){
+    }else{
       ti = parent->tis->get(termEnum->term(false));
     }
     

Modified: trunk/src/CLucene/index/SegmentTermEnum.h
===================================================================
--- trunk/src/CLucene/index/SegmentTermEnum.h	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/index/SegmentTermEnum.h	2006-06-29 12:45:03 UTC (rev 2242)
@@ -93,6 +93,9 @@
         //Returns a clone of this instance
 		SegmentTermEnum* clone() const;
 
+		const char* getObjectName(){ return SegmentTermEnum::getClassName(); }
+		static const char* getClassName(){ return "SegmentTermEnum"; }
+
     private:
 		//Reads the next term in the enumeration
 		Term* readTerm(Term* reuse);

Modified: trunk/src/CLucene/index/Terms.h
===================================================================
--- trunk/src/CLucene/index/Terms.h	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/index/Terms.h	2006-06-29 12:45:03 UTC (rev 2242)
@@ -135,6 +135,11 @@
 		} while (target->compareTo(term(false)) > 0);
 		return true;
 	}
+
+	/**
+	* Because we need to know how to cast the object, we need the objects name.
+	*/
+	virtual const char* getObjectName() = 0;
 };
 
 

Modified: trunk/src/CLucene/search/BooleanScorer.cpp
===================================================================
--- trunk/src/CLucene/search/BooleanScorer.cpp	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/search/BooleanScorer.cpp	2006-06-29 12:45:03 UTC (rev 2242)
@@ -48,23 +48,23 @@
 			// check prohibited & required
 			if ((current->bits & prohibitedMask) == 0 && 
 				(current->bits & requiredMask) == requiredMask) {
-			return true;
+				return true;
+			}
 		}
-	}
-
-	// refill the queue
-	more = false;
-	end += BooleanScorer::BucketTable_SIZE;
-	for (SubScorer* sub = scorers; sub != NULL; sub = sub->next) {
-		Scorer* scorer = sub->scorer;
-		while (!sub->done && scorer->doc() < end) {
-			sub->collector->collect(scorer->doc(), scorer->score());
-			sub->done = !scorer->next();
+	
+		// refill the queue
+		more = false;
+		end += BooleanScorer::BucketTable_SIZE;
+		for (SubScorer* sub = scorers; sub != NULL; sub = sub->next) {
+			Scorer* scorer = sub->scorer;
+			while (!sub->done && scorer->doc() < end) {
+				sub->collector->collect(scorer->doc(), scorer->score());
+				sub->done = !scorer->next();
+			}
+			if (!sub->done) {
+				more = true;
+			}
 		}
-		if (!sub->done) {
-			more = true;
-		}
-	}
 	} while (bucketTable->first != NULL || more);
 
 	return false;
@@ -226,9 +226,6 @@
     BucketTable* table = bucketTable;
     const int32_t i = doc & (BooleanScorer::BucketTable_SIZE-1);
     Bucket* bucket = &table->buckets[i];
-    //if (bucket == NULL)
-    //	bucket = _CLNEW Bucket();
-    //	table.buckets[i] = bucket;
 
     if (bucket->doc != doc) {			  // invalid bucket
       bucket->doc = doc;			  // set doc

Modified: trunk/src/CLucene/search/FuzzyQuery.h
===================================================================
--- trunk/src/CLucene/search/FuzzyQuery.h	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/search/FuzzyQuery.h	2006-06-29 12:45:03 UTC (rev 2242)
@@ -149,7 +149,10 @@
 		*  multiplied by the scale factor
 		*/
 		float_t difference();
-	  
+
+		
+		const char* getObjectName(){ return FuzzyTermEnum::getClassName(); }
+		static const char* getClassName(){ return "FuzzyTermEnum"; }
   };
 CL_NS_END
 #endif

Modified: trunk/src/CLucene/search/WildcardQuery.cpp
===================================================================
--- trunk/src/CLucene/search/WildcardQuery.cpp	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/search/WildcardQuery.cpp	2006-06-29 12:45:03 UTC (rev 2242)
@@ -7,8 +7,10 @@
 #include "CLucene/StdHeader.h"
 #ifndef NO_WILDCARD_QUERY
 #include "WildcardQuery.h"
+#include "CLucene/util/BitSet.h"
 
 CL_NS_USE(index)
+CL_NS_USE(util)
 CL_NS_DEF(search)
 
 
@@ -64,5 +66,82 @@
 		return (this->getBoost() == tq->getBoost())
 			&& getTerm()->equals(tq->getTerm());
 	}
+	
+	
+	
+	
+	
+	
+	
+	
+	
+	
+	
+
+WildcardFilter::WildcardFilter( Term* term )
+{
+	this->term = _CL_POINTER(term);
+}
+
+WildcardFilter::~WildcardFilter()
+{
+	_CLDECDELETE(term);
+}
+
+WildcardFilter::WildcardFilter( const WildcardFilter& copy ) : 
+	term( _CL_POINTER(copy.term) )
+{
+}
+
+Filter* WildcardFilter::clone() const {
+	return _CLNEW WildcardFilter(*this );
+}
+
+
+TCHAR* WildcardFilter::toString()
+{
+	//Instantiate a stringbuffer buffer to store the readable version temporarily
+    CL_NS(util)::StringBuffer buffer;
+    //check if field equal to the field of prefix
+    if( term->field() != NULL != 0 ) {
+	  //Append the field of prefix to the buffer
+      buffer.append(term->field());
+	  //Append a colon
+      buffer.append(_T(":") );
+    }
+    //Append the text of the prefix
+    buffer.append(term->text());
+
+	//Convert StringBuffer buffer to TCHAR block and return it
+    return buffer.toString();
+}
+
+
+/** Returns a BitSet with true for documents which should be permitted in
+search results, and false for those that should not. */
+BitSet* WildcardFilter::bits( IndexReader* reader )
+{
+	BitSet* bts = _CLNEW BitSet( reader->maxDoc() );
+	
+	WildcardTermEnum termEnum (reader, term);
+	TermDocs* termDocs = reader->termDocs();
+
+	try{
+		do{
+			termDocs->seek(&termEnum);
+
+			while (termDocs->next()) {
+			  bts->set(termDocs->doc());
+			}
+		}while(termEnum.next());
+	} _CLFINALLY(
+      termDocs->close();
+      _CLDELETE(termDocs);
+      termEnum.close();
+    )
+
+	return bts;
+}
+
 CL_NS_END
 #endif

Modified: trunk/src/CLucene/search/WildcardQuery.h
===================================================================
--- trunk/src/CLucene/search/WildcardQuery.h	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/search/WildcardQuery.h	2006-06-29 12:45:03 UTC (rev 2242)
@@ -44,7 +44,29 @@
 		bool equals(Query* other) const;
 		Query* clone() const;
     };
+    
+    
+    
+class WildcardFilter: public Filter 
+{
+private:
+	CL_NS(index)::Term* term;
+protected:
+	WildcardFilter( const WildcardFilter& copy );
+	
+public:
+	WildcardFilter(CL_NS(index)::Term* term);
+	~WildcardFilter();
 
+	/** Returns a BitSet with true for documents which should be permitted in
+	search results, and false for those that should not. */
+	CL_NS(util)::BitSet* bits( CL_NS(index)::IndexReader* reader );
+	
+	Filter* clone() const;
+	TCHAR* toString();
+};
+
+
 CL_NS_END
 #endif
 #endif

Modified: trunk/src/CLucene/search/WildcardTermEnum.h
===================================================================
--- trunk/src/CLucene/search/WildcardTermEnum.h	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/search/WildcardTermEnum.h	2006-06-29 12:45:03 UTC (rev 2242)
@@ -61,6 +61,9 @@
         static bool wildcardEquals(const TCHAR* pattern, int32_t patternLen, int32_t patternIdx, const TCHAR* str, int32_t strLen, int32_t stringIdx);
 
         void close();
+
+		const char* getObjectName(){ return WildcardTermEnum::getClassName(); }
+		static const char* getClassName(){  return "WildcardTermEnum"; }
     };
 CL_NS_END
 #endif

Modified: trunk/src/CLucene/util/BitSet.cpp
===================================================================
--- trunk/src/CLucene/util/BitSet.cpp	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/src/CLucene/util/BitSet.cpp	2006-06-29 12:45:03 UTC (rev 2242)
@@ -22,7 +22,8 @@
 
 ///Create a bitset with the specified size
 BitSet::BitSet ( int32_t size ):
-  _size(size)
+  _size(size),
+  _count(-1)
 {
 	int32_t len = (_size >> 3) + 1;
 	bits = _CL_NEWARRAY(uint8_t, len);

Modified: trunk/test/search/TestWildcard.cpp
===================================================================
--- trunk/test/search/TestWildcard.cpp	2006-06-29 12:43:33 UTC (rev 2241)
+++ trunk/test/search/TestWildcard.cpp	2006-06-29 12:45:03 UTC (rev 2242)
@@ -8,6 +8,27 @@
 
 #ifndef NO_WILDCARD_QUERY
 
+void _testWildcard(CuTest* tc, IndexSearcher* searcher, const TCHAR* qt, int expectedLen){
+	Term* term = _CLNEW Term(_T("body"), qt);
+	Query* query = _CLNEW WildcardQuery(term);
+
+	//test the wildcardquery
+	Hits* result = searcher->search(query);
+	CLUCENE_ASSERT(expectedLen == result->length());
+	_CLDELETE(result);
+	_CLDELETE(query);
+
+
+	//now test wildcardfilter
+	Filter* filter = _CLNEW WildcardFilter(term);
+	BitSet* bits = filter->bits(searcher->getReader());
+	CLUCENE_ASSERT(expectedLen == bits->count());
+	_CLDELETE(filter);
+	_CLDELETE(bits);
+
+	_CLDECDELETE(term);
+}
+
 	void testFuzzyQuery(CuTest *tc){
 		RAMDirectory ram;
 		
@@ -44,7 +65,7 @@
 		_CLDELETE(doc);
 
 		//---
-		writer->optimize();
+		//writer->optimize(); //test unoptimized
 		writer->close();
 		_CLDELETE(writer);
 
@@ -86,44 +107,21 @@
 		IndexReader* reader = IndexReader::open(&indexStore);
 		IndexSearcher* searcher = _CLNEW IndexSearcher(reader);
 
+		_testWildcard(tc, searcher, _T("metal*"), 2);
+		_testWildcard(tc, searcher, _T("m*tal"), 1);
+		_testWildcard(tc, searcher, _T("m*tal*"), 2);
+			
+
 		Term* term = _CLNEW Term(_T("body"), _T("metal"));
-			Query* query1 = _CLNEW TermQuery(term);
+		Query* query1 = _CLNEW TermQuery(term);
 		_CLDECDELETE(term);
 
-		term = _CLNEW Term(_T("body"), _T("metal*"));
-		Query* query2 = _CLNEW WildcardQuery(term);
-		_CLDECDELETE(term);
-			
-		term = _CLNEW Term(_T("body"), _T("m*tal"));
-		Query* query3 = _CLNEW WildcardQuery(term);
-		_CLDECDELETE(term);
-			
-		term = _CLNEW Term(_T("body"), _T("m*tal*"));
-		Query* query4 = _CLNEW WildcardQuery(term);
-		_CLDECDELETE(term);
-
-    	Hits* result = NULL;
-	
-    	result = searcher->search(query1);
+    	Hits* result = searcher->search(query1);
     	CLUCENE_ASSERT(1 == result->length());
 		_CLDELETE(result);
 		_CLDELETE(query1);
 
-    	result = searcher->search(query2);
-    	CLUCENE_ASSERT(2 == result->length());
-		_CLDELETE(result);
-		_CLDELETE(query2);
 	
-    	result = searcher->search(query3);
-    	CLUCENE_ASSERT(1 == result->length());
-		_CLDELETE(result);
-		_CLDELETE(query3);
-	
-    	result = searcher->search(query4);
-    	CLUCENE_ASSERT(2 == result->length());
-		_CLDELETE(result);
-		_CLDELETE(query4);
-
 		indexStore.close();
 		searcher->close();
 		reader->close();
@@ -155,45 +153,12 @@
 		IndexReader* reader = IndexReader::open(&indexStore);
 		IndexSearcher* searcher = _CLNEW IndexSearcher(reader);
 		
-		Term* term = _CLNEW Term(_T("body"), _T("m?tal"));
-		Query* query1 = _CLNEW WildcardQuery(term); 
-		_CLDECDELETE(term);
 
-			term = _CLNEW Term(_T("body"), _T("metal?"));
-		Query* query2 = _CLNEW WildcardQuery(term);
-			_CLDECDELETE(term);
+		_testWildcard(tc, searcher, _T("m?tal"), 1);
+		_testWildcard(tc, searcher, _T("metal?"), 2);
+		_testWildcard(tc, searcher, _T("metals?"), 1);
+		_testWildcard(tc, searcher, _T("m?t?ls"), 3);
 
-			term = _CLNEW Term(_T("body"), _T("metals?"));
-		Query* query3 = _CLNEW WildcardQuery(term);
-		_CLDECDELETE(term);
-
-		term = _CLNEW Term(_T("body"), _T("m?t?ls"));
-		Query* query4 = _CLNEW WildcardQuery(term);
-		_CLDECDELETE(term);
-
-
-    	Hits* result = NULL;
-	
-    	result = searcher->search(query1);
-    	CLUCENE_ASSERT(1 == result->length());
-		_CLDELETE(result);
-		_CLDELETE(query1);
-
-    	result = searcher->search(query2);
-    	CLUCENE_ASSERT(2 == result->length());
-		_CLDELETE(result);
-		_CLDELETE(query2);
-	
-    	result = searcher->search(query3);
-    	CLUCENE_ASSERT(1 == result->length());
-		_CLDELETE(result);
-		_CLDELETE(query3);
-	
-    	result = searcher->search(query4);
-    	CLUCENE_ASSERT(3 == result->length());
-		_CLDELETE(result);
-		_CLDELETE(query4);
-	
 		indexStore.close();
 		reader->close();
 		searcher->close();


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.