public void TestBoostOnlyRewrite()
{
Directory directory = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), directory, Similarity, TimeZone);
addDoc("Lucene", writer);
addDoc("Lucene", writer);
addDoc("Lucenne", writer);
IndexReader reader = writer.Reader;
IndexSearcher searcher = NewSearcher(reader);
writer.Dispose();
SlowFuzzyQuery query = new SlowFuzzyQuery(new Term("field", "lucene"));
query.SetRewriteMethod(new MultiTermQuery.TopTermsBoostOnlyBooleanQueryRewrite(50));
ScoreDoc[] hits = searcher.Search(query, null, 1000).ScoreDocs;
assertEquals(3, hits.Length);
// normally, 'Lucenne' would be the first result as IDF will skew the score.
assertEquals("Lucene", reader.Document(hits[0].Doc).Get("field"));
assertEquals("Lucene", reader.Document(hits[1].Doc).Get("field"));
assertEquals("Lucenne", reader.Document(hits[2].Doc).Get("field"));
reader.Dispose();
directory.Dispose();
}
public void assertFromTestData(int[] codePointTable)
{
if (VERBOSE)
{
Console.WriteLine("TEST: codePointTable=" + codePointTable);
}
//Stream stream = GetType().getResourceAsStream("fuzzyTestData.txt");
Stream stream = GetType().Assembly.GetManifestResourceStream("Lucene.Net.Sandbox.Queries.fuzzyTestData.txt");
TextReader reader = new StreamReader(stream, Encoding.UTF8);
int bits = int.Parse(reader.ReadLine(), CultureInfo.InvariantCulture);
int terms = (int)Math.Pow(2, bits);
Store.Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random(), MockTokenizer.KEYWORD, false)).SetMergePolicy(NewLogMergePolicy()));
Document doc = new Document();
Field field = NewTextField("field", "", Field.Store.NO);
doc.Add(field);
for (int i = 0; i < terms; i++)
{
field.StringValue = (MapInt(codePointTable, i));
writer.AddDocument(doc);
}
IndexReader r = writer.Reader;
IndexSearcher searcher = NewSearcher(r);
if (VERBOSE)
{
Console.WriteLine("TEST: searcher=" + searcher);
}
// even though this uses a boost-only rewrite, this test relies upon queryNorm being the default implementation,
// otherwise scores are different!
searcher.Similarity = (new DefaultSimilarity());
writer.Dispose();
String line;
int lineNum = 0;
while ((line = reader.ReadLine()) != null)
{
lineNum++;
String[] @params = line.Split(new char[] { ',' }, StringSplitOptions.RemoveEmptyEntries);
String query = MapInt(codePointTable, int.Parse(@params[0], CultureInfo.InvariantCulture));
int prefix = int.Parse(@params[1], CultureInfo.InvariantCulture);
int pqSize = int.Parse(@params[2], CultureInfo.InvariantCulture);
float minScore = float.Parse(@params[3], CultureInfo.InvariantCulture);
#pragma warning disable 612, 618
SlowFuzzyQuery q = new SlowFuzzyQuery(new Term("field", query), minScore, prefix);
#pragma warning restore 612, 618
q.SetRewriteMethod(new MultiTermQuery.TopTermsBoostOnlyBooleanQueryRewrite(pqSize));
int expectedResults = int.Parse(reader.ReadLine(), CultureInfo.InvariantCulture);
TopDocs docs = searcher.Search(q, expectedResults);
assertEquals(expectedResults, docs.TotalHits);
for (int i = 0; i < expectedResults; i++)
{
String[] scoreDoc = reader.ReadLine().Split(new char[] { ',' }, StringSplitOptions.RemoveEmptyEntries);
assertEquals(int.Parse(scoreDoc[0], CultureInfo.InvariantCulture), docs.ScoreDocs[i].Doc);
assertEquals(float.Parse(scoreDoc[1], CultureInfo.InvariantCulture), docs.ScoreDocs[i].Score, epsilon);
}
}
r.Dispose();
dir.Dispose();
}
public void assertFromTestData(int[] codePointTable)
{
if (VERBOSE)
{
Console.WriteLine("TEST: codePointTable=" + codePointTable);
}
//Stream stream = GetType().getResourceAsStream("fuzzyTestData.txt");
Stream stream = GetType().Assembly.GetManifestResourceStream("Lucene.Net.Sandbox.Queries.fuzzyTestData.txt");
TextReader reader = new StreamReader(stream, Encoding.UTF8);
int bits = int.Parse(reader.ReadLine(), CultureInfo.InvariantCulture);
int terms = (int)Math.Pow(2, bits);
Store.Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random(), MockTokenizer.KEYWORD, false)).SetMergePolicy(NewLogMergePolicy()));
Document doc = new Document();
Field field = NewTextField("field", "", Field.Store.NO);
doc.Add(field);
for (int i = 0; i < terms; i++)
{
field.StringValue = (MapInt(codePointTable, i));
writer.AddDocument(doc);
}
IndexReader r = writer.Reader;
IndexSearcher searcher = NewSearcher(r);
if (VERBOSE)
{
Console.WriteLine("TEST: searcher=" + searcher);
}
// even though this uses a boost-only rewrite, this test relies upon queryNorm being the default implementation,
// otherwise scores are different!
searcher.Similarity = (new DefaultSimilarity());
writer.Dispose();
String line;
int lineNum = 0;
while ((line = reader.ReadLine()) != null)
{
lineNum++;
String[] @params = line.Split(new char[] { ',' }, StringSplitOptions.RemoveEmptyEntries);
String query = MapInt(codePointTable, int.Parse(@params[0], CultureInfo.InvariantCulture));
int prefix = int.Parse(@params[1], CultureInfo.InvariantCulture);
int pqSize = int.Parse(@params[2], CultureInfo.InvariantCulture);
float minScore = float.Parse(@params[3], CultureInfo.InvariantCulture);
#pragma warning disable 612, 618
SlowFuzzyQuery q = new SlowFuzzyQuery(new Term("field", query), minScore, prefix);
#pragma warning restore 612, 618
q.SetRewriteMethod(new MultiTermQuery.TopTermsBoostOnlyBooleanQueryRewrite(pqSize));
int expectedResults = int.Parse(reader.ReadLine(), CultureInfo.InvariantCulture);
TopDocs docs = searcher.Search(q, expectedResults);
assertEquals(expectedResults, docs.TotalHits);
for (int i = 0; i < expectedResults; i++)
{
String[] scoreDoc = reader.ReadLine().Split(new char[] { ',' }, StringSplitOptions.RemoveEmptyEntries);
assertEquals(int.Parse(scoreDoc[0], CultureInfo.InvariantCulture), docs.ScoreDocs[i].Doc);
assertEquals(float.Parse(scoreDoc[1], CultureInfo.InvariantCulture), docs.ScoreDocs[i].Score, epsilon);
}
}
r.Dispose();
dir.Dispose();
}
public void TestBoostOnlyRewrite()
{
Directory directory = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), directory, Similarity, TimeZone);
addDoc("Lucene", writer);
addDoc("Lucene", writer);
addDoc("Lucenne", writer);
IndexReader reader = writer.Reader;
IndexSearcher searcher = NewSearcher(reader);
writer.Dispose();
SlowFuzzyQuery query = new SlowFuzzyQuery(new Term("field", "lucene"));
query.SetRewriteMethod(new MultiTermQuery.TopTermsBoostOnlyBooleanQueryRewrite(50));
ScoreDoc[] hits = searcher.Search(query, null, 1000).ScoreDocs;
assertEquals(3, hits.Length);
// normally, 'Lucenne' would be the first result as IDF will skew the score.
assertEquals("Lucene", reader.Document(hits[0].Doc).Get("field"));
assertEquals("Lucene", reader.Document(hits[1].Doc).Get("field"));
assertEquals("Lucenne", reader.Document(hits[2].Doc).Get("field"));
reader.Dispose();
directory.Dispose();
}