public virtual void TestEmptyIndex()
{
// creating two directories for indices
Directory indexStoreA = new MockRAMDirectory();
Directory indexStoreB = new MockRAMDirectory();
// creating a document to store
Document lDoc = new Document();
lDoc.Add(new Field("fulltext", "Once upon a time.....", Field.Store.YES, Field.Index.ANALYZED));
lDoc.Add(new Field("id", "doc1", Field.Store.YES, Field.Index.NOT_ANALYZED));
lDoc.Add(new Field("handle", "1", Field.Store.YES, Field.Index.NOT_ANALYZED));
// creating a document to store
Document lDoc2 = new Document();
lDoc2.Add(new Field("fulltext", "in a galaxy far far away.....", Field.Store.YES, Field.Index.ANALYZED));
lDoc2.Add(new Field("id", "doc2", Field.Store.YES, Field.Index.NOT_ANALYZED));
lDoc2.Add(new Field("handle", "1", Field.Store.YES, Field.Index.NOT_ANALYZED));
// creating a document to store
Document lDoc3 = new Document();
lDoc3.Add(new Field("fulltext", "a bizarre bug manifested itself....", Field.Store.YES, Field.Index.ANALYZED));
lDoc3.Add(new Field("id", "doc3", Field.Store.YES, Field.Index.NOT_ANALYZED));
lDoc3.Add(new Field("handle", "1", Field.Store.YES, Field.Index.NOT_ANALYZED));
// creating an index writer for the first index
IndexWriter writerA = new IndexWriter(indexStoreA, new StandardAnalyzer(Util.Version.LUCENE_CURRENT), true, IndexWriter.MaxFieldLength.LIMITED);
// creating an index writer for the second index, but writing nothing
IndexWriter writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(Util.Version.LUCENE_CURRENT), true, IndexWriter.MaxFieldLength.LIMITED);
//--------------------------------------------------------------------
// scenario 1
//--------------------------------------------------------------------
// writing the documents to the first index
writerA.AddDocument(lDoc);
writerA.AddDocument(lDoc2);
writerA.AddDocument(lDoc3);
writerA.Optimize();
writerA.Close();
// closing the second index
writerB.Close();
// creating the query
QueryParser parser = new QueryParser(Util.Version.LUCENE_CURRENT, "fulltext", new StandardAnalyzer(Util.Version.LUCENE_CURRENT));
Query query = parser.Parse("handle:1");
// building the searchables
Searcher[] searchers = new Searcher[2];
// VITAL STEP:adding the searcher for the empty index first, before the searcher for the populated index
searchers[0] = new IndexSearcher(indexStoreB, true);
searchers[1] = new IndexSearcher(indexStoreA, true);
// creating the multiSearcher
Searcher mSearcher = GetMultiSearcherInstance(searchers);
// performing the search
ScoreDoc[] hits = mSearcher.Search(query, null, 1000).ScoreDocs;
Assert.AreEqual(3, hits.Length);
// iterating over the hit documents
for (int i = 0; i < hits.Length; i++)
{
mSearcher.Doc(hits[i].Doc);
}
mSearcher.Close();
//--------------------------------------------------------------------
// scenario 2
//--------------------------------------------------------------------
// adding one document to the empty index
writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(Util.Version.LUCENE_CURRENT), false, IndexWriter.MaxFieldLength.LIMITED);
writerB.AddDocument(lDoc);
writerB.Optimize();
writerB.Close();
// building the searchables
Searcher[] searchers2 = new Searcher[2];
// VITAL STEP:adding the searcher for the empty index first, before the searcher for the populated index
searchers2[0] = new IndexSearcher(indexStoreB, true);
searchers2[1] = new IndexSearcher(indexStoreA, true);
// creating the mulitSearcher
MultiSearcher mSearcher2 = GetMultiSearcherInstance(searchers2);
// performing the same search
ScoreDoc[] hits2 = mSearcher2.Search(query, null, 1000).ScoreDocs;
Assert.AreEqual(4, hits2.Length);
// iterating over the hit documents
for (int i = 0; i < hits2.Length; i++)
{
// no exception should happen at this point
mSearcher2.Doc(hits2[i].Doc);
}
// test the subSearcher() method:
Query subSearcherQuery = parser.Parse("id:doc1");
hits2 = mSearcher2.Search(subSearcherQuery, null, 1000).ScoreDocs;
Assert.AreEqual(2, hits2.Length);
Assert.AreEqual(0, mSearcher2.SubSearcher(hits2[0].Doc)); // hit from searchers2[0]
Assert.AreEqual(1, mSearcher2.SubSearcher(hits2[1].Doc)); // hit from searchers2[1]
subSearcherQuery = parser.Parse("id:doc2");
hits2 = mSearcher2.Search(subSearcherQuery, null, 1000).ScoreDocs;
Assert.AreEqual(1, hits2.Length);
Assert.AreEqual(1, mSearcher2.SubSearcher(hits2[0].Doc)); // hit from searchers2[1]
mSearcher2.Close();
//--------------------------------------------------------------------
// scenario 3
//--------------------------------------------------------------------
// deleting the document just added, this will cause a different exception to take place
Term term = new Term("id", "doc1");
IndexReader readerB = IndexReader.Open(indexStoreB, false);
readerB.DeleteDocuments(term);
readerB.Close();
// optimizing the index with the writer
writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(Util.Version.LUCENE_CURRENT), false, IndexWriter.MaxFieldLength.LIMITED);
writerB.Optimize();
writerB.Close();
// building the searchables
Searcher[] searchers3 = new Searcher[2];
searchers3[0] = new IndexSearcher(indexStoreB, true);
searchers3[1] = new IndexSearcher(indexStoreA, true);
// creating the mulitSearcher
Searcher mSearcher3 = GetMultiSearcherInstance(searchers3);
// performing the same search
ScoreDoc[] hits3 = mSearcher3.Search(query, null, 1000).ScoreDocs;
Assert.AreEqual(3, hits3.Length);
// iterating over the hit documents
for (int i = 0; i < hits3.Length; i++)
{
mSearcher3.Doc(hits3[i].Doc);
}
mSearcher3.Close();
indexStoreA.Close();
indexStoreB.Close();
}
private void TestNormalization(int nDocs, System.String message)
{
Query query = new TermQuery(new Term("contents", "doc0"));
RAMDirectory ramDirectory1;
IndexSearcher indexSearcher1;
ScoreDoc[] hits;
ramDirectory1 = new MockRAMDirectory();
// First put the documents in the same index
InitIndex(ramDirectory1, nDocs, true, null); // documents with a single token "doc0", "doc1", etc...
InitIndex(ramDirectory1, nDocs, false, "x"); // documents with two tokens "doc0" and "x", "doc1" and x, etc...
indexSearcher1 = new IndexSearcher(ramDirectory1, true);
indexSearcher1.SetDefaultFieldSortScoring(true, true);
hits = indexSearcher1.Search(query, null, 1000).ScoreDocs;
Assert.AreEqual(2, hits.Length, message);
// Store the scores for use later
float[] scores = new float[] { hits[0].Score, hits[1].Score };
Assert.IsTrue(scores[0] > scores[1], message);
indexSearcher1.Close();
ramDirectory1.Close();
hits = null;
RAMDirectory ramDirectory2;
IndexSearcher indexSearcher2;
ramDirectory1 = new MockRAMDirectory();
ramDirectory2 = new MockRAMDirectory();
// Now put the documents in a different index
InitIndex(ramDirectory1, nDocs, true, null); // documents with a single token "doc0", "doc1", etc...
InitIndex(ramDirectory2, nDocs, true, "x"); // documents with two tokens "doc0" and "x", "doc1" and x, etc...
indexSearcher1 = new IndexSearcher(ramDirectory1, true);
indexSearcher1.SetDefaultFieldSortScoring(true, true);
indexSearcher2 = new IndexSearcher(ramDirectory2, true);
indexSearcher2.SetDefaultFieldSortScoring(true, true);
Searcher searcher = GetMultiSearcherInstance(new Searcher[] { indexSearcher1, indexSearcher2 });
hits = searcher.Search(query, null, 1000).ScoreDocs;
Assert.AreEqual(2, hits.Length, message);
// The scores should be the same (within reason)
Assert.AreEqual(scores[0], hits[0].Score, 1e-6, message); // This will a document from ramDirectory1
Assert.AreEqual(scores[1], hits[1].Score, 1e-6, message); // This will a document from ramDirectory2
// Adding a Sort.RELEVANCE object should not change anything
hits = searcher.Search(query, null, 1000, Sort.RELEVANCE).ScoreDocs;
Assert.AreEqual(2, hits.Length, message);
Assert.AreEqual(scores[0], hits[0].Score, 1e-6, message); // This will a document from ramDirectory1
Assert.AreEqual(scores[1], hits[1].Score, 1e-6, message); // This will a document from ramDirectory2
searcher.Close();
ramDirectory1.Close();
ramDirectory2.Close();
}
public virtual void TestEmptyIndex()
{
// creating two directories for indices
Directory indexStoreA = new RAMDirectory();
Directory indexStoreB = new RAMDirectory();
// creating a document to store
Document lDoc = new Document();
lDoc.Add(Field.Text("fulltext", "Once upon a time....."));
lDoc.Add(Field.Keyword("id", "doc1"));
lDoc.Add(Field.Keyword("handle", "1"));
// creating a document to store
Document lDoc2 = new Document();
lDoc2.Add(Field.Text("fulltext", "in a galaxy far far away....."));
lDoc2.Add(Field.Keyword("id", "doc2"));
lDoc2.Add(Field.Keyword("handle", "1"));
// creating a document to store
Document lDoc3 = new Document();
lDoc3.Add(Field.Text("fulltext", "a bizarre bug manifested itself...."));
lDoc3.Add(Field.Keyword("id", "doc3"));
lDoc3.Add(Field.Keyword("handle", "1"));
// creating an index writer for the first index
IndexWriter writerA = new IndexWriter(indexStoreA, new StandardAnalyzer(), true);
// creating an index writer for the second index, but writing nothing
IndexWriter writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(), true);
//--------------------------------------------------------------------
// scenario 1
//--------------------------------------------------------------------
// writing the documents to the first index
writerA.AddDocument(lDoc);
writerA.AddDocument(lDoc2);
writerA.AddDocument(lDoc3);
writerA.Optimize();
writerA.Close();
// closing the second index
writerB.Close();
// creating the query
Query query = Lucene.Net.QueryParsers.QueryParser.Parse("handle:1", "fulltext", new StandardAnalyzer());
// building the searchables
Searcher[] searchers = new Searcher[2];
// VITAL STEP:adding the searcher for the empty index first, before the searcher for the populated index
searchers[0] = new IndexSearcher(indexStoreB);
searchers[1] = new IndexSearcher(indexStoreA);
// creating the multiSearcher
Searcher mSearcher = GetMultiSearcherInstance(searchers);
// performing the search
Hits hits = mSearcher.Search(query);
Assert.AreEqual(3, hits.Length());
try
{
// iterating over the hit documents
for (int i = 0; i < hits.Length(); i++)
{
Document d = hits.Doc(i);
}
}
catch (System.IndexOutOfRangeException e)
{
Assert.Fail("ArrayIndexOutOfBoundsException thrown: " + e.Message);
System.Console.Error.WriteLine(e.Source);
}
finally
{
mSearcher.Close();
}
//--------------------------------------------------------------------
// scenario 2
//--------------------------------------------------------------------
// adding one document to the empty index
writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(), false);
writerB.AddDocument(lDoc);
writerB.Optimize();
writerB.Close();
// building the searchables
Searcher[] searchers2 = new Searcher[2];
// VITAL STEP:adding the searcher for the empty index first, before the searcher for the populated index
searchers2[0] = new IndexSearcher(indexStoreB);
searchers2[1] = new IndexSearcher(indexStoreA);
// creating the mulitSearcher
Searcher mSearcher2 = GetMultiSearcherInstance(searchers2);
// performing the same search
Hits hits2 = mSearcher2.Search(query);
Assert.AreEqual(4, hits2.Length());
try
{
// iterating over the hit documents
for (int i = 0; i < hits2.Length(); i++)
{
// no exception should happen at this point
Document d = hits2.Doc(i);
}
}
catch (System.Exception e)
{
Assert.Fail("Exception thrown: " + e.Message);
System.Console.Error.WriteLine(e.Source);
}
finally
{
mSearcher2.Close();
}
//--------------------------------------------------------------------
// scenario 3
//--------------------------------------------------------------------
// deleting the document just added, this will cause a different exception to take place
Term term = new Term("id", "doc1");
IndexReader readerB = IndexReader.Open(indexStoreB);
readerB.Delete(term);
readerB.Close();
// optimizing the index with the writer
writerB = new IndexWriter(indexStoreB, new StandardAnalyzer(), false);
writerB.Optimize();
writerB.Close();
// building the searchables
Searcher[] searchers3 = new Searcher[2];
searchers3[0] = new IndexSearcher(indexStoreB);
searchers3[1] = new IndexSearcher(indexStoreA);
// creating the mulitSearcher
Searcher mSearcher3 = GetMultiSearcherInstance(searchers3);
// performing the same search
Hits hits3 = mSearcher3.Search(query);
Assert.AreEqual(3, hits3.Length());
try
{
// iterating over the hit documents
for (int i = 0; i < hits3.Length(); i++)
{
Document d = hits3.Doc(i);
}
}
catch (System.IO.IOException e)
{
Assert.Fail("IOException thrown: " + e.Message);
System.Console.Error.WriteLine(e.Source);
}
finally
{
mSearcher3.Close();
}
}