public virtual void TestFilterWorks()
{
Directory dir = new RAMDirectory();
IndexWriter writer = new IndexWriter(dir, new SimpleAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
for (int i = 0; i < 500; i++)
{
Document document = new Document();
document.Add(new Field("field", English.IntToEnglish(i) + " equals " + English.IntToEnglish(i), Field.Store.NO, Field.Index.ANALYZED));
writer.AddDocument(document);
}
writer.Close();
IndexReader reader = IndexReader.Open(dir, true);
SpanTermQuery query = new SpanTermQuery(new Term("field", English.IntToEnglish(10).Trim()));
SpanQueryFilter filter = new SpanQueryFilter(query);
SpanFilterResult result = filter.BitSpans(reader);
DocIdSet docIdSet = result.DocIdSet;
Assert.IsTrue(docIdSet != null, "docIdSet is null and it shouldn't be");
AssertContainsDocId("docIdSet doesn't contain docId 10", docIdSet, 10);
var spans = result.Positions;
Assert.IsTrue(spans != null, "spans is null and it shouldn't be");
int size = GetDocIdSetSize(docIdSet);
Assert.IsTrue(spans.Count == size, "spans Size: " + spans.Count + " is not: " + size);
for (System.Collections.IEnumerator iterator = spans.GetEnumerator(); iterator.MoveNext(); )
{
SpanFilterResult.PositionInfo info = (SpanFilterResult.PositionInfo) iterator.Current;
Assert.IsTrue(info != null, "info is null and it shouldn't be");
//The doc should indicate the bit is on
AssertContainsDocId("docIdSet doesn't contain docId " + info.Doc, docIdSet, info.Doc);
//There should be two positions in each
Assert.IsTrue(info.Positions.Count == 2, "info.getPositions() Size: " + info.Positions.Count + " is not: " + 2);
}
reader.Close();
}
public override Query Rewrite(IndexReader reader)
{
Query orig = new RegexQuery(term).Rewrite(reader);
// RegexQuery (via MultiTermQuery).rewrite always returns a BooleanQuery
BooleanQuery bq = (BooleanQuery) orig;
BooleanClause[] clauses = bq.GetClauses();
SpanQuery[] sqs = new SpanQuery[clauses.Length];
for (int i = 0; i < clauses.Length; i++)
{
BooleanClause clause = clauses[i];
// Clauses from RegexQuery.rewrite are always TermQuery's
TermQuery tq = (TermQuery) clause.GetQuery();
sqs[i] = new SpanTermQuery(tq.GetTerm());
sqs[i].SetBoost(tq.GetBoost());
}
SpanOrQuery query = new SpanOrQuery(sqs);
query.SetBoost(orig.GetBoost());
return query;
}
public override Query Rewrite(IndexReader reader)
{
Query orig = new RegexQuery(term).Rewrite(reader);
// RegexQuery (via MultiTermQuery).rewrite always returns a BooleanQuery
BooleanQuery bq = (BooleanQuery)orig;
BooleanClause[] clauses = bq.GetClauses();
SpanQuery[] sqs = new SpanQuery[clauses.Length];
for (int i = 0; i < clauses.Length; i++)
{
BooleanClause clause = clauses[i];
// Clauses from RegexQuery.rewrite are always TermQuery's
TermQuery tq = (TermQuery)clause.GetQuery();
sqs[i] = new SpanTermQuery(tq.GetTerm());
sqs[i].SetBoost(tq.GetBoost());
}
SpanOrQuery query = new SpanOrQuery(sqs);
query.SetBoost(orig.GetBoost());
return(query);
}
public virtual void TestQuery()
{
PayloadTermQuery boostingFuncTermQuery = new PayloadTermQuery(new Term(PayloadHelper.MULTI_FIELD, "seventy"), new MaxPayloadFunction());
QueryUtils.Check(boostingFuncTermQuery);
SpanTermQuery spanTermQuery = new SpanTermQuery(new Term(PayloadHelper.MULTI_FIELD, "seventy"));
Assert.IsTrue(boostingFuncTermQuery.Equals(spanTermQuery) == spanTermQuery.Equals(boostingFuncTermQuery));
PayloadTermQuery boostingFuncTermQuery2 = new PayloadTermQuery(new Term(PayloadHelper.MULTI_FIELD, "seventy"), new AveragePayloadFunction());
QueryUtils.CheckUnequal(boostingFuncTermQuery, boostingFuncTermQuery2);
}
public virtual void TestFilterWorks()
{
Directory dir = new RAMDirectory();
IndexWriter writer = new IndexWriter(dir, new SimpleAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED, null);
for (int i = 0; i < 500; i++)
{
Document document = new Document();
document.Add(new Field("field", English.IntToEnglish(i) + " equals " + English.IntToEnglish(i), Field.Store.NO, Field.Index.ANALYZED));
writer.AddDocument(document, null);
}
writer.Close();
IndexReader reader = IndexReader.Open(dir, true, null);
SpanTermQuery query = new SpanTermQuery(new Term("field", English.IntToEnglish(10).Trim()));
SpanQueryFilter filter = new SpanQueryFilter(query);
SpanFilterResult result = filter.BitSpans(reader, null);
DocIdSet docIdSet = result.DocIdSet;
Assert.IsTrue(docIdSet != null, "docIdSet is null and it shouldn't be");
AssertContainsDocId("docIdSet doesn't contain docId 10", docIdSet, 10);
var spans = result.Positions;
Assert.IsTrue(spans != null, "spans is null and it shouldn't be");
int size = GetDocIdSetSize(docIdSet);
Assert.IsTrue(spans.Count == size, "spans Size: " + spans.Count + " is not: " + size);
for (System.Collections.IEnumerator iterator = spans.GetEnumerator(); iterator.MoveNext();)
{
SpanFilterResult.PositionInfo info = (SpanFilterResult.PositionInfo)iterator.Current;
Assert.IsTrue(info != null, "info is null and it shouldn't be");
//The doc should indicate the bit is on
AssertContainsDocId("docIdSet doesn't contain docId " + info.Doc, docIdSet, info.Doc);
//There should be two positions in each
Assert.IsTrue(info.Positions.Count == 2, "info.getPositions() Size: " + info.Positions.Count + " is not: " + 2);
}
reader.Close();
}
public virtual void TestFilterWorks()
{
Directory dir = new RAMDirectory();
IndexWriter writer = new IndexWriter(dir, new SimpleAnalyzer(), true);
for (int i = 0; i < 500; i++)
{
Document document = new Document();
document.Add(new Field("field", English.IntToEnglish(i) + " equals " + English.IntToEnglish(i), Field.Store.NO, Field.Index.TOKENIZED));
writer.AddDocument(document);
}
writer.Close();
IndexReader reader = IndexReader.Open(dir);
SpanTermQuery query = new SpanTermQuery(new Term("field", English.IntToEnglish(10).Trim()));
SpanQueryFilter filter = new SpanQueryFilter(query);
SpanFilterResult result = filter.BitSpans(reader);
System.Collections.BitArray bits = result.GetBits();
Assert.IsTrue(bits != null, "bits is null and it shouldn't be");
Assert.IsTrue(bits.Get(10), "tenth bit is not on");
System.Collections.IList spans = result.GetPositions();
Assert.IsTrue(spans != null, "spans is null and it shouldn't be");
int cardinality = 0;
for (int i = 0; i < bits.Count; i++)
{
if (bits.Get(i)) cardinality++;
}
Assert.IsTrue(spans.Count == cardinality, "spans Size: " + spans.Count + " is not: " + cardinality);
for (System.Collections.IEnumerator iterator = spans.GetEnumerator(); iterator.MoveNext(); )
{
SpanFilterResult.PositionInfo info = (SpanFilterResult.PositionInfo) iterator.Current;
Assert.IsTrue(info != null, "info is null and it shouldn't be");
//The doc should indicate the bit is on
Assert.IsTrue(bits.Get(info.GetDoc()), "Bit is not on and it should be");
//There should be two positions in each
Assert.IsTrue(info.GetPositions().Count == 2, "info.getPositions() Size: " + info.GetPositions().Count + " is not: " + 2);
}
reader.Close();
}
public virtual void TestQuery()
{
PayloadTermQuery boostingFuncTermQuery = new PayloadTermQuery(new Term(PayloadHelper.MULTI_FIELD, "seventy"), new MaxPayloadFunction());
QueryUtils.Check(boostingFuncTermQuery);
SpanTermQuery spanTermQuery = new SpanTermQuery(new Term(PayloadHelper.MULTI_FIELD, "seventy"));
Assert.IsTrue(boostingFuncTermQuery.Equals(spanTermQuery) == spanTermQuery.Equals(boostingFuncTermQuery));
PayloadTermQuery boostingFuncTermQuery2 = new PayloadTermQuery(new Term(PayloadHelper.MULTI_FIELD, "seventy"), new AveragePayloadFunction());
QueryUtils.CheckUnequal(boostingFuncTermQuery, boostingFuncTermQuery2);
}
private void QueryToSpanQuery(Query query, ICollection <byte[]> payloads)
{
if (query is BooleanQuery)
{
BooleanClause[] queryClauses = ((BooleanQuery)query).GetClauses();
for (int i = 0; i < queryClauses.Length; i++)
{
if (!queryClauses[i].IsProhibited())
{
QueryToSpanQuery(queryClauses[i].GetQuery(), payloads);
}
}
}
else if (query is PhraseQuery)
{
Term[] phraseQueryTerms = ((PhraseQuery)query).GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = ((PhraseQuery)query).GetSlop();
bool inorder = false;
if (slop == 0)
{
inorder = true;
}
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.SetBoost(query.GetBoost());
GetPayloads(payloads, sp);
}
else if (query is TermQuery)
{
SpanTermQuery stq = new SpanTermQuery(((TermQuery)query).GetTerm());
stq.SetBoost(query.GetBoost());
GetPayloads(payloads, stq);
}
else if (query is SpanQuery)
{
GetPayloads(payloads, (SpanQuery)query);
}
else if (query is FilteredQuery)
{
QueryToSpanQuery(((FilteredQuery)query).GetQuery(), payloads);
}
else if (query is DisjunctionMaxQuery)
{
for (System.Collections.IEnumerator iterator = ((DisjunctionMaxQuery)query).Iterator(); iterator.MoveNext();)
{
QueryToSpanQuery((Query)iterator.Current, payloads);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery)query;
System.Collections.IList termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
System.Collections.ArrayList[] disjunctLists = new System.Collections.ArrayList[maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = (Term[])termArrays[i];
System.Collections.ArrayList disjuncts = disjunctLists[positions[i]];
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new System.Collections.ArrayList(termArray.Length));
++distinctPositions;
}
for (int j = 0; j < termArray.Length; ++j)
{
disjuncts.Add(new SpanTermQuery(termArray[j]));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
for (int i = 0; i < disjunctLists.Length; ++i)
{
System.Collections.ArrayList disjuncts = disjunctLists[i];
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery((SpanQuery[])(disjuncts.ToArray(typeof(SpanQuery[]))));
}
else
{
++positionGaps;
}
}
int slop = mpq.GetSlop();
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.SetBoost(query.GetBoost());
GetPayloads(payloads, sp);
}
}
}
public virtual void TestBooleanSpanQuery()
{
bool failed = false;
int hits = 0;
Directory directory = NewDirectory();
Analyzer indexerAnalyzer = new MockAnalyzer(Random());
IndexWriterConfig config = new IndexWriterConfig(TEST_VERSION_CURRENT, indexerAnalyzer);
IndexWriter writer = new IndexWriter(directory, config);
string FIELD = "content";
Document d = new Document();
d.Add(new TextField(FIELD, "clockwork orange", Field.Store.YES));
writer.AddDocument(d);
writer.Dispose();
IndexReader indexReader = DirectoryReader.Open(directory);
IndexSearcher searcher = NewSearcher(indexReader);
BooleanQuery query = new BooleanQuery();
SpanQuery sq1 = new SpanTermQuery(new Term(FIELD, "clockwork"));
SpanQuery sq2 = new SpanTermQuery(new Term(FIELD, "clckwork"));
query.Add(sq1, BooleanClause.Occur.SHOULD);
query.Add(sq2, BooleanClause.Occur.SHOULD);
TopScoreDocCollector collector = TopScoreDocCollector.Create(1000, true);
searcher.Search(query, collector);
hits = collector.TopDocs().ScoreDocs.Length;
foreach (ScoreDoc scoreDoc in collector.TopDocs().ScoreDocs)
{
Console.WriteLine(scoreDoc.Doc);
}
indexReader.Dispose();
Assert.AreEqual(failed, false, "Bug in boolean query composed of span queries");
Assert.AreEqual(hits, 1, "Bug in boolean query composed of span queries");
directory.Dispose();
}
public void TestNegativePositions()
{
SinkTokenizer tokens = new SinkTokenizer();
Token t = new Token();
t.SetTermText("a");
t.SetPositionIncrement(0);
tokens.Add(t);
t.SetTermText("b");
t.SetPositionIncrement(1);
tokens.Add(t);
t.SetTermText("c");
tokens.Add(t);
MockRAMDirectory dir = new MockRAMDirectory();
IndexWriter w = new IndexWriter(dir, false, new WhitespaceAnalyzer(), true);
Document doc = new Document();
doc.Add(new Field("field", tokens));
w.AddDocument(doc);
w.Close();
IndexSearcher s = new IndexSearcher(dir);
PhraseQuery pq = new PhraseQuery();
pq.Add(new Term("field", "a"));
pq.Add(new Term("field", "b"));
pq.Add(new Term("field", "c"));
Hits hits = s.Search(pq);
Assert.AreEqual(1, hits.Length());
Query q = new SpanTermQuery(new Term("field", "a"));
hits = s.Search(q);
Assert.AreEqual(1, hits.Length());
TermPositions tps = s.GetIndexReader().TermPositions(new Term("field", "a"));
Assert.IsTrue(tps.Next());
Assert.AreEqual(1, tps.Freq());
Assert.AreEqual(-1, tps.NextPosition());
Assert.IsTrue(_TestUtil.CheckIndex(dir));
s.Close();
dir.Close();
}
private void QueryToSpanQuery(Query query, ICollection<byte[]> payloads)
{
if (query is BooleanQuery)
{
BooleanClause[] queryClauses = ((BooleanQuery) query).GetClauses();
for (int i = 0; i < queryClauses.Length; i++)
{
if (!queryClauses[i].IsProhibited())
{
QueryToSpanQuery(queryClauses[i].GetQuery(), payloads);
}
}
}
else if (query is PhraseQuery)
{
Term[] phraseQueryTerms = ((PhraseQuery) query).GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = ((PhraseQuery) query).GetSlop();
bool inorder = false;
if (slop == 0)
{
inorder = true;
}
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.SetBoost(query.GetBoost());
GetPayloads(payloads, sp);
}
else if (query is TermQuery)
{
SpanTermQuery stq = new SpanTermQuery(((TermQuery) query).GetTerm());
stq.SetBoost(query.GetBoost());
GetPayloads(payloads, stq);
}
else if (query is SpanQuery)
{
GetPayloads(payloads, (SpanQuery) query);
}
else if (query is FilteredQuery)
{
QueryToSpanQuery(((FilteredQuery) query).GetQuery(), payloads);
}
else if (query is DisjunctionMaxQuery)
{
for (System.Collections.IEnumerator iterator = ((DisjunctionMaxQuery) query).Iterator(); iterator.MoveNext(); )
{
QueryToSpanQuery((Query) iterator.Current, payloads);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery) query;
System.Collections.IList termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
System.Collections.ArrayList[] disjunctLists = new System.Collections.ArrayList[maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = (Term[]) termArrays[i];
System.Collections.IList disjuncts = disjunctLists[positions[i]];
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new System.Collections.ArrayList(termArray.Length));
++distinctPositions;
}
for (int j = 0; j < termArray.Length; ++j)
{
disjuncts.Add(new SpanTermQuery(termArray[j]));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
for (int i = 0; i < disjunctLists.Length; ++i)
{
System.Collections.ArrayList disjuncts = disjunctLists[i];
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery((SpanQuery[]) (disjuncts.ToArray(typeof(SpanQuery[]))));
}
else
{
++positionGaps;
}
}
int slop = mpq.GetSlop();
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.SetBoost(query.GetBoost());
GetPayloads(payloads, sp);
}
}
}
public virtual void TestNegativePositions()
{
TokenStream tokens = new AnonymousClassTokenStream(this);
MockRAMDirectory dir = new MockRAMDirectory();
IndexWriter w = new IndexWriter(dir, new WhitespaceAnalyzer(), true, IndexWriter.MaxFieldLength.UNLIMITED);
Document doc = new Document();
doc.Add(new Field("field", tokens));
w.AddDocument(doc);
w.Commit();
IndexSearcher s = new IndexSearcher(dir);
PhraseQuery pq = new PhraseQuery();
pq.Add(new Term("field", "a"));
pq.Add(new Term("field", "b"));
pq.Add(new Term("field", "c"));
ScoreDoc[] hits = s.Search(pq, null, 1000).scoreDocs;
Assert.AreEqual(1, hits.Length);
Query q = new SpanTermQuery(new Term("field", "a"));
hits = s.Search(q, null, 1000).scoreDocs;
Assert.AreEqual(1, hits.Length);
TermPositions tps = s.GetIndexReader().TermPositions(new Term("field", "a"));
Assert.IsTrue(tps.Next());
Assert.AreEqual(1, tps.Freq());
Assert.AreEqual(0, tps.NextPosition());
w.Close();
Assert.IsTrue(_TestUtil.CheckIndex(dir));
s.Close();
dir.Close();
}
public override Query VisitSpanTermQuery(Lucene.Net.Search.Spans.SpanTermQuery spanTermq)
{
throw new SnNotSupportedException();
}
private void QueryToSpanQuery(Query query, ICollection<byte[]> payloads)
{
if (query is BooleanQuery)
{
BooleanClause[] queryClauses = ((BooleanQuery) query).GetClauses();
for (int i = 0; i < queryClauses.Length; i++)
{
if (!queryClauses[i].IsProhibited)
{
QueryToSpanQuery(queryClauses[i].Query, payloads);
}
}
}
else if (query is PhraseQuery)
{
Term[] phraseQueryTerms = ((PhraseQuery) query).GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = ((PhraseQuery) query).Slop;
bool inorder = false;
if (slop == 0)
{
inorder = true;
}
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.Boost = query.Boost;
GetPayloads(payloads, sp);
}
else if (query is TermQuery)
{
SpanTermQuery stq = new SpanTermQuery(((TermQuery) query).Term);
stq.Boost = query.Boost;
GetPayloads(payloads, stq);
}
else if (query is SpanQuery)
{
GetPayloads(payloads, (SpanQuery) query);
}
else if (query is FilteredQuery)
{
QueryToSpanQuery(((FilteredQuery) query).Query, payloads);
}
else if (query is DisjunctionMaxQuery)
{
for (IEnumerator<Query> iterator = ((DisjunctionMaxQuery)query).GetEnumerator(); iterator.MoveNext(); )
{
QueryToSpanQuery(iterator.Current, payloads);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery) query;
System.Collections.Generic.IList<Term[]> termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
IList<Query>[] disjunctLists = new IList<Query>[maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = termArrays[i];
IList<Query> disjuncts = disjunctLists[positions[i]];
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new List<Query>(termArray.Length));
++distinctPositions;
}
foreach(Term term in termArray)
{
disjuncts.Add(new SpanTermQuery(term));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
for (int i = 0; i < disjunctLists.Length; ++i)
{
IList<Query> disjuncts = disjunctLists[i];
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery((SpanQuery[]) (disjuncts.ToArray()));
}
else
{
++positionGaps;
}
}
int slop = mpq.Slop;
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.Boost = query.Boost;
GetPayloads(payloads, sp);
}
}
}
public virtual void TestPayloadsPos0()
{
for (int x = 0; x < 2; x++)
{
Directory dir = new MockRAMDirectory();
IndexWriter writer = new IndexWriter(dir, new TestPayloadAnalyzer(), true, IndexWriter.MaxFieldLength.LIMITED);
if (x == 1)
{
writer.SetAllowMinus1Position();
}
Document doc = new Document();
System.IO.MemoryStream ms = new System.IO.MemoryStream();
System.IO.StreamWriter sw = new System.IO.StreamWriter(ms);
sw.Write("a a b c d e a f g h i j a b k k");
// flush to stream & reset it's position so it can be read
sw.Flush();
ms.Position = 0;
doc.Add(new Field("content", new System.IO.StreamReader(ms)));
writer.AddDocument(doc);
IndexReader r = writer.GetReader();
TermPositions tp = r.TermPositions(new Term("content", "a"));
int count = 0;
Assert.IsTrue(tp.Next());
// "a" occurs 4 times
Assert.AreEqual(4, tp.Freq());
int expected;
if (x == 1)
{
expected = System.Int32.MaxValue;
}
else
{
expected = 0;
}
Assert.AreEqual(expected, tp.NextPosition());
if (x == 1)
{
continue;
}
Assert.AreEqual(1, tp.NextPosition());
Assert.AreEqual(3, tp.NextPosition());
Assert.AreEqual(6, tp.NextPosition());
// only one doc has "a"
Assert.IsFalse(tp.Next());
IndexSearcher is_Renamed = new IndexSearcher(r);
SpanTermQuery stq1 = new SpanTermQuery(new Term("content", "a"));
SpanTermQuery stq2 = new SpanTermQuery(new Term("content", "k"));
SpanQuery[] sqs = new SpanQuery[]{stq1, stq2};
SpanNearQuery snq = new SpanNearQuery(sqs, 30, false);
count = 0;
bool sawZero = false;
//System.out.println("\ngetPayloadSpans test");
Lucene.Net.Search.Spans.Spans pspans = snq.GetSpans(is_Renamed.GetIndexReader());
while (pspans.Next())
{
//System.out.println(pspans.doc() + " - " + pspans.start() + " - "+ pspans.end());
System.Collections.Generic.ICollection<byte[]> payloads = pspans.GetPayload();
sawZero |= pspans.Start() == 0;
for (System.Collections.IEnumerator it = payloads.GetEnumerator(); it.MoveNext(); )
{
count++;
System.Object generatedAux2 = it.Current;
//System.out.println(new String((byte[]) it.next()));
}
}
Assert.AreEqual(5, count);
Assert.IsTrue(sawZero);
//System.out.println("\ngetSpans test");
Lucene.Net.Search.Spans.Spans spans = snq.GetSpans(is_Renamed.GetIndexReader());
count = 0;
sawZero = false;
while (spans.Next())
{
count++;
sawZero |= spans.Start() == 0;
//System.out.println(spans.doc() + " - " + spans.start() + " - " + spans.end());
}
Assert.AreEqual(4, count);
Assert.IsTrue(sawZero);
//System.out.println("\nPayloadSpanUtil test");
sawZero = false;
PayloadSpanUtil psu = new PayloadSpanUtil(is_Renamed.GetIndexReader());
System.Collections.Generic.ICollection<byte[]> pls = psu.GetPayloadsForQuery(snq);
count = pls.Count;
for (System.Collections.IEnumerator it = pls.GetEnumerator(); it.MoveNext(); )
{
System.String s = new System.String(System.Text.UTF8Encoding.UTF8.GetChars((byte[]) it.Current));
//System.out.println(s);
sawZero |= s.Equals("pos: 0");
}
Assert.AreEqual(5, count);
Assert.IsTrue(sawZero);
writer.Close();
is_Renamed.GetIndexReader().Close();
dir.Close();
}
}
public virtual void TestPayloadsPos0()
{
Directory dir = new MockRAMDirectory();
IndexWriter writer = new IndexWriter(dir, new TestPayloadAnalyzer(), true,
IndexWriter.MaxFieldLength.LIMITED);
Document doc = new Document();
System.IO.MemoryStream ms = new System.IO.MemoryStream();
System.IO.StreamWriter sw = new System.IO.StreamWriter(ms);
sw.Write("a a b c d e a f g h i j a b k k");
// flush to stream & reset it's position so it can be read
sw.Flush();
ms.Position = 0;
doc.Add(new Field("content", new System.IO.StreamReader(ms)));
writer.AddDocument(doc);
IndexReader r = writer.GetReader();
TermPositions tp = r.TermPositions(new Term("content", "a"));
int count = 0;
Assert.IsTrue(tp.Next());
// "a" occurs 4 times
Assert.AreEqual(4, tp.Freq);
int expected = 0;
Assert.AreEqual(expected, tp.NextPosition());
Assert.AreEqual(1, tp.NextPosition());
Assert.AreEqual(3, tp.NextPosition());
Assert.AreEqual(6, tp.NextPosition());
// only one doc has "a"
Assert.IsFalse(tp.Next());
IndexSearcher is_Renamed = new IndexSearcher(r);
SpanTermQuery stq1 = new SpanTermQuery(new Term("content", "a"));
SpanTermQuery stq2 = new SpanTermQuery(new Term("content", "k"));
SpanQuery[] sqs = new SpanQuery[] { stq1, stq2 };
SpanNearQuery snq = new SpanNearQuery(sqs, 30, false);
count = 0;
bool sawZero = false;
//System.out.println("\ngetPayloadSpans test");
Lucene.Net.Search.Spans.Spans pspans = snq.GetSpans(is_Renamed.IndexReader);
while (pspans.Next())
{
//System.out.println(pspans.doc() + " - " + pspans.start() + " - "+ pspans.end());
System.Collections.Generic.ICollection <byte[]> payloads = pspans.GetPayload();
sawZero |= pspans.Start() == 0;
for (System.Collections.IEnumerator it = payloads.GetEnumerator(); it.MoveNext();)
{
count++;
System.Object generatedAux2 = it.Current;
//System.out.println(new String((byte[]) it.next()));
}
}
Assert.AreEqual(5, count);
Assert.IsTrue(sawZero);
//System.out.println("\ngetSpans test");
Lucene.Net.Search.Spans.Spans spans = snq.GetSpans(is_Renamed.IndexReader);
count = 0;
sawZero = false;
while (spans.Next())
{
count++;
sawZero |= spans.Start() == 0;
//System.out.println(spans.doc() + " - " + spans.start() + " - " + spans.end());
}
Assert.AreEqual(4, count);
Assert.IsTrue(sawZero);
//System.out.println("\nPayloadSpanUtil test");
sawZero = false;
PayloadSpanUtil psu = new PayloadSpanUtil(is_Renamed.IndexReader);
System.Collections.Generic.ICollection <byte[]> pls = psu.GetPayloadsForQuery(snq);
count = pls.Count;
for (System.Collections.IEnumerator it = pls.GetEnumerator(); it.MoveNext();)
{
System.String s = new System.String(System.Text.UTF8Encoding.UTF8.GetChars((byte[])it.Current));
//System.out.println(s);
sawZero |= s.Equals("pos: 0");
}
Assert.AreEqual(5, count);
Assert.IsTrue(sawZero);
writer.Close();
is_Renamed.IndexReader.Close();
dir.Close();
}
private void QueryToSpanQuery(Query query, ICollection <byte[]> payloads, IState state)
{
if (query is BooleanQuery)
{
BooleanClause[] queryClauses = ((BooleanQuery)query).GetClauses();
for (int i = 0; i < queryClauses.Length; i++)
{
if (!queryClauses[i].IsProhibited)
{
QueryToSpanQuery(queryClauses[i].Query, payloads, state);
}
}
}
else if (query is PhraseQuery)
{
Term[] phraseQueryTerms = ((PhraseQuery)query).GetTerms();
SpanQuery[] clauses = new SpanQuery[phraseQueryTerms.Length];
for (int i = 0; i < phraseQueryTerms.Length; i++)
{
clauses[i] = new SpanTermQuery(phraseQueryTerms[i]);
}
int slop = ((PhraseQuery)query).Slop;
bool inorder = false;
if (slop == 0)
{
inorder = true;
}
SpanNearQuery sp = new SpanNearQuery(clauses, slop, inorder);
sp.Boost = query.Boost;
GetPayloads(payloads, sp, state);
}
else if (query is TermQuery)
{
SpanTermQuery stq = new SpanTermQuery(((TermQuery)query).Term);
stq.Boost = query.Boost;
GetPayloads(payloads, stq, state);
}
else if (query is SpanQuery)
{
GetPayloads(payloads, (SpanQuery)query, state);
}
else if (query is FilteredQuery)
{
QueryToSpanQuery(((FilteredQuery)query).Query, payloads, state);
}
else if (query is DisjunctionMaxQuery)
{
for (IEnumerator <Query> iterator = ((DisjunctionMaxQuery)query).GetEnumerator(); iterator.MoveNext();)
{
QueryToSpanQuery(iterator.Current, payloads, state);
}
}
else if (query is MultiPhraseQuery)
{
MultiPhraseQuery mpq = (MultiPhraseQuery)query;
System.Collections.Generic.IList <Term[]> termArrays = mpq.GetTermArrays();
int[] positions = mpq.GetPositions();
if (positions.Length > 0)
{
int maxPosition = positions[positions.Length - 1];
for (int i = 0; i < positions.Length - 1; ++i)
{
if (positions[i] > maxPosition)
{
maxPosition = positions[i];
}
}
IList <Query>[] disjunctLists = new IList <Query> [maxPosition + 1];
int distinctPositions = 0;
for (int i = 0; i < termArrays.Count; ++i)
{
Term[] termArray = termArrays[i];
IList <Query> disjuncts = disjunctLists[positions[i]];
if (disjuncts == null)
{
disjuncts = (disjunctLists[positions[i]] = new List <Query>(termArray.Length));
++distinctPositions;
}
foreach (Term term in termArray)
{
disjuncts.Add(new SpanTermQuery(term));
}
}
int positionGaps = 0;
int position = 0;
SpanQuery[] clauses = new SpanQuery[distinctPositions];
for (int i = 0; i < disjunctLists.Length; ++i)
{
IList <Query> disjuncts = disjunctLists[i];
if (disjuncts != null)
{
clauses[position++] = new SpanOrQuery((SpanQuery[])(disjuncts.ToArray()));
}
else
{
++positionGaps;
}
}
int slop = mpq.Slop;
bool inorder = (slop == 0);
SpanNearQuery sp = new SpanNearQuery(clauses, slop + positionGaps, inorder);
sp.Boost = query.Boost;
GetPayloads(payloads, sp, state);
}
}
}
private PayloadNearQuery NewPhraseQuery(string fieldName, string phrase, bool inOrder, PayloadFunction function)
{
var words = _whiteSpaceRegex.Split(phrase);
var clauses = new SpanQuery[words.Length];
for (var i = 0; i < clauses.Length; i++)
{
clauses[i] = new SpanTermQuery(new Term(fieldName, words[i]));
}
return new PayloadNearQuery(clauses, 0, inOrder, function);
}
public virtual void TestCrazySpans()
{
// The problem: "normal" lucene queries create scorers, returning null if terms dont exist
// this means they never score a term that does not exist.
// however with spans, there is only one scorer for the whole hierarchy:
// inner queries are not real queries, their boosts are ignored, etc.
Directory dir = NewDirectory();
RandomIndexWriter iw = new RandomIndexWriter(Random(), dir);
Document doc = new Document();
FieldType ft = new FieldType(TextField.TYPE_NOT_STORED);
doc.Add(NewField("foo", "bar", ft));
iw.AddDocument(doc);
IndexReader ir = iw.Reader;
iw.Dispose();
IndexSearcher @is = NewSearcher(ir);
foreach (Similarity sim in Sims)
{
@is.Similarity = sim;
SpanTermQuery s1 = new SpanTermQuery(new Term("foo", "bar"));
SpanTermQuery s2 = new SpanTermQuery(new Term("foo", "baz"));
Query query = new SpanOrQuery(s1, s2);
TopDocs td = @is.Search(query, 10);
Assert.AreEqual(1, td.TotalHits);
float score = td.ScoreDocs[0].Score;
Assert.IsTrue(score >= 0.0f);
Assert.IsFalse(float.IsInfinity(score), "inf score for " + sim);
}
ir.Dispose();
dir.Dispose();
}
public virtual void TestPayloadsPos0()
{
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(Random(), dir, new MockPayloadAnalyzer());
Document doc = new Document();
doc.Add(new TextField("content", new StringReader("a a b c d e a f g h i j a b k k")));
writer.AddDocument(doc);
IndexReader readerFromWriter = writer.Reader;
AtomicReader r = SlowCompositeReaderWrapper.Wrap(readerFromWriter);
DocsAndPositionsEnum tp = r.TermPositionsEnum(new Term("content", "a"));
int count = 0;
Assert.IsTrue(tp.NextDoc() != DocIdSetIterator.NO_MORE_DOCS);
// "a" occurs 4 times
Assert.AreEqual(4, tp.Freq());
Assert.AreEqual(0, tp.NextPosition());
Assert.AreEqual(1, tp.NextPosition());
Assert.AreEqual(3, tp.NextPosition());
Assert.AreEqual(6, tp.NextPosition());
// only one doc has "a"
Assert.AreEqual(DocIdSetIterator.NO_MORE_DOCS, tp.NextDoc());
IndexSearcher @is = NewSearcher(readerFromWriter);
SpanTermQuery stq1 = new SpanTermQuery(new Term("content", "a"));
SpanTermQuery stq2 = new SpanTermQuery(new Term("content", "k"));
SpanQuery[] sqs = new SpanQuery[] { stq1, stq2 };
SpanNearQuery snq = new SpanNearQuery(sqs, 30, false);
count = 0;
bool sawZero = false;
if (VERBOSE)
{
Console.WriteLine("\ngetPayloadSpans test");
}
Search.Spans.Spans pspans = MultiSpansWrapper.Wrap(@is.TopReaderContext, snq);
while (pspans.Next())
{
if (VERBOSE)
{
Console.WriteLine("doc " + pspans.Doc() + ": span " + pspans.Start() + " to " + pspans.End());
}
var payloads = pspans.Payload;
sawZero |= pspans.Start() == 0;
foreach (var bytes in payloads)
{
count++;
if (VERBOSE)
{
Console.WriteLine(" payload: " + Encoding.UTF8.GetString((byte[])(Array)bytes));
}
}
}
Assert.IsTrue(sawZero);
Assert.AreEqual(5, count);
// System.out.println("\ngetSpans test");
Search.Spans.Spans spans = MultiSpansWrapper.Wrap(@is.TopReaderContext, snq);
count = 0;
sawZero = false;
while (spans.Next())
{
count++;
sawZero |= spans.Start() == 0;
// System.out.println(spans.Doc() + " - " + spans.Start() + " - " +
// spans.End());
}
Assert.AreEqual(4, count);
Assert.IsTrue(sawZero);
// System.out.println("\nPayloadSpanUtil test");
sawZero = false;
PayloadSpanUtil psu = new PayloadSpanUtil(@is.TopReaderContext);
var pls = psu.GetPayloadsForQuery(snq);
count = pls.Count;
foreach (var bytes in pls)
{
string s = Encoding.UTF8.GetString(bytes);
//System.out.println(s);
sawZero |= s.Equals("pos: 0");
}
Assert.AreEqual(5, count);
Assert.IsTrue(sawZero);
writer.Dispose();
@is.IndexReader.Dispose();
dir.Dispose();
}
public override Query VisitSpanTermQuery(Lucene.Net.Search.Spans.SpanTermQuery spanTermq)
{
throw new NotImplementedException();
}
