public override DocIdSet GetDocIdSet(AtomicReaderContext context, IBits acceptDocs)
{
SortedDocValues fcsi = FieldCache.GetTermsIndex((context.AtomicReader), field);
FixedBitSet bits = new FixedBitSet(fcsi.ValueCount);
for (int i = 0; i < terms.Length; i++)
{
int ord = fcsi.LookupTerm(terms[i]);
if (ord >= 0)
{
bits.Set(ord);
}
}
return(new FieldCacheDocIdSet(context.Reader.MaxDoc, acceptDocs, (doc) =>
{
int ord = fcsi.GetOrd(doc);
if (ord == -1)
{
// missing
return false;
}
else
{
return bits.Get(ord);
}
}));
}
public FieldCacheDocIdSetAnonymousInnerClassHelper(FieldCacheTermsFilter outerInstance, int maxDoc, IBits acceptDocs, SortedDocValues fcsi, FixedBitSet bits)
: base(maxDoc, acceptDocs)
{
this.outerInstance = outerInstance;
this.fcsi = fcsi;
this.bits = bits;
}
public NumericDocValuesFieldUpdates(string field, int maxDoc)
: base(field, Type_e.NUMERIC)
{
DocsWithField = new FixedBitSet(64);
Docs = new PagedMutable(1, 1024, PackedInts.BitsRequired(maxDoc - 1), PackedInts.COMPACT);
Values = new PagedGrowableWriter(1, 1024, 1, PackedInts.FAST);
Size = 0;
}
public NumericDocValuesWriter(FieldInfo fieldInfo, Counter iwBytesUsed, bool trackDocsWithField)
{
Pending = new AppendingDeltaPackedLongBuffer(PackedInts.COMPACT);
DocsWithField = trackDocsWithField ? new FixedBitSet(64) : null;
BytesUsed = Pending.RamBytesUsed() + DocsWithFieldBytesUsed();
this.FieldInfo = fieldInfo;
this.IwBytesUsed = iwBytesUsed;
iwBytesUsed.AddAndGet(BytesUsed);
}
public override DocIdSet GetDocIdSet(AtomicReaderContext context, Bits acceptDocs)
{
FixedBitSet bits = new FixedBitSet(context.Reader.MaxDoc);
bits.Set(Doc);
if (acceptDocs != null && !acceptDocs.Get(Doc))
{
bits.Clear(Doc);
}
return bits;
}
public BinaryDocValuesFieldUpdates(string field, int maxDoc)
: base(field, Type_e.BINARY)
{
DocsWithField = new FixedBitSet(64);
Docs = new PagedMutable(1, 1024, PackedInts.BitsRequired(maxDoc - 1), PackedInts.COMPACT);
Offsets = new PagedGrowableWriter(1, 1024, 1, PackedInts.FAST);
Lengths = new PagedGrowableWriter(1, 1024, 1, PackedInts.FAST);
Values = new BytesRef(16); // start small
Size = 0;
}
public override DocIdSet GetDocIdSet(AtomicReaderContext context, Bits acceptDocs)
{
FixedBitSet bits = new FixedBitSet(context.Reader.MaxDoc);
bits.Set(Doc);
if (acceptDocs != null && !acceptDocs.Get(Doc))
{
bits.Clear(Doc);
}
return(bits);
}
public override DocIdSet GetDocIdSet(AtomicReaderContext context, IBits acceptDocs)
{
SortedDocValues fcsi = FieldCache.GetTermsIndex((context.AtomicReader), field);
FixedBitSet bits = new FixedBitSet(fcsi.ValueCount);
for (int i = 0; i < terms.Length; i++)
{
int ord = fcsi.LookupTerm(terms[i]);
if (ord >= 0)
{
bits.Set(ord);
}
}
return(new FieldCacheDocIdSetAnonymousInnerClassHelper(this, context.Reader.MaxDoc, acceptDocs, fcsi, bits));
}
public override DocIdSet GetDocIdSet(AtomicReaderContext context, Bits acceptDocs)
{
SortedDocValues fcsi = FieldCache.GetTermsIndex((context.AtomicReader), Field);
FixedBitSet bits = new FixedBitSet(fcsi.ValueCount);
for (int i = 0; i < Terms.Length; i++)
{
int ord = fcsi.LookupTerm(Terms[i]);
if (ord >= 0)
{
bits.Set(ord);
}
}
return new FieldCacheDocIdSetAnonymousInnerClassHelper(this, context.Reader.MaxDoc, acceptDocs, fcsi, bits);
}
public InPlaceMergeSorterAnonymousInnerClassHelper(NumericDocValuesFieldUpdates outerInstance, PagedMutable docs, PagedGrowableWriter values, FixedBitSet docsWithField)
{
this.OuterInstance = outerInstance;
this.Docs = docs;
this.Values = values;
this.DocsWithField = docsWithField;
}
internal Iterator(int size, PagedGrowableWriter values, FixedBitSet docsWithField, PagedMutable docs)
{
this.Size = size;
this.Values = values;
this.DocsWithField = docsWithField;
this.Docs = docs;
}
public override void Merge(DocValuesFieldUpdates other)
{
Debug.Assert(other is NumericDocValuesFieldUpdates);
NumericDocValuesFieldUpdates otherUpdates = (NumericDocValuesFieldUpdates)other;
if (Size + otherUpdates.Size > int.MaxValue)
{
throw new InvalidOperationException("cannot support more than Integer.MAX_VALUE doc/value entries; size=" + Size + " other.size=" + otherUpdates.Size);
}
Docs = Docs.Grow(Size + otherUpdates.Size);
Values = Values.Grow(Size + otherUpdates.Size);
DocsWithField = FixedBitSet.EnsureCapacity(DocsWithField, (int)Docs.Size());
for (int i = 0; i < otherUpdates.Size; i++)
{
int doc = (int)otherUpdates.Docs.Get(i);
if (otherUpdates.DocsWithField.Get(i))
{
DocsWithField.Set(Size);
}
Docs.Set(Size, doc);
Values.Set(Size, otherUpdates.Values.Get(i));
++Size;
}
}
public static void AfterClass()
{
AllTerms = null;
FieldInfos = null;
Fields = null;
GlobalLiveDocs = null;
}
/// <summary>
/// Test term vectors.
/// @lucene.experimental
/// </summary>
public static Status.TermVectorStatus TestTermVectors(AtomicReader reader, TextWriter infoStream, bool verbose, bool crossCheckTermVectors)
{
Status.TermVectorStatus status = new Status.TermVectorStatus();
FieldInfos fieldInfos = reader.FieldInfos;
Bits onlyDocIsDeleted = new FixedBitSet(1);
try
{
if (infoStream != null)
{
infoStream.Write(" test: term vectors........");
}
DocsEnum docs = null;
DocsAndPositionsEnum postings = null;
// Only used if crossCheckTermVectors is true:
DocsEnum postingsDocs = null;
DocsAndPositionsEnum postingsPostings = null;
Bits liveDocs = reader.LiveDocs;
Fields postingsFields;
// TODO: testTermsIndex
if (crossCheckTermVectors)
{
postingsFields = reader.Fields;
}
else
{
postingsFields = null;
}
TermsEnum termsEnum = null;
TermsEnum postingsTermsEnum = null;
for (int j = 0; j < reader.MaxDoc; ++j)
{
// Intentionally pull/visit (but don't count in
// stats) deleted documents to make sure they too
// are not corrupt:
Fields tfv = reader.GetTermVectors(j);
// TODO: can we make a IS(FIR) that searches just
// this term vector... to pass for searcher?
if (tfv != null)
{
// First run with no deletions:
CheckFields(tfv, null, 1, fieldInfos, false, true, infoStream, verbose);
// Again, with the one doc deleted:
CheckFields(tfv, onlyDocIsDeleted, 1, fieldInfos, false, true, infoStream, verbose);
// Only agg stats if the doc is live:
bool doStats = liveDocs == null || liveDocs.Get(j);
if (doStats)
{
status.DocCount++;
}
foreach (string field in tfv)
{
if (doStats)
{
status.TotVectors++;
}
// Make sure FieldInfo thinks this field is vector'd:
FieldInfo fieldInfo = fieldInfos.FieldInfo(field);
if (!fieldInfo.HasVectors())
{
throw new Exception("docID=" + j + " has term vectors for field=" + field + " but FieldInfo has storeTermVector=false");
}
if (crossCheckTermVectors)
{
Terms terms = tfv.Terms(field);
termsEnum = terms.Iterator(termsEnum);
bool postingsHasFreq = fieldInfo.FieldIndexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS;
bool postingsHasPayload = fieldInfo.HasPayloads();
bool vectorsHasPayload = terms.HasPayloads();
Terms postingsTerms = postingsFields.Terms(field);
if (postingsTerms == null)
{
throw new Exception("vector field=" + field + " does not exist in postings; doc=" + j);
}
postingsTermsEnum = postingsTerms.Iterator(postingsTermsEnum);
bool hasProx = terms.HasOffsets() || terms.HasPositions();
BytesRef term = null;
while ((term = termsEnum.Next()) != null)
{
if (hasProx)
{
postings = termsEnum.DocsAndPositions(null, postings);
Debug.Assert(postings != null);
docs = null;
}
else
{
docs = termsEnum.Docs(null, docs);
Debug.Assert(docs != null);
postings = null;
}
DocsEnum docs2;
if (hasProx)
{
Debug.Assert(postings != null);
docs2 = postings;
}
else
{
Debug.Assert(docs != null);
docs2 = docs;
}
DocsEnum postingsDocs2;
if (!postingsTermsEnum.SeekExact(term))
{
throw new Exception("vector term=" + term + " field=" + field + " does not exist in postings; doc=" + j);
}
postingsPostings = postingsTermsEnum.DocsAndPositions(null, postingsPostings);
if (postingsPostings == null)
{
// Term vectors were indexed w/ pos but postings were not
postingsDocs = postingsTermsEnum.Docs(null, postingsDocs);
if (postingsDocs == null)
{
throw new Exception("vector term=" + term + " field=" + field + " does not exist in postings; doc=" + j);
}
}
if (postingsPostings != null)
{
postingsDocs2 = postingsPostings;
}
else
{
postingsDocs2 = postingsDocs;
}
int advanceDoc = postingsDocs2.Advance(j);
if (advanceDoc != j)
{
throw new Exception("vector term=" + term + " field=" + field + ": doc=" + j + " was not found in postings (got: " + advanceDoc + ")");
}
int doc = docs2.NextDoc();
if (doc != 0)
{
throw new Exception("vector for doc " + j + " didn't return docID=0: got docID=" + doc);
}
if (postingsHasFreq)
{
int tf = docs2.Freq();
if (postingsHasFreq && postingsDocs2.Freq() != tf)
{
throw new Exception("vector term=" + term + " field=" + field + " doc=" + j + ": freq=" + tf + " differs from postings freq=" + postingsDocs2.Freq());
}
if (hasProx)
{
for (int i = 0; i < tf; i++)
{
int pos = postings.NextPosition();
if (postingsPostings != null)
{
int postingsPos = postingsPostings.NextPosition();
if (terms.HasPositions() && pos != postingsPos)
{
throw new Exception("vector term=" + term + " field=" + field + " doc=" + j + ": pos=" + pos + " differs from postings pos=" + postingsPos);
}
}
// Call the methods to at least make
// sure they don't throw exc:
int startOffset = postings.StartOffset();
int endOffset = postings.EndOffset();
// TODO: these are too anal...?
/*
if (endOffset < startOffset) {
throw new RuntimeException("vector startOffset=" + startOffset + " is > endOffset=" + endOffset);
}
if (startOffset < lastStartOffset) {
throw new RuntimeException("vector startOffset=" + startOffset + " is < prior startOffset=" + lastStartOffset);
}
lastStartOffset = startOffset;
*/
if (postingsPostings != null)
{
int postingsStartOffset = postingsPostings.StartOffset();
int postingsEndOffset = postingsPostings.EndOffset();
if (startOffset != -1 && postingsStartOffset != -1 && startOffset != postingsStartOffset)
{
throw new Exception("vector term=" + term + " field=" + field + " doc=" + j + ": startOffset=" + startOffset + " differs from postings startOffset=" + postingsStartOffset);
}
if (endOffset != -1 && postingsEndOffset != -1 && endOffset != postingsEndOffset)
{
throw new Exception("vector term=" + term + " field=" + field + " doc=" + j + ": endOffset=" + endOffset + " differs from postings endOffset=" + postingsEndOffset);
}
}
BytesRef payload = postings.Payload;
if (payload != null)
{
Debug.Assert(vectorsHasPayload);
}
if (postingsHasPayload && vectorsHasPayload)
{
Debug.Assert(postingsPostings != null);
if (payload == null)
{
// we have payloads, but not at this position.
// postings has payloads too, it should not have one at this position
if (postingsPostings.Payload != null)
{
throw new Exception("vector term=" + term + " field=" + field + " doc=" + j + " has no payload but postings does: " + postingsPostings.Payload);
}
}
else
{
// we have payloads, and one at this position
// postings should also have one at this position, with the same bytes.
if (postingsPostings.Payload == null)
{
throw new Exception("vector term=" + term + " field=" + field + " doc=" + j + " has payload=" + payload + " but postings does not.");
}
BytesRef postingsPayload = postingsPostings.Payload;
if (!payload.Equals(postingsPayload))
{
throw new Exception("vector term=" + term + " field=" + field + " doc=" + j + " has payload=" + payload + " but differs from postings payload=" + postingsPayload);
}
}
}
}
}
}
}
}
}
}
}
float vectorAvg = status.DocCount == 0 ? 0 : status.TotVectors / (float)status.DocCount;
Msg(infoStream, "OK [" + status.TotVectors + " total vector count; avg " + vectorAvg.ToString(CultureInfo.InvariantCulture.NumberFormat) + " term/freq vector fields per doc]");
}
catch (Exception e)
{
Msg(infoStream, "ERROR [" + Convert.ToString(e.Message) + "]");
status.Error = e;
if (infoStream != null)
{
// LUCENENET NOTE: Some tests rely on the error type being in
// the message. We can't get the error type with StackTrace, we
// need ToString() for that.
infoStream.WriteLine(e.ToString());
//infoStream.WriteLine(e.StackTrace);
}
}
return status;
}
internal RandomBits(int maxDoc, double pctLive, Random random)
{
Bits = new FixedBitSet(maxDoc);
for (int i = 0; i < maxDoc; i++)
{
if (random.NextDouble() <= pctLive)
{
Bits.Set(i);
}
}
}
public FieldCacheDocIdSetAnonymousInnerClassHelper(FieldCacheTermsFilter outerInstance, int maxDoc, Bits acceptDocs, SortedDocValues fcsi, FixedBitSet bits)
: base(maxDoc, acceptDocs)
{
this.OuterInstance = outerInstance;
this.Fcsi = fcsi;
this.Bits = bits;
}
public virtual void TestBuildDocMap()
{
int maxDoc = TestUtil.NextInt(Random(), 1, 128);
int numDocs = TestUtil.NextInt(Random(), 0, maxDoc);
int numDeletedDocs = maxDoc - numDocs;
FixedBitSet liveDocs = new FixedBitSet(maxDoc);
for (int i = 0; i < numDocs; ++i)
{
while (true)
{
int docID = Random().Next(maxDoc);
if (!liveDocs.Get(docID))
{
liveDocs.Set(docID);
break;
}
}
}
MergeState.DocMap docMap = MergeState.DocMap.Build(maxDoc, liveDocs);
Assert.AreEqual(maxDoc, docMap.MaxDoc);
Assert.AreEqual(numDocs, docMap.NumDocs);
Assert.AreEqual(numDeletedDocs, docMap.NumDeletedDocs);
// assert the mapping is compact
for (int i = 0, del = 0; i < maxDoc; ++i)
{
if (!liveDocs.Get(i))
{
Assert.AreEqual(-1, docMap.Get(i));
++del;
}
else
{
Assert.AreEqual(i - del, docMap.Get(i));
}
}
}
protected internal virtual void AssertEquals(RandomTokenStream tk, FieldType ft, Terms terms)
{
Assert.AreEqual(1, terms.DocCount);
int termCount = (new HashSet<string>(Arrays.AsList(tk.Terms))).Count;
Assert.AreEqual(termCount, terms.Size());
Assert.AreEqual(termCount, terms.SumDocFreq);
Assert.AreEqual(ft.StoreTermVectorPositions, terms.HasPositions());
Assert.AreEqual(ft.StoreTermVectorOffsets, terms.HasOffsets());
Assert.AreEqual(ft.StoreTermVectorPayloads && tk.HasPayloads(), terms.HasPayloads());
HashSet<BytesRef> uniqueTerms = new HashSet<BytesRef>();
foreach (string term in tk.Freqs.Keys)
{
uniqueTerms.Add(new BytesRef(term));
}
BytesRef[] sortedTerms = uniqueTerms.ToArray(/*new BytesRef[0]*/);
Array.Sort(sortedTerms, terms.Comparator);
TermsEnum termsEnum = terms.Iterator(Random().NextBoolean() ? null : this.termsEnum.Value);
this.termsEnum.Value = termsEnum;
for (int i = 0; i < sortedTerms.Length; ++i)
{
BytesRef nextTerm = termsEnum.Next();
Assert.AreEqual(sortedTerms[i], nextTerm);
Assert.AreEqual(sortedTerms[i], termsEnum.Term());
Assert.AreEqual(1, termsEnum.DocFreq());
FixedBitSet bits = new FixedBitSet(1);
DocsEnum docsEnum = termsEnum.Docs(bits, Random().NextBoolean() ? null : this.docsEnum.Value);
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.NextDoc());
bits.Set(0);
docsEnum = termsEnum.Docs(Random().NextBoolean() ? bits : null, Random().NextBoolean() ? null : docsEnum);
Assert.IsNotNull(docsEnum);
Assert.AreEqual(0, docsEnum.NextDoc());
Assert.AreEqual(0, docsEnum.DocID());
Assert.AreEqual(tk.Freqs[termsEnum.Term().Utf8ToString()], (int?)docsEnum.Freq());
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsEnum.NextDoc());
this.docsEnum.Value = docsEnum;
bits.Clear(0);
DocsAndPositionsEnum docsAndPositionsEnum = termsEnum.DocsAndPositions(bits, Random().NextBoolean() ? null : this.docsAndPositionsEnum.Value);
Assert.AreEqual(ft.StoreTermVectorOffsets || ft.StoreTermVectorPositions, docsAndPositionsEnum != null);
if (docsAndPositionsEnum != null)
{
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsAndPositionsEnum.NextDoc());
}
bits.Set(0);
docsAndPositionsEnum = termsEnum.DocsAndPositions(Random().NextBoolean() ? bits : null, Random().NextBoolean() ? null : docsAndPositionsEnum);
Assert.AreEqual(ft.StoreTermVectorOffsets || ft.StoreTermVectorPositions, docsAndPositionsEnum != null);
if (terms.HasPositions() || terms.HasOffsets())
{
Assert.AreEqual(0, docsAndPositionsEnum.NextDoc());
int freq = docsAndPositionsEnum.Freq();
Assert.AreEqual(tk.Freqs[termsEnum.Term().Utf8ToString()], (int?)freq);
if (docsAndPositionsEnum != null)
{
for (int k = 0; k < freq; ++k)
{
int position = docsAndPositionsEnum.NextPosition();
ISet<int?> indexes;
if (terms.HasPositions())
{
indexes = tk.PositionToTerms[position];
Assert.IsNotNull(indexes);
}
else
{
indexes = tk.StartOffsetToTerms[docsAndPositionsEnum.StartOffset()];
Assert.IsNotNull(indexes);
}
if (terms.HasPositions())
{
bool foundPosition = false;
foreach (int index in indexes)
{
if (tk.TermBytes[index].Equals(termsEnum.Term()) && tk.Positions[index] == position)
{
foundPosition = true;
break;
}
}
Assert.IsTrue(foundPosition);
}
if (terms.HasOffsets())
{
bool foundOffset = false;
foreach (int index in indexes)
{
if (tk.TermBytes[index].Equals(termsEnum.Term()) && tk.StartOffsets[index] == docsAndPositionsEnum.StartOffset() && tk.EndOffsets[index] == docsAndPositionsEnum.EndOffset())
{
foundOffset = true;
break;
}
}
Assert.IsTrue(foundOffset);
}
if (terms.HasPayloads())
{
bool foundPayload = false;
foreach (int index in indexes)
{
if (tk.TermBytes[index].Equals(termsEnum.Term()) && Equals(tk.Payloads[index], docsAndPositionsEnum.Payload))
{
foundPayload = true;
break;
}
}
Assert.IsTrue(foundPayload);
}
}
try
{
docsAndPositionsEnum.NextPosition();
Assert.Fail();
}
catch (Exception e)
{
// ok
}
}
Assert.AreEqual(DocsEnum.NO_MORE_DOCS, docsAndPositionsEnum.NextDoc());
}
this.docsAndPositionsEnum.Value = docsAndPositionsEnum;
}
Assert.IsNull(termsEnum.Next());
for (int i = 0; i < 5; ++i)
{
if (Random().NextBoolean())
{
Assert.IsTrue(termsEnum.SeekExact(RandomInts.RandomFrom(Random(), tk.TermBytes)));
}
else
{
Assert.AreEqual(SeekStatus.FOUND, termsEnum.SeekCeil(RandomInts.RandomFrom(Random(), tk.TermBytes)));
}
}
}
// maxAllowed = the "highest" we can index, but we will still
// randomly index at lower IndexOption
private FieldsProducer BuildIndex(Directory dir, FieldInfo.IndexOptions maxAllowed, bool allowPayloads, bool alwaysTestMax)
{
Codec codec = Codec;
SegmentInfo segmentInfo = new SegmentInfo(dir, Constants.LUCENE_MAIN_VERSION, "_0", MaxDoc, false, codec, null);
int maxIndexOption = Enum.GetValues(typeof(FieldInfo.IndexOptions)).Cast<FieldInfo.IndexOptions>().ToList().IndexOf(maxAllowed);
if (VERBOSE)
{
Console.WriteLine("\nTEST: now build index");
}
int maxIndexOptionNoOffsets = Enum.GetValues(typeof(FieldInfo.IndexOptions)).Cast<FieldInfo.IndexOptions>().ToList().IndexOf(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS);
// TODO use allowPayloads
var newFieldInfoArray = new FieldInfo[Fields.Count];
for (int fieldUpto = 0; fieldUpto < Fields.Count; fieldUpto++)
{
FieldInfo oldFieldInfo = FieldInfos.FieldInfo(fieldUpto);
string pf = TestUtil.GetPostingsFormat(codec, oldFieldInfo.Name);
int fieldMaxIndexOption;
if (DoesntSupportOffsets.Contains(pf))
{
fieldMaxIndexOption = Math.Min(maxIndexOptionNoOffsets, maxIndexOption);
}
else
{
fieldMaxIndexOption = maxIndexOption;
}
// Randomly picked the IndexOptions to index this
// field with:
FieldInfo.IndexOptions indexOptions = Enum.GetValues(typeof(FieldInfo.IndexOptions)).Cast<FieldInfo.IndexOptions>().ToArray()[alwaysTestMax ? fieldMaxIndexOption : Random().Next(1 + fieldMaxIndexOption)];
bool doPayloads = indexOptions.CompareTo(FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) >= 0 && allowPayloads;
newFieldInfoArray[fieldUpto] = new FieldInfo(oldFieldInfo.Name, true, fieldUpto, false, false, doPayloads, indexOptions, null, DocValuesType.NUMERIC, null);
}
FieldInfos newFieldInfos = new FieldInfos(newFieldInfoArray);
// Estimate that flushed segment size will be 25% of
// what we use in RAM:
long bytes = TotalPostings * 8 + TotalPayloadBytes;
SegmentWriteState writeState = new SegmentWriteState(null, dir, segmentInfo, newFieldInfos, 32, null, new IOContext(new FlushInfo(MaxDoc, bytes)));
FieldsConsumer fieldsConsumer = codec.PostingsFormat().FieldsConsumer(writeState);
foreach (KeyValuePair<string, SortedDictionary<BytesRef, long>> fieldEnt in Fields)
{
string field = fieldEnt.Key;
IDictionary<BytesRef, long> terms = fieldEnt.Value;
FieldInfo fieldInfo = newFieldInfos.FieldInfo(field);
FieldInfo.IndexOptions? indexOptions = fieldInfo.FieldIndexOptions;
if (VERBOSE)
{
Console.WriteLine("field=" + field + " indexOtions=" + indexOptions);
}
bool doFreq = indexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS;
bool doPos = indexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS;
bool doPayloads = indexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS && allowPayloads;
bool doOffsets = indexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS;
TermsConsumer termsConsumer = fieldsConsumer.AddField(fieldInfo);
long sumTotalTF = 0;
long sumDF = 0;
FixedBitSet seenDocs = new FixedBitSet(MaxDoc);
foreach (KeyValuePair<BytesRef, long> termEnt in terms)
{
BytesRef term = termEnt.Key;
SeedPostings postings = GetSeedPostings(term.Utf8ToString(), termEnt.Value, false, maxAllowed);
if (VERBOSE)
{
Console.WriteLine(" term=" + field + ":" + term.Utf8ToString() + " docFreq=" + postings.DocFreq + " seed=" + termEnt.Value);
}
PostingsConsumer postingsConsumer = termsConsumer.StartTerm(term);
long totalTF = 0;
int docID = 0;
while ((docID = postings.NextDoc()) != DocsEnum.NO_MORE_DOCS)
{
int freq = postings.Freq();
if (VERBOSE)
{
Console.WriteLine(" " + postings.Upto + ": docID=" + docID + " freq=" + postings.Freq_Renamed);
}
postingsConsumer.StartDoc(docID, doFreq ? postings.Freq_Renamed : -1);
seenDocs.Set(docID);
if (doPos)
{
totalTF += postings.Freq_Renamed;
for (int posUpto = 0; posUpto < freq; posUpto++)
{
int pos = postings.NextPosition();
BytesRef payload = postings.Payload;
if (VERBOSE)
{
if (doPayloads)
{
Console.WriteLine(" pos=" + pos + " payload=" + (payload == null ? "null" : payload.Length + " bytes"));
}
else
{
Console.WriteLine(" pos=" + pos);
}
}
postingsConsumer.AddPosition(pos, doPayloads ? payload : null, doOffsets ? postings.StartOffset() : -1, doOffsets ? postings.EndOffset() : -1);
}
}
else if (doFreq)
{
totalTF += freq;
}
else
{
totalTF++;
}
postingsConsumer.FinishDoc();
}
termsConsumer.FinishTerm(term, new TermStats(postings.DocFreq, doFreq ? totalTF : -1));
sumTotalTF += totalTF;
sumDF += postings.DocFreq;
}
termsConsumer.Finish(doFreq ? sumTotalTF : -1, sumDF, seenDocs.Cardinality());
}
fieldsConsumer.Dispose();
if (VERBOSE)
{
Console.WriteLine("TEST: after indexing: files=");
foreach (string file in dir.ListAll())
{
Console.WriteLine(" " + file + ": " + dir.FileLength(file) + " bytes");
}
}
CurrentFieldInfos = newFieldInfos;
SegmentReadState readState = new SegmentReadState(dir, segmentInfo, newFieldInfos, IOContext.READ, 1);
return codec.PostingsFormat().FieldsProducer(readState);
}
public static void CreatePostings()
{
TotalPostings = 0;
TotalPayloadBytes = 0;
Fields = new SortedDictionary<string, SortedDictionary<BytesRef, long>>();
int numFields = TestUtil.NextInt(Random(), 1, 5);
if (VERBOSE)
{
Console.WriteLine("TEST: " + numFields + " fields");
}
MaxDoc = 0;
FieldInfo[] fieldInfoArray = new FieldInfo[numFields];
int fieldUpto = 0;
while (fieldUpto < numFields)
{
string field = TestUtil.RandomSimpleString(Random());
if (Fields.ContainsKey(field))
{
continue;
}
fieldInfoArray[fieldUpto] = new FieldInfo(field, true, fieldUpto, false, false, true, FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS, null, DocValuesType.NUMERIC, null);
fieldUpto++;
SortedDictionary<BytesRef, long> postings = new SortedDictionary<BytesRef, long>();
Fields[field] = postings;
HashSet<string> seenTerms = new HashSet<string>();
int numTerms;
if (Random().Next(10) == 7)
{
numTerms = AtLeast(50);
}
else
{
numTerms = TestUtil.NextInt(Random(), 2, 20);
}
for (int termUpto = 0; termUpto < numTerms; termUpto++)
{
string term = TestUtil.RandomSimpleString(Random());
if (seenTerms.Contains(term))
{
continue;
}
seenTerms.Add(term);
if (TEST_NIGHTLY && termUpto == 0 && fieldUpto == 1)
{
// Make 1 big term:
term = "big_" + term;
}
else if (termUpto == 1 && fieldUpto == 1)
{
// Make 1 medium term:
term = "medium_" + term;
}
else if (Random().NextBoolean())
{
// Low freq term:
term = "low_" + term;
}
else
{
// Very low freq term (don't multiply by RANDOM_MULTIPLIER):
term = "verylow_" + term;
}
long termSeed = Random().NextLong();
postings[new BytesRef(term)] = termSeed;
// NOTE: sort of silly: we enum all the docs just to
// get the maxDoc
DocsEnum docsEnum = GetSeedPostings(term, termSeed, false, FieldInfo.IndexOptions.DOCS_ONLY);
int doc;
int lastDoc = 0;
while ((doc = docsEnum.NextDoc()) != DocsEnum.NO_MORE_DOCS)
{
lastDoc = doc;
}
MaxDoc = Math.Max(lastDoc, MaxDoc);
}
}
FieldInfos = new FieldInfos(fieldInfoArray);
// It's the count, not the last docID:
MaxDoc++;
GlobalLiveDocs = new FixedBitSet(MaxDoc);
double liveRatio = Random().NextDouble();
for (int i = 0; i < MaxDoc; i++)
{
if (Random().NextDouble() <= liveRatio)
{
GlobalLiveDocs.Set(i);
}
}
AllTerms = new List<FieldAndTerm>();
foreach (KeyValuePair<string, SortedDictionary<BytesRef, long>> fieldEnt in Fields)
{
string field = fieldEnt.Key;
foreach (KeyValuePair<BytesRef, long> termEnt in fieldEnt.Value.EntrySet())
{
AllTerms.Add(new FieldAndTerm(field, termEnt.Key));
}
}
if (VERBOSE)
{
Console.WriteLine("TEST: done init postings; " + AllTerms.Count + " total terms, across " + FieldInfos.Size() + " fields");
}
}
/// <summary>
/// checks Fields api is consistent with itself.
/// searcher is optional, to verify with queries. Can be null.
/// </summary>
private static Status.TermIndexStatus CheckFields(Fields fields, Bits liveDocs, int maxDoc, FieldInfos fieldInfos, bool doPrint, bool isVectors, TextWriter infoStream, bool verbose)
{
// TODO: we should probably return our own stats thing...?!
Status.TermIndexStatus status = new Status.TermIndexStatus();
int computedFieldCount = 0;
if (fields == null)
{
Msg(infoStream, "OK [no fields/terms]");
return status;
}
DocsEnum docs = null;
DocsEnum docsAndFreqs = null;
DocsAndPositionsEnum postings = null;
string lastField = null;
foreach (string field in fields)
{
// MultiFieldsEnum relies upon this order...
if (lastField != null && field.CompareTo(lastField) <= 0)
{
throw new Exception("fields out of order: lastField=" + lastField + " field=" + field);
}
lastField = field;
// check that the field is in fieldinfos, and is indexed.
// TODO: add a separate test to check this for different reader impls
FieldInfo fieldInfo = fieldInfos.FieldInfo(field);
if (fieldInfo == null)
{
throw new Exception("fieldsEnum inconsistent with fieldInfos, no fieldInfos for: " + field);
}
if (!fieldInfo.Indexed)
{
throw new Exception("fieldsEnum inconsistent with fieldInfos, isIndexed == false for: " + field);
}
// TODO: really the codec should not return a field
// from FieldsEnum if it has no Terms... but we do
// this today:
// assert fields.terms(field) != null;
computedFieldCount++;
Terms terms = fields.Terms(field);
if (terms == null)
{
continue;
}
bool hasFreqs = terms.HasFreqs();
bool hasPositions = terms.HasPositions();
bool hasPayloads = terms.HasPayloads();
bool hasOffsets = terms.HasOffsets();
// term vectors cannot omit TF:
bool expectedHasFreqs = (isVectors || fieldInfo.FieldIndexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS);
if (hasFreqs != expectedHasFreqs)
{
throw new Exception("field \"" + field + "\" should have hasFreqs=" + expectedHasFreqs + " but got " + hasFreqs);
}
if (hasFreqs == false)
{
if (terms.SumTotalTermFreq != -1)
{
throw new Exception("field \"" + field + "\" hasFreqs is false, but Terms.getSumTotalTermFreq()=" + terms.SumTotalTermFreq + " (should be -1)");
}
}
if (!isVectors)
{
bool expectedHasPositions = fieldInfo.FieldIndexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS;
if (hasPositions != expectedHasPositions)
{
throw new Exception("field \"" + field + "\" should have hasPositions=" + expectedHasPositions + " but got " + hasPositions);
}
bool expectedHasPayloads = fieldInfo.HasPayloads();
if (hasPayloads != expectedHasPayloads)
{
throw new Exception("field \"" + field + "\" should have hasPayloads=" + expectedHasPayloads + " but got " + hasPayloads);
}
bool expectedHasOffsets = fieldInfo.FieldIndexOptions >= FieldInfo.IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS;
if (hasOffsets != expectedHasOffsets)
{
throw new Exception("field \"" + field + "\" should have hasOffsets=" + expectedHasOffsets + " but got " + hasOffsets);
}
}
TermsEnum termsEnum = terms.Iterator(null);
bool hasOrd = true;
long termCountStart = status.DelTermCount + status.TermCount;
BytesRef lastTerm = null;
IComparer<BytesRef> termComp = terms.Comparator;
long sumTotalTermFreq = 0;
long sumDocFreq = 0;
FixedBitSet visitedDocs = new FixedBitSet(maxDoc);
while (true)
{
BytesRef term = termsEnum.Next();
if (term == null)
{
break;
}
Debug.Assert(term.Valid);
// make sure terms arrive in order according to
// the comp
if (lastTerm == null)
{
lastTerm = BytesRef.DeepCopyOf(term);
}
else
{
if (termComp.Compare(lastTerm, term) >= 0)
{
throw new Exception("terms out of order: lastTerm=" + lastTerm + " term=" + term);
}
lastTerm.CopyBytes(term);
}
int docFreq = termsEnum.DocFreq();
if (docFreq <= 0)
{
throw new Exception("docfreq: " + docFreq + " is out of bounds");
}
sumDocFreq += docFreq;
docs = termsEnum.Docs(liveDocs, docs);
postings = termsEnum.DocsAndPositions(liveDocs, postings);
if (hasFreqs == false)
{
if (termsEnum.TotalTermFreq() != -1)
{
throw new Exception("field \"" + field + "\" hasFreqs is false, but TermsEnum.totalTermFreq()=" + termsEnum.TotalTermFreq() + " (should be -1)");
}
}
if (hasOrd)
{
long ord = -1;
try
{
ord = termsEnum.Ord();
}
catch (System.NotSupportedException uoe)
{
hasOrd = false;
}
if (hasOrd)
{
long ordExpected = status.DelTermCount + status.TermCount - termCountStart;
if (ord != ordExpected)
{
throw new Exception("ord mismatch: TermsEnum has ord=" + ord + " vs actual=" + ordExpected);
}
}
}
DocsEnum docs2;
if (postings != null)
{
docs2 = postings;
}
else
{
docs2 = docs;
}
int lastDoc = -1;
int docCount = 0;
long totalTermFreq = 0;
while (true)
{
int doc = docs2.NextDoc();
if (doc == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
status.TotFreq++;
visitedDocs.Set(doc);
int freq = -1;
if (hasFreqs)
{
freq = docs2.Freq();
if (freq <= 0)
{
throw new Exception("term " + term + ": doc " + doc + ": freq " + freq + " is out of bounds");
}
status.TotPos += freq;
totalTermFreq += freq;
}
else
{
// When a field didn't index freq, it must
// consistently "lie" and pretend that freq was
// 1:
if (docs2.Freq() != 1)
{
throw new Exception("term " + term + ": doc " + doc + ": freq " + freq + " != 1 when Terms.hasFreqs() is false");
}
}
docCount++;
if (doc <= lastDoc)
{
throw new Exception("term " + term + ": doc " + doc + " <= lastDoc " + lastDoc);
}
if (doc >= maxDoc)
{
throw new Exception("term " + term + ": doc " + doc + " >= maxDoc " + maxDoc);
}
lastDoc = doc;
int lastPos = -1;
int lastOffset = 0;
if (hasPositions)
{
for (int j = 0; j < freq; j++)
{
int pos = postings.NextPosition();
if (pos < 0)
{
throw new Exception("term " + term + ": doc " + doc + ": pos " + pos + " is out of bounds");
}
if (pos < lastPos)
{
throw new Exception("term " + term + ": doc " + doc + ": pos " + pos + " < lastPos " + lastPos);
}
lastPos = pos;
BytesRef payload = postings.Payload;
if (payload != null)
{
Debug.Assert(payload.Valid);
}
if (payload != null && payload.Length < 1)
{
throw new Exception("term " + term + ": doc " + doc + ": pos " + pos + " payload length is out of bounds " + payload.Length);
}
if (hasOffsets)
{
int startOffset = postings.StartOffset();
int endOffset = postings.EndOffset();
// NOTE: we cannot enforce any bounds whatsoever on vectors... they were a free-for-all before?
// but for offsets in the postings lists these checks are fine: they were always enforced by IndexWriter
if (!isVectors)
{
if (startOffset < 0)
{
throw new Exception("term " + term + ": doc " + doc + ": pos " + pos + ": startOffset " + startOffset + " is out of bounds");
}
if (startOffset < lastOffset)
{
throw new Exception("term " + term + ": doc " + doc + ": pos " + pos + ": startOffset " + startOffset + " < lastStartOffset " + lastOffset);
}
if (endOffset < 0)
{
throw new Exception("term " + term + ": doc " + doc + ": pos " + pos + ": endOffset " + endOffset + " is out of bounds");
}
if (endOffset < startOffset)
{
throw new Exception("term " + term + ": doc " + doc + ": pos " + pos + ": endOffset " + endOffset + " < startOffset " + startOffset);
}
}
lastOffset = startOffset;
}
}
}
}
if (docCount != 0)
{
status.TermCount++;
}
else
{
status.DelTermCount++;
}
long totalTermFreq2 = termsEnum.TotalTermFreq();
bool hasTotalTermFreq = hasFreqs && totalTermFreq2 != -1;
// Re-count if there are deleted docs:
if (liveDocs != null)
{
if (hasFreqs)
{
DocsEnum docsNoDel = termsEnum.Docs(null, docsAndFreqs);
docCount = 0;
totalTermFreq = 0;
while (docsNoDel.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
visitedDocs.Set(docsNoDel.DocID());
docCount++;
totalTermFreq += docsNoDel.Freq();
}
}
else
{
DocsEnum docsNoDel = termsEnum.Docs(null, docs, DocsEnum.FLAG_NONE);
docCount = 0;
totalTermFreq = -1;
while (docsNoDel.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
visitedDocs.Set(docsNoDel.DocID());
docCount++;
}
}
}
if (docCount != docFreq)
{
throw new Exception("term " + term + " docFreq=" + docFreq + " != tot docs w/o deletions " + docCount);
}
if (hasTotalTermFreq)
{
if (totalTermFreq2 <= 0)
{
throw new Exception("totalTermFreq: " + totalTermFreq2 + " is out of bounds");
}
sumTotalTermFreq += totalTermFreq;
if (totalTermFreq != totalTermFreq2)
{
throw new Exception("term " + term + " totalTermFreq=" + totalTermFreq2 + " != recomputed totalTermFreq=" + totalTermFreq);
}
}
// Test skipping
if (hasPositions)
{
for (int idx = 0; idx < 7; idx++)
{
int skipDocID = (int)(((idx + 1) * (long)maxDoc) / 8);
postings = termsEnum.DocsAndPositions(liveDocs, postings);
int docID = postings.Advance(skipDocID);
if (docID == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
else
{
if (docID < skipDocID)
{
throw new Exception("term " + term + ": advance(docID=" + skipDocID + ") returned docID=" + docID);
}
int freq = postings.Freq();
if (freq <= 0)
{
throw new Exception("termFreq " + freq + " is out of bounds");
}
int lastPosition = -1;
int lastOffset = 0;
for (int posUpto = 0; posUpto < freq; posUpto++)
{
int pos = postings.NextPosition();
if (pos < 0)
{
throw new Exception("position " + pos + " is out of bounds");
}
if (pos < lastPosition)
{
throw new Exception("position " + pos + " is < lastPosition " + lastPosition);
}
lastPosition = pos;
if (hasOffsets)
{
int startOffset = postings.StartOffset();
int endOffset = postings.EndOffset();
// NOTE: we cannot enforce any bounds whatsoever on vectors... they were a free-for-all before?
// but for offsets in the postings lists these checks are fine: they were always enforced by IndexWriter
if (!isVectors)
{
if (startOffset < 0)
{
throw new Exception("term " + term + ": doc " + docID + ": pos " + pos + ": startOffset " + startOffset + " is out of bounds");
}
if (startOffset < lastOffset)
{
throw new Exception("term " + term + ": doc " + docID + ": pos " + pos + ": startOffset " + startOffset + " < lastStartOffset " + lastOffset);
}
if (endOffset < 0)
{
throw new Exception("term " + term + ": doc " + docID + ": pos " + pos + ": endOffset " + endOffset + " is out of bounds");
}
if (endOffset < startOffset)
{
throw new Exception("term " + term + ": doc " + docID + ": pos " + pos + ": endOffset " + endOffset + " < startOffset " + startOffset);
}
}
lastOffset = startOffset;
}
}
int nextDocID = postings.NextDoc();
if (nextDocID == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
if (nextDocID <= docID)
{
throw new Exception("term " + term + ": advance(docID=" + skipDocID + "), then .next() returned docID=" + nextDocID + " vs prev docID=" + docID);
}
}
}
}
else
{
for (int idx = 0; idx < 7; idx++)
{
int skipDocID = (int)(((idx + 1) * (long)maxDoc) / 8);
docs = termsEnum.Docs(liveDocs, docs, DocsEnum.FLAG_NONE);
int docID = docs.Advance(skipDocID);
if (docID == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
else
{
if (docID < skipDocID)
{
throw new Exception("term " + term + ": advance(docID=" + skipDocID + ") returned docID=" + docID);
}
int nextDocID = docs.NextDoc();
if (nextDocID == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
if (nextDocID <= docID)
{
throw new Exception("term " + term + ": advance(docID=" + skipDocID + "), then .next() returned docID=" + nextDocID + " vs prev docID=" + docID);
}
}
}
}
}
Terms fieldTerms = fields.Terms(field);
if (fieldTerms == null)
{
// Unusual: the FieldsEnum returned a field but
// the Terms for that field is null; this should
// only happen if it's a ghost field (field with
// no terms, eg there used to be terms but all
// docs got deleted and then merged away):
}
else
{
if (fieldTerms is BlockTreeTermsReader.FieldReader)
{
BlockTreeTermsReader.Stats stats = ((BlockTreeTermsReader.FieldReader)fieldTerms).ComputeStats();
Debug.Assert(stats != null);
if (status.BlockTreeStats == null)
{
status.BlockTreeStats = new Dictionary<string, BlockTreeTermsReader.Stats>();
}
status.BlockTreeStats[field] = stats;
}
if (sumTotalTermFreq != 0)
{
long v = fields.Terms(field).SumTotalTermFreq;
if (v != -1 && sumTotalTermFreq != v)
{
throw new Exception("sumTotalTermFreq for field " + field + "=" + v + " != recomputed sumTotalTermFreq=" + sumTotalTermFreq);
}
}
if (sumDocFreq != 0)
{
long v = fields.Terms(field).SumDocFreq;
if (v != -1 && sumDocFreq != v)
{
throw new Exception("sumDocFreq for field " + field + "=" + v + " != recomputed sumDocFreq=" + sumDocFreq);
}
}
if (fieldTerms != null)
{
int v = fieldTerms.DocCount;
if (v != -1 && visitedDocs.Cardinality() != v)
{
throw new Exception("docCount for field " + field + "=" + v + " != recomputed docCount=" + visitedDocs.Cardinality());
}
}
// Test seek to last term:
if (lastTerm != null)
{
if (termsEnum.SeekCeil(lastTerm) != TermsEnum.SeekStatus.FOUND)
{
throw new Exception("seek to last term " + lastTerm + " failed");
}
int expectedDocFreq = termsEnum.DocFreq();
DocsEnum d = termsEnum.Docs(null, null, DocsEnum.FLAG_NONE);
int docFreq = 0;
while (d.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
docFreq++;
}
if (docFreq != expectedDocFreq)
{
throw new Exception("docFreq for last term " + lastTerm + "=" + expectedDocFreq + " != recomputed docFreq=" + docFreq);
}
}
// check unique term count
long termCount = -1;
if ((status.DelTermCount + status.TermCount) - termCountStart > 0)
{
termCount = fields.Terms(field).Size();
if (termCount != -1 && termCount != status.DelTermCount + status.TermCount - termCountStart)
{
throw new Exception("termCount mismatch " + (status.DelTermCount + termCount) + " vs " + (status.TermCount - termCountStart));
}
}
// Test seeking by ord
if (hasOrd && status.TermCount - termCountStart > 0)
{
int seekCount = (int)Math.Min(10000L, termCount);
if (seekCount > 0)
{
BytesRef[] seekTerms = new BytesRef[seekCount];
// Seek by ord
for (int i = seekCount - 1; i >= 0; i--)
{
long ord = i * (termCount / seekCount);
termsEnum.SeekExact(ord);
seekTerms[i] = BytesRef.DeepCopyOf(termsEnum.Term());
}
// Seek by term
long totDocCount = 0;
for (int i = seekCount - 1; i >= 0; i--)
{
if (termsEnum.SeekCeil(seekTerms[i]) != TermsEnum.SeekStatus.FOUND)
{
throw new Exception("seek to existing term " + seekTerms[i] + " failed");
}
docs = termsEnum.Docs(liveDocs, docs, DocsEnum.FLAG_NONE);
if (docs == null)
{
throw new Exception("null DocsEnum from to existing term " + seekTerms[i]);
}
while (docs.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
totDocCount++;
}
}
long totDocCountNoDeletes = 0;
long totDocFreq = 0;
for (int i = 0; i < seekCount; i++)
{
if (!termsEnum.SeekExact(seekTerms[i]))
{
throw new Exception("seek to existing term " + seekTerms[i] + " failed");
}
totDocFreq += termsEnum.DocFreq();
docs = termsEnum.Docs(null, docs, DocsEnum.FLAG_NONE);
if (docs == null)
{
throw new Exception("null DocsEnum from to existing term " + seekTerms[i]);
}
while (docs.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
totDocCountNoDeletes++;
}
}
if (totDocCount > totDocCountNoDeletes)
{
throw new Exception("more postings with deletes=" + totDocCount + " than without=" + totDocCountNoDeletes);
}
if (totDocCountNoDeletes != totDocFreq)
{
throw new Exception("docfreqs=" + totDocFreq + " != recomputed docfreqs=" + totDocCountNoDeletes);
}
}
}
}
}
int fieldCount = fields.Size;
if (fieldCount != -1)
{
if (fieldCount < 0)
{
throw new Exception("invalid fieldCount: " + fieldCount);
}
if (fieldCount != computedFieldCount)
{
throw new Exception("fieldCount mismatch " + fieldCount + " vs recomputed field count " + computedFieldCount);
}
}
// for most implementations, this is boring (just the sum across all fields)
// but codecs that don't work per-field like preflex actually implement this,
// but don't implement it on Terms, so the check isn't redundant.
long uniqueTermCountAllFields = fields.UniqueTermCount;
if (uniqueTermCountAllFields != -1 && status.TermCount + status.DelTermCount != uniqueTermCountAllFields)
{
throw new Exception("termCount mismatch " + uniqueTermCountAllFields + " vs " + (status.TermCount + status.DelTermCount));
}
if (doPrint)
{
Msg(infoStream, "OK [" + status.TermCount + " terms; " + status.TotFreq + " terms/docs pairs; " + status.TotPos + " tokens]");
}
if (verbose && status.BlockTreeStats != null && infoStream != null && status.TermCount > 0)
{
foreach (KeyValuePair<string, BlockTreeTermsReader.Stats> ent in status.BlockTreeStats)
{
infoStream.WriteLine(" field \"" + ent.Key + "\":");
infoStream.WriteLine(" " + ent.Value.ToString().Replace("\n", "\n "));
}
}
return status;
}
private static void CheckSortedDocValues(string fieldName, AtomicReader reader, SortedDocValues dv, Bits docsWithField)
{
CheckBinaryDocValues(fieldName, reader, dv, docsWithField);
int maxOrd = dv.ValueCount - 1;
FixedBitSet seenOrds = new FixedBitSet(dv.ValueCount);
int maxOrd2 = -1;
for (int i = 0; i < reader.MaxDoc; i++)
{
int ord = dv.GetOrd(i);
if (ord == -1)
{
if (docsWithField.Get(i))
{
throw new Exception("dv for field: " + fieldName + " has -1 ord but is not marked missing for doc: " + i);
}
}
else if (ord < -1 || ord > maxOrd)
{
throw new Exception("ord out of bounds: " + ord);
}
else
{
if (!docsWithField.Get(i))
{
throw new Exception("dv for field: " + fieldName + " is missing but has ord=" + ord + " for doc: " + i);
}
maxOrd2 = Math.Max(maxOrd2, ord);
seenOrds.Set(ord);
}
}
if (maxOrd != maxOrd2)
{
throw new Exception("dv for field: " + fieldName + " reports wrong maxOrd=" + maxOrd + " but this is not the case: " + maxOrd2);
}
if (seenOrds.Cardinality() != dv.ValueCount)
{
throw new Exception("dv for field: " + fieldName + " has holes in its ords, valueCount=" + dv.ValueCount + " but only used: " + seenOrds.Cardinality());
}
BytesRef lastValue = null;
BytesRef scratch = new BytesRef();
for (int i = 0; i <= maxOrd; i++)
{
dv.LookupOrd(i, scratch);
Debug.Assert(scratch.Valid);
if (lastValue != null)
{
if (scratch.CompareTo(lastValue) <= 0)
{
throw new Exception("dv for field: " + fieldName + " has ords out of order: " + lastValue + " >=" + scratch);
}
}
lastValue = BytesRef.DeepCopyOf(scratch);
}
}
internal Iterator(int size, PagedGrowableWriter offsets, PagedGrowableWriter lengths, PagedMutable docs, BytesRef values, FixedBitSet docsWithField)
{
this.Offsets = offsets;
this.Size = size;
this.Lengths = lengths;
this.Docs = docs;
this.DocsWithField = docsWithField;
Value_Renamed = (BytesRef)values.Clone();
}
private void VerifyCount(IndexReader ir)
{
Fields fields = MultiFields.GetFields(ir);
if (fields == null)
{
return;
}
foreach (string field in fields)
{
Terms terms = fields.Terms(field);
if (terms == null)
{
continue;
}
int docCount = terms.DocCount;
FixedBitSet visited = new FixedBitSet(ir.MaxDoc);
TermsEnum te = terms.Iterator(null);
while (te.Next() != null)
{
DocsEnum de = TestUtil.Docs(Random(), te, null, null, DocsEnum.FLAG_NONE);
while (de.NextDoc() != DocIdSetIterator.NO_MORE_DOCS)
{
visited.Set(de.DocID());
}
}
Assert.AreEqual(visited.Cardinality(), docCount);
}
}
public override DocIdSet GetDocIdSet(AtomicReaderContext context, Bits acceptDocs)
{
Assert.IsNull(acceptDocs, "acceptDocs should be null, as we have an index without deletions");
FixedBitSet set = new FixedBitSet(context.Reader.MaxDoc);
int docBase = context.DocBase;
int limit = docBase + context.Reader.MaxDoc;
for (int index = 0; index < Docs.Length; index++)
{
int docId = Docs[index];
if (docId >= docBase && docId < limit)
{
set.Set(docId - docBase);
}
}
return set.Cardinality() == 0 ? null : set;
}
public virtual void AddValue(int docID, long value)
{
if (docID < Pending.Size())
{
throw new System.ArgumentException("DocValuesField \"" + FieldInfo.Name + "\" appears more than once in this document (only one value is allowed per field)");
}
// Fill in any holes:
for (int i = (int)Pending.Size(); i < docID; ++i)
{
Pending.Add(MISSING);
}
Pending.Add(value);
if (DocsWithField != null)
{
DocsWithField = FixedBitSet.EnsureCapacity(DocsWithField, docID);
DocsWithField.Set(docID);
}
UpdateBytesUsed();
}
public virtual void DoTestLongPostingsNoPositions(FieldInfo.IndexOptions options)
{
// Don't use TestUtil.getTempDir so that we own the
// randomness (ie same seed will point to same dir):
Directory dir = NewFSDirectory(CreateTempDir("longpostings" + "." + Random().NextLong()));
int NUM_DOCS = AtLeast(2000);
if (VERBOSE)
{
Console.WriteLine("TEST: NUM_DOCS=" + NUM_DOCS);
}
string s1 = GetRandomTerm(null);
string s2 = GetRandomTerm(s1);
if (VERBOSE)
{
Console.WriteLine("\nTEST: s1=" + s1 + " s2=" + s2);
/*
for(int idx=0;idx<s1.Length();idx++) {
System.out.println(" s1 ch=0x" + Integer.toHexString(s1.charAt(idx)));
}
for(int idx=0;idx<s2.Length();idx++) {
System.out.println(" s2 ch=0x" + Integer.toHexString(s2.charAt(idx)));
}
*/
}
FixedBitSet isS1 = new FixedBitSet(NUM_DOCS);
for (int idx = 0; idx < NUM_DOCS; idx++)
{
if (Random().NextBoolean())
{
isS1.Set(idx);
}
}
IndexReader r;
if (true)
{
IndexWriterConfig iwc = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())).SetOpenMode(IndexWriterConfig.OpenMode_e.CREATE).SetMergePolicy(NewLogMergePolicy());
iwc.SetRAMBufferSizeMB(16.0 + 16.0 * Random().NextDouble());
iwc.SetMaxBufferedDocs(-1);
RandomIndexWriter riw = new RandomIndexWriter(Random(), dir, iwc);
FieldType ft = new FieldType(TextField.TYPE_NOT_STORED);
ft.IndexOptions = options;
for (int idx = 0; idx < NUM_DOCS; idx++)
{
Document doc = new Document();
string s = isS1.Get(idx) ? s1 : s2;
Field f = NewField("field", s, ft);
int count = TestUtil.NextInt(Random(), 1, 4);
for (int ct = 0; ct < count; ct++)
{
doc.Add(f);
}
riw.AddDocument(doc);
}
r = riw.Reader;
riw.Dispose();
}
else
{
r = DirectoryReader.Open(dir);
}
/*
if (VERBOSE) {
System.out.println("TEST: terms");
TermEnum termEnum = r.Terms();
while(termEnum.Next()) {
System.out.println(" term=" + termEnum.Term() + " len=" + termEnum.Term().Text().Length());
Assert.IsTrue(termEnum.DocFreq() > 0);
System.out.println(" s1?=" + (termEnum.Term().Text().equals(s1)) + " s1len=" + s1.Length());
System.out.println(" s2?=" + (termEnum.Term().Text().equals(s2)) + " s2len=" + s2.Length());
final String s = termEnum.Term().Text();
for(int idx=0;idx<s.Length();idx++) {
System.out.println(" ch=0x" + Integer.toHexString(s.charAt(idx)));
}
}
}
*/
Assert.AreEqual(NUM_DOCS, r.NumDocs);
Assert.IsTrue(r.DocFreq(new Term("field", s1)) > 0);
Assert.IsTrue(r.DocFreq(new Term("field", s2)) > 0);
int num = AtLeast(1000);
for (int iter = 0; iter < num; iter++)
{
string term;
bool doS1;
if (Random().NextBoolean())
{
term = s1;
doS1 = true;
}
else
{
term = s2;
doS1 = false;
}
if (VERBOSE)
{
Console.WriteLine("\nTEST: iter=" + iter + " doS1=" + doS1 + " term=" + term);
}
DocsEnum docs;
DocsEnum postings;
if (options == FieldInfo.IndexOptions.DOCS_ONLY)
{
docs = TestUtil.Docs(Random(), r, "field", new BytesRef(term), null, null, DocsEnum.FLAG_NONE);
postings = null;
}
else
{
docs = postings = TestUtil.Docs(Random(), r, "field", new BytesRef(term), null, null, DocsEnum.FLAG_FREQS);
Debug.Assert(postings != null);
}
Debug.Assert(docs != null);
int docID = -1;
while (docID < DocIdSetIterator.NO_MORE_DOCS)
{
int what = Random().Next(3);
if (what == 0)
{
if (VERBOSE)
{
Console.WriteLine("TEST: docID=" + docID + "; do next()");
}
// nextDoc
int expected = docID + 1;
while (true)
{
if (expected == NUM_DOCS)
{
expected = int.MaxValue;
break;
}
else if (isS1.Get(expected) == doS1)
{
break;
}
else
{
expected++;
}
}
docID = docs.NextDoc();
if (VERBOSE)
{
Console.WriteLine(" got docID=" + docID);
}
Assert.AreEqual(expected, docID);
if (docID == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
if (Random().Next(6) == 3 && postings != null)
{
int freq = postings.Freq();
Assert.IsTrue(freq >= 1 && freq <= 4);
}
}
else
{
// advance
int targetDocID;
if (docID == -1)
{
targetDocID = Random().Next(NUM_DOCS + 1);
}
else
{
targetDocID = docID + TestUtil.NextInt(Random(), 1, NUM_DOCS - docID);
}
if (VERBOSE)
{
Console.WriteLine("TEST: docID=" + docID + "; do advance(" + targetDocID + ")");
}
int expected = targetDocID;
while (true)
{
if (expected == NUM_DOCS)
{
expected = int.MaxValue;
break;
}
else if (isS1.Get(expected) == doS1)
{
break;
}
else
{
expected++;
}
}
docID = docs.Advance(targetDocID);
if (VERBOSE)
{
Console.WriteLine(" got docID=" + docID);
}
Assert.AreEqual(expected, docID);
if (docID == DocIdSetIterator.NO_MORE_DOCS)
{
break;
}
if (Random().Next(6) == 3 && postings != null)
{
int freq = postings.Freq();
Assert.IsTrue(freq >= 1 && freq <= 4, "got invalid freq=" + freq);
}
}
}
}
r.Dispose();
dir.Dispose();
}
public override void Merge(DocValuesFieldUpdates other)
{
BinaryDocValuesFieldUpdates otherUpdates = (BinaryDocValuesFieldUpdates)other;
int newSize = Size + otherUpdates.Size;
if (newSize > int.MaxValue)
{
throw new InvalidOperationException("cannot support more than Integer.MAX_VALUE doc/value entries; size=" + Size + " other.size=" + otherUpdates.Size);
}
Docs = Docs.Grow(newSize);
Offsets = Offsets.Grow(newSize);
Lengths = Lengths.Grow(newSize);
DocsWithField = FixedBitSet.EnsureCapacity(DocsWithField, (int)Docs.Size());
for (int i = 0; i < otherUpdates.Size; i++)
{
int doc = (int)otherUpdates.Docs.Get(i);
if (otherUpdates.DocsWithField.Get(i))
{
DocsWithField.Set(Size);
}
Docs.Set(Size, doc);
Offsets.Set(Size, Values.Length + otherUpdates.Offsets.Get(i)); // correct relative offset
Lengths.Set(Size, otherUpdates.Lengths.Get(i));
++Size;
}
Values.Append(otherUpdates.Values);
}
protected override DocIdSet CacheImpl(DocIdSetIterator iterator, AtomicReader reader)
{
var cached = new FixedBitSet(reader.MaxDoc);
filterWasUsed.Set(true);
cached.Or(iterator);
return cached;
}
public InPlaceMergeSorterAnonymousInnerClassHelper(BinaryDocValuesFieldUpdates outerInstance, PagedMutable docs, PagedGrowableWriter offsets, PagedGrowableWriter lengths, FixedBitSet docsWithField)
{
this.OuterInstance = outerInstance;
this.Docs = docs;
this.Offsets = offsets;
this.Lengths = lengths;
this.DocsWithField = docsWithField;
}
public CollectorAnonymousInnerClassHelper(TestBooleanOr outerInstance, BulkScorer scorer, FixedBitSet hits, AtomicInteger end)
{
this.OuterInstance = outerInstance;
this.scorer = scorer;
this.Hits = hits;
this.End = end;
}
public override void Add(int doc, object value)
{
// TODO: if the Sorter interface changes to take long indexes, we can remove that limitation
if (Size == int.MaxValue)
{
throw new InvalidOperationException("cannot support more than Integer.MAX_VALUE doc/value entries");
}
BytesRef val = (BytesRef)value;
if (val == null)
{
val = BinaryDocValuesUpdate.MISSING;
}
// grow the structures to have room for more elements
if (Docs.Size() == Size)
{
Docs = Docs.Grow(Size + 1);
Offsets = Offsets.Grow(Size + 1);
Lengths = Lengths.Grow(Size + 1);
DocsWithField = FixedBitSet.EnsureCapacity(DocsWithField, (int)Docs.Size());
}
if (val != BinaryDocValuesUpdate.MISSING)
{
// only mark the document as having a value in that field if the value wasn't set to null (MISSING)
DocsWithField.Set(Size);
}
Docs.Set(Size, doc);
Offsets.Set(Size, Values.Length);
Lengths.Set(Size, val.Length);
Values.Append(val);
++Size;
}
public virtual void TestBooleanScorerMax()
{
Directory dir = NewDirectory();
RandomIndexWriter riw = new RandomIndexWriter(Random(), dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random())));
int docCount = AtLeast(10000);
for (int i = 0; i < docCount; i++)
{
Document doc = new Document();
doc.Add(NewField("field", "a", TextField.TYPE_NOT_STORED));
riw.AddDocument(doc);
}
riw.ForceMerge(1);
IndexReader r = riw.Reader;
riw.Dispose();
IndexSearcher s = NewSearcher(r);
BooleanQuery bq = new BooleanQuery();
bq.Add(new TermQuery(new Term("field", "a")), BooleanClause.Occur.SHOULD);
bq.Add(new TermQuery(new Term("field", "a")), BooleanClause.Occur.SHOULD);
Weight w = s.CreateNormalizedWeight(bq);
Assert.AreEqual(1, s.IndexReader.Leaves.Count);
BulkScorer scorer = w.BulkScorer(s.IndexReader.Leaves[0], false, null);
FixedBitSet hits = new FixedBitSet(docCount);
AtomicInteger end = new AtomicInteger();
Collector c = new CollectorAnonymousInnerClassHelper(this, scorer, hits, end);
while (end.Get() < docCount)
{
int inc = TestUtil.NextInt(Random(), 1, 1000);
end.AddAndGet(inc);
scorer.Score(c, end.Get());
}
Assert.AreEqual(docCount, hits.Cardinality());
r.Dispose();
dir.Dispose();
}
