[Clucene-cvs] SF.net SVN: clucene:[3020] branches/lucene2_3_2/src

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

Revision: 3020
          http://clucene.svn.sourceforge.net/clucene/?rev=3020&view=rev
Author:   ustramooner
Date:     2009-07-08 10:10:52 +0000 (Wed, 08 Jul 2009)

Log Message:
-----------
Isidor Zeuner's fix & test for norms not being merged

Modified Paths:
--------------
    branches/lucene2_3_2/src/core/CLucene/index/IndexReader.h
    branches/lucene2_3_2/src/core/CLucene/index/SegmentMerger.cpp
    branches/lucene2_3_2/src/test/index/TestIndexWriter.cpp

Modified: branches/lucene2_3_2/src/core/CLucene/index/IndexReader.h
===================================================================
--- branches/lucene2_3_2/src/core/CLucene/index/IndexReader.h	2009-07-08 10:10:09 UTC (rev 3019)
+++ branches/lucene2_3_2/src/core/CLucene/index/IndexReader.h	2009-07-08 10:10:52 UTC (rev 3020)
@@ -270,6 +270,8 @@
 	/** Reads the byte-encoded normalization factor for the named field of every
 	*  document.  This is used by the search code to score documents.
 	*
+	* The size of bytes must be the size of the IndexReader->maxDoc()
+	*
 	* @see Field#setBoost(float_t)
 	*/
 	virtual void norms(const TCHAR* field, uint8_t* bytes) = 0;

Modified: branches/lucene2_3_2/src/core/CLucene/index/SegmentMerger.cpp
===================================================================
--- branches/lucene2_3_2/src/core/CLucene/index/SegmentMerger.cpp	2009-07-08 10:10:09 UTC (rev 3019)
+++ branches/lucene2_3_2/src/core/CLucene/index/SegmentMerger.cpp	2009-07-08 10:10:52 UTC (rev 3020)
@@ -769,6 +769,7 @@
             normBuffer.resize(maxDoc);
             memset(normBuffer.values,0,sizeof(uint8_t) * maxDoc);
 			    }
+          reader->norms(fi->name, normBuffer.values);
 
           if (!reader->hasDeletions()) {
             //optimized case for segments without deleted docs

Modified: branches/lucene2_3_2/src/test/index/TestIndexWriter.cpp
===================================================================
--- branches/lucene2_3_2/src/test/index/TestIndexWriter.cpp	2009-07-08 10:10:09 UTC (rev 3019)
+++ branches/lucene2_3_2/src/test/index/TestIndexWriter.cpp	2009-07-08 10:10:52 UTC (rev 3020)
@@ -6,7 +6,66 @@
 ------------------------------------------------------------------------------*/
 #include "test.h"
 #include <iostream>
+#include <sys/stat.h>
+#include <sys/types.h>
 
+//checks if a merged index finds phrases correctly
+void testIWmergePhraseSegments(CuTest *tc){
+	char fsdir[CL_MAX_PATH];
+	sprintf(fsdir,"%s/%s",cl_tempDir, "test.indexwriter");
+	mkdir(fsdir, 0777);
+	SimpleAnalyzer a;
+
+	IndexWriter ndx2(fsdir,&a,true);
+	ndx2.setUseCompoundFile(false);
+	Document doc0;
+	doc0.add(
+		*_CLNEW Field(
+			_T("field0"),
+			_T("value0 value1"),
+			Field::STORE_YES | Field::INDEX_TOKENIZED
+		)
+	);
+	ndx2.addDocument(&doc0);
+	ndx2.optimize();
+	ndx2.close();
+
+	IndexWriter ndx(fsdir,&a,false);
+	ndx.setUseCompoundFile(false);
+	Document doc1;
+	doc1.add(
+		*_CLNEW Field(
+			_T("field0"),
+			_T("value1 value0"),
+			Field::STORE_YES | Field::INDEX_TOKENIZED
+		)
+	);
+	ndx.addDocument(&doc1);
+	ndx.optimize();
+	ndx.close();
+
+	//test the index querying
+	IndexSearcher searcher(fsdir);
+	Query* query0 = QueryParser::parse(
+		_T("\"value0 value1\""),
+		_T("field0"),
+		&a
+	);
+	Hits* hits0 = searcher.search(query0);
+	CLUCENE_ASSERT(hits0->length() > 0);
+	Query* query1 = QueryParser::parse(
+		_T("\"value1 value0\""),
+		_T("field0"),
+		&a
+	);
+	Hits* hits1 = searcher.search(query1);
+	CLUCENE_ASSERT(hits1->length() > 0);
+	_CLDELETE(query0);
+	_CLDELETE(query1);
+	_CLDELETE(hits0);
+	_CLDELETE(hits1);
+}
+
 //checks that adding more than the min_merge value goes ok...
 //checks for a mem leak that used to occur
 void testIWmergeSegments1(CuTest *tc){
@@ -45,6 +104,60 @@
 	_CLDECDELETE(term);
 	_CLDELETE(reader2);
 }
+
+//checks if appending to an index works correctly
+void testIWmergeSegments2(CuTest *tc){
+	char fsdir[CL_MAX_PATH];
+	sprintf(fsdir,"%s/%s",cl_tempDir, "test.indexwriter");
+	mkdir(fsdir, 0777);
+	SimpleAnalyzer a;
+
+	IndexWriter ndx2(fsdir,&a,true);
+	ndx2.setUseCompoundFile(false);
+	Document doc0;
+	doc0.add(
+		*_CLNEW Field(
+			_T("field0"),
+			_T("value0"),
+			Field::STORE_YES | Field::INDEX_TOKENIZED
+		)
+	);
+	ndx2.addDocument(&doc0);
+	ndx2.optimize();
+	ndx2.close();
+
+	IndexWriter ndx(fsdir,&a,false);
+	ndx.setUseCompoundFile(false);
+	Document doc1;
+	doc1.add(
+		*_CLNEW Field(
+			_T("field0"),
+			_T("value1"),
+			Field::STORE_YES | Field::INDEX_TOKENIZED
+		)
+	);
+	ndx.addDocument(&doc1);
+	ndx.optimize();
+	ndx.close();
+
+	//test the ram querying
+	IndexSearcher searcher(fsdir);
+	Term* term0 = _CLNEW Term(_T("field0"),_T("value1"));
+	Query* query0 = QueryParser::parse(_T("value0"),_T("field0"),&a);
+	Hits* hits0 = searcher.search(query0);
+	CLUCENE_ASSERT(hits0->length() > 0);
+	Term* term1 = _CLNEW Term(_T("field0"),_T("value0"));
+	Query* query1 = QueryParser::parse(_T("value1"),_T("field0"),&a);
+	Hits* hits1 = searcher.search(query1);
+	CLUCENE_ASSERT(hits1->length() > 0);
+	_CLDELETE(query0);
+	_CLDELETE(query1);
+	_CLDELETE(hits0);
+  _CLDELETE(hits1);
+	_CLDECDELETE(term0);
+	_CLDECDELETE(term1);
+}
+
 void testAddIndexes(CuTest *tc){
 	char reuters_origdirectory[1024];
   strcpy(reuters_origdirectory, clucene_data_location);
@@ -73,11 +186,14 @@
     CLUCENE_ASSERT(w.docCount()==62); //31 docs in reuters...
   }
 }
+
 CuSuite *testindexwriter(void)
 {
 	CuSuite *suite = CuSuiteNew(_T("CLucene IndexWriter Test"));
 	SUITE_ADD_TEST(suite, testAddIndexes);
 	SUITE_ADD_TEST(suite, testIWmergeSegments1);
+  SUITE_ADD_TEST(suite, testIWmergeSegments2);
+	SUITE_ADD_TEST(suite, testIWmergePhraseSegments);
 
   return suite;
 }


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