/// <summary>
/// add a term </summary>
public virtual void Add(Term term)
{
Debug.Assert(LastTerm.Equals(new Term("")) || term.CompareTo(LastTerm) > 0);
try
{
int prefix = SharedPrefix(LastTerm.Bytes_Renamed, term.Bytes_Renamed);
int suffix = term.Bytes_Renamed.Length - prefix;
if (term.Field_Renamed.Equals(LastTerm.Field_Renamed))
{
Output.WriteVInt(prefix << 1);
}
else
{
Output.WriteVInt(prefix << 1 | 1);
Output.WriteString(term.Field_Renamed);
}
Output.WriteVInt(suffix);
Output.WriteBytes(term.Bytes_Renamed.Bytes, term.Bytes_Renamed.Offset + prefix, suffix);
LastTerm.Bytes_Renamed.CopyBytes(term.Bytes_Renamed);
LastTerm.Field_Renamed = term.Field_Renamed;
}
catch (IOException e)
{
throw new Exception(e.Message, e);
}
}
/// <summary>
/// add a term </summary>
public virtual void Add(Term term)
{
Debug.Assert(lastTerm.Equals(new Term("")) || term.CompareTo(lastTerm) > 0);
try
{
int prefix = SharedPrefix(lastTerm.Bytes, term.Bytes);
int suffix = term.Bytes.Length - prefix;
if (term.Field.Equals(lastTerm.Field, StringComparison.Ordinal))
{
output.WriteVInt32(prefix << 1);
}
else
{
output.WriteVInt32(prefix << 1 | 1);
output.WriteString(term.Field);
}
output.WriteVInt32(suffix);
output.WriteBytes(term.Bytes.Bytes, term.Bytes.Offset + prefix, suffix);
lastTerm.Bytes.CopyBytes(term.Bytes);
lastTerm.Field = term.Field;
}
catch (IOException e)
{
throw new Exception(e.ToString(), e);
}
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: @Override public void addPosition(int pos, util.BytesRef payload, int startOffset, int endOffset) throws java.io.IOException
public override void addPosition(int pos, BytesRef payload, int startOffset, int endOffset)
{
Debug.Assert(payload == null || outerInstance.field.hasPayloads());
//System.out.println(" addPos pos=" + pos + " payload=" + payload);
//JAVA TO C# CONVERTER WARNING: The original Java variable was marked 'final':
//ORIGINAL LINE: final int delta = pos - lastPos;
int delta = pos - lastPos;
Debug.Assert(delta >= 0);
lastPos = pos;
int payloadLen = 0;
if (outerInstance.field.hasPayloads())
{
payloadLen = payload == null ? 0 : payload.length;
if (payloadLen != lastPayloadLen)
{
lastPayloadLen = payloadLen;
buffer.WriteVInt((delta << 1) | 1);
buffer.WriteVInt(payloadLen);
}
else
{
buffer.WriteVInt(delta << 1);
}
}
else
{
buffer.WriteVInt(delta);
}
if (outerInstance.field.IndexOptions.compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) >= 0)
{
// don't use startOffset - lastEndOffset, because this creates lots of negative vints for synonyms,
// and the numbers aren't that much smaller anyways.
int offsetDelta = startOffset - lastOffset;
int offsetLength = endOffset - startOffset;
if (offsetLength != lastOffsetLength)
{
buffer.WriteVInt(offsetDelta << 1 | 1);
buffer.WriteVInt(offsetLength);
}
else
{
buffer.WriteVInt(offsetDelta << 1);
}
lastOffset = startOffset;
lastOffsetLength = offsetLength;
}
if (payloadLen > 0)
{
buffer.WriteBytes(payload.bytes, payload.offset, payloadLen);
}
}
public override void AddPosition(int pos, BytesRef payload, int startOffset, int endOffset)
{
Debug.Assert(payload == null || outerInstance.field.HasPayloads());
//System.out.println(" addPos pos=" + pos + " payload=" + payload);
int delta = pos - lastPos;
Debug.Assert(delta >= 0);
lastPos = pos;
int payloadLen = 0;
if (outerInstance.field.HasPayloads())
{
payloadLen = payload == null ? 0 : payload.Length;
if (payloadLen != lastPayloadLen)
{
lastPayloadLen = payloadLen;
buffer.WriteVInt((delta << 1) | 1);
buffer.WriteVInt(payloadLen);
}
else
{
buffer.WriteVInt(delta << 1);
}
}
else
{
buffer.WriteVInt(delta);
}
if (outerInstance.field.FieldIndexOptions.Value.CompareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) >= 0)
{
// don't use startOffset - lastEndOffset, because this creates lots of negative vints for synonyms,
// and the numbers aren't that much smaller anyways.
int offsetDelta = startOffset - lastOffset;
int offsetLength = endOffset - startOffset;
if (offsetLength != lastOffsetLength)
{
buffer.WriteVInt(offsetDelta << 1 | 1);
buffer.WriteVInt(offsetLength);
}
else
{
buffer.WriteVInt(offsetDelta << 1);
}
lastOffset = startOffset;
lastOffsetLength = offsetLength;
}
if (payloadLen > 0)
{
buffer.WriteBytes(payload.Bytes, payload.Offset, payloadLen);
}
}
/// <summary>
/// Called when we are done adding docs to this term
/// </summary>
/// <param name="state"></param>
public override void FinishTerm(BlockTermState state)
{
var state2 = (PulsingTermState)state;
Debug.Assert(_pendingCount > 0 || _pendingCount == -1);
if (_pendingCount == -1)
{
state2.wrappedState.DocFreq = state2.DocFreq;
state2.wrappedState.TotalTermFreq = state2.TotalTermFreq;
state2.bytes = null;
_wrappedPostingsWriter.FinishTerm(state2.wrappedState);
}
else
{
// There were few enough total occurrences for this
// term, so we fully inline our postings data into
// terms dict, now:
// TODO: it'd be better to share this encoding logic
// in some inner codec that knows how to write a
// single doc / single position, etc. This way if a
// given codec wants to store other interesting
// stuff, it could use this pulsing codec to do so
if (_indexOptions.CompareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) >= 0)
{
var lastDocID = 0;
var pendingIDX = 0;
var lastPayloadLength = -1;
var lastOffsetLength = -1;
while (pendingIDX < _pendingCount)
{
var doc = _pending[pendingIDX];
var delta = doc.docID - lastDocID;
lastDocID = doc.docID;
// if (DEBUG) System.out.println(" write doc=" + doc.docID + " freq=" + doc.termFreq);
if (doc.termFreq == 1)
{
_buffer.WriteVInt32((delta << 1) | 1);
}
else
{
_buffer.WriteVInt32(delta << 1);
_buffer.WriteVInt32(doc.termFreq);
}
var lastPos = 0;
var lastOffset = 0;
for (var posIDX = 0; posIDX < doc.termFreq; posIDX++)
{
var pos = _pending[pendingIDX++];
Debug.Assert(pos.docID == doc.docID);
var posDelta = pos.pos - lastPos;
lastPos = pos.pos;
var payloadLength = pos.payload == null ? 0 : pos.payload.Length;
if (_storePayloads)
{
if (payloadLength != lastPayloadLength)
{
_buffer.WriteVInt32((posDelta << 1) | 1);
_buffer.WriteVInt32(payloadLength);
lastPayloadLength = payloadLength;
}
else
{
_buffer.WriteVInt32(posDelta << 1);
}
}
else
{
_buffer.WriteVInt32(posDelta);
}
if (_indexOptions.CompareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) >= 0)
{
//System.out.println("write=" + pos.startOffset + "," + pos.endOffset);
var offsetDelta = pos.startOffset - lastOffset;
var offsetLength = pos.endOffset - pos.startOffset;
if (offsetLength != lastOffsetLength)
{
_buffer.WriteVInt32(offsetDelta << 1 | 1);
_buffer.WriteVInt32(offsetLength);
}
else
{
_buffer.WriteVInt32(offsetDelta << 1);
}
lastOffset = pos.startOffset;
lastOffsetLength = offsetLength;
}
if (payloadLength > 0)
{
Debug.Assert(_storePayloads);
_buffer.WriteBytes(pos.payload.Bytes, 0, pos.payload.Length);
}
}
}
}
else if (_indexOptions == IndexOptions.DOCS_AND_FREQS)
{
int lastDocId = 0;
for (int posIdx = 0; posIdx < _pendingCount; posIdx++)
{
Position doc = _pending[posIdx];
int delta = doc.docID - lastDocId;
Debug.Assert(doc.termFreq != 0);
if (doc.termFreq == 1)
{
_buffer.WriteVInt32((delta << 1) | 1);
}
else
{
_buffer.WriteVInt32(delta << 1);
_buffer.WriteVInt32(doc.termFreq);
}
lastDocId = doc.docID;
}
}
else if (_indexOptions == IndexOptions.DOCS_ONLY)
{
int lastDocId = 0;
for (int posIdx = 0; posIdx < _pendingCount; posIdx++)
{
Position doc = _pending[posIdx];
_buffer.WriteVInt32(doc.docID - lastDocId);
lastDocId = doc.docID;
}
}
state2.bytes = new byte[(int)_buffer.GetFilePointer()];
_buffer.WriteTo(state2.bytes, 0);
_buffer.Reset();
}
_pendingCount = 0;
}
public override void AddPosition(int pos, BytesRef payload, int startOffset, int endOffset)
{
if (Debugging.AssertsEnabled)
{
Debugging.Assert(payload == null || outerInstance.field.HasPayloads);
}
//System.out.println(" addPos pos=" + pos + " payload=" + payload);
int delta = pos - lastPos;
if (Debugging.AssertsEnabled)
{
Debugging.Assert(delta >= 0);
}
lastPos = pos;
int payloadLen = 0;
if (outerInstance.field.HasPayloads)
{
payloadLen = payload == null ? 0 : payload.Length;
if (payloadLen != lastPayloadLen)
{
lastPayloadLen = payloadLen;
buffer.WriteVInt32((delta << 1) | 1);
buffer.WriteVInt32(payloadLen);
}
else
{
buffer.WriteVInt32(delta << 1);
}
}
else
{
buffer.WriteVInt32(delta);
}
// LUCENENET specific - to avoid boxing, changed from CompareTo() to IndexOptionsComparer.Compare()
if (IndexOptionsComparer.Default.Compare(outerInstance.field.IndexOptions, IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) >= 0)
{
// don't use startOffset - lastEndOffset, because this creates lots of negative vints for synonyms,
// and the numbers aren't that much smaller anyways.
int offsetDelta = startOffset - lastOffset;
int offsetLength = endOffset - startOffset;
if (offsetLength != lastOffsetLength)
{
buffer.WriteVInt32(offsetDelta << 1 | 1);
buffer.WriteVInt32(offsetLength);
}
else
{
buffer.WriteVInt32(offsetDelta << 1);
}
lastOffset = startOffset;
lastOffsetLength = offsetLength;
}
if (payloadLen > 0)
{
buffer.WriteBytes(payload.Bytes, payload.Offset, payloadLen);
}
}
private void FlushBlock()
{
//System.out.println("BTW.flushBlock seg=" + segment + " pendingCount=" + pendingCount + " fp=" + out.getFilePointer());
// First pass: compute common prefix for all terms
// in the block, against term before first term in
// this block:
int commonPrefix = SharedPrefix(lastPrevTerm, pendingTerms[0].Term);
for (int termCount = 1; termCount < pendingCount; termCount++)
{
commonPrefix = Math.Min(commonPrefix,
SharedPrefix(lastPrevTerm,
pendingTerms[termCount].Term));
}
outerInstance.m_output.WriteVInt32(pendingCount);
outerInstance.m_output.WriteVInt32(commonPrefix);
// 2nd pass: write suffixes, as separate byte[] blob
for (int termCount = 0; termCount < pendingCount; termCount++)
{
int suffix = pendingTerms[termCount].Term.Length - commonPrefix;
// TODO: cutover to better intblock codec, instead
// of interleaving here:
bytesWriter.WriteVInt32(suffix);
bytesWriter.WriteBytes(pendingTerms[termCount].Term.Bytes, commonPrefix, suffix);
}
outerInstance.m_output.WriteVInt32((int)bytesWriter.Position); // LUCENENET specific: Renamed from getFilePointer() to match FileStream
bytesWriter.WriteTo(outerInstance.m_output);
bytesWriter.Reset();
// 3rd pass: write the freqs as byte[] blob
// TODO: cutover to better intblock codec. simple64?
// write prefix, suffix first:
for (int termCount = 0; termCount < pendingCount; termCount++)
{
BlockTermState state = pendingTerms[termCount].State;
if (Debugging.AssertsEnabled)
{
Debugging.Assert(state != null);
}
bytesWriter.WriteVInt32(state.DocFreq);
if (fieldInfo.IndexOptions != IndexOptions.DOCS_ONLY)
{
bytesWriter.WriteVInt64(state.TotalTermFreq - state.DocFreq);
}
}
outerInstance.m_output.WriteVInt32((int)bytesWriter.Position); // LUCENENET specific: Renamed from getFilePointer() to match FileStream
bytesWriter.WriteTo(outerInstance.m_output);
bytesWriter.Reset();
// 4th pass: write the metadata
long[] longs = new long[longsSize];
bool absolute = true;
for (int termCount = 0; termCount < pendingCount; termCount++)
{
BlockTermState state = pendingTerms[termCount].State;
postingsWriter.EncodeTerm(longs, bufferWriter, fieldInfo, state, absolute);
for (int i = 0; i < longsSize; i++)
{
bytesWriter.WriteVInt64(longs[i]);
}
bufferWriter.WriteTo(bytesWriter);
bufferWriter.Reset();
absolute = false;
}
outerInstance.m_output.WriteVInt32((int)bytesWriter.Position); // LUCENENET specific: Renamed from getFilePointer() to match FileStream
bytesWriter.WriteTo(outerInstance.m_output);
bytesWriter.Reset();
lastPrevTerm.CopyBytes(pendingTerms[pendingCount - 1].Term);
pendingCount = 0;
}
public DirectField(SegmentReadState state, string field, Terms termsIn, int minSkipCount, int lowFreqCutoff)
{
FieldInfo fieldInfo = state.FieldInfos.FieldInfo(field);
sumTotalTermFreq = termsIn.SumTotalTermFreq;
sumDocFreq = termsIn.SumDocFreq;
docCount = termsIn.DocCount;
int numTerms = (int) termsIn.Size();
if (numTerms == -1)
{
throw new System.ArgumentException("codec does not provide Terms.size()");
}
terms = new TermAndSkip[numTerms];
termOffsets = new int[1 + numTerms];
byte[] termBytes = new byte[1024];
this.minSkipCount = minSkipCount;
hasFreq = fieldInfo.FieldIndexOptions.Value.CompareTo(IndexOptions.DOCS_ONLY) > 0;
hasPos = fieldInfo.FieldIndexOptions.Value.CompareTo(IndexOptions.DOCS_AND_FREQS) > 0;
hasOffsets_Renamed = fieldInfo.FieldIndexOptions.Value.CompareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) > 0;
hasPayloads_Renamed = fieldInfo.HasPayloads();
BytesRef term;
DocsEnum docsEnum = null;
DocsAndPositionsEnum docsAndPositionsEnum = null;
TermsEnum termsEnum = termsIn.Iterator(null);
int termOffset = 0;
IntArrayWriter scratch = new IntArrayWriter();
// Used for payloads, if any:
RAMOutputStream ros = new RAMOutputStream();
// if (DEBUG) {
// System.out.println("\nLOAD terms seg=" + state.segmentInfo.name + " field=" + field + " hasOffsets=" + hasOffsets + " hasFreq=" + hasFreq + " hasPos=" + hasPos + " hasPayloads=" + hasPayloads);
// }
while ((term = termsEnum.Next()) != null)
{
int docFreq = termsEnum.DocFreq();
long totalTermFreq = termsEnum.TotalTermFreq();
// if (DEBUG) {
// System.out.println(" term=" + term.utf8ToString());
// }
termOffsets[count] = termOffset;
if (termBytes.Length < (termOffset + term.Length))
{
termBytes = ArrayUtil.Grow(termBytes, termOffset + term.Length);
}
Array.Copy(term.Bytes, term.Offset, termBytes, termOffset, term.Length);
termOffset += term.Length;
termOffsets[count + 1] = termOffset;
if (hasPos)
{
docsAndPositionsEnum = termsEnum.DocsAndPositions(null, docsAndPositionsEnum);
}
else
{
docsEnum = termsEnum.Docs(null, docsEnum);
}
TermAndSkip ent;
DocsEnum docsEnum2;
docsEnum2 = hasPos ? docsAndPositionsEnum : docsEnum;
int docID;
if (docFreq <= lowFreqCutoff)
{
ros.Reset();
// Pack postings for low-freq terms into a single int[]:
while ((docID = docsEnum2.NextDoc()) != DocsEnum.NO_MORE_DOCS)
{
scratch.Add(docID);
if (hasFreq)
{
int freq = docsEnum2.Freq();
scratch.Add(freq);
if (hasPos)
{
for (int pos = 0; pos < freq; pos++)
{
scratch.Add(docsAndPositionsEnum.NextPosition());
if (hasOffsets_Renamed)
{
scratch.Add(docsAndPositionsEnum.StartOffset());
scratch.Add(docsAndPositionsEnum.EndOffset());
}
if (hasPayloads_Renamed)
{
BytesRef payload = docsAndPositionsEnum.Payload;
if (payload != null)
{
scratch.Add(payload.Length);
ros.WriteBytes(payload.Bytes, payload.Offset, payload.Length);
}
else
{
scratch.Add(0);
}
}
}
}
}
}
byte[] payloads;
if (hasPayloads_Renamed)
{
ros.Flush();
payloads = new byte[(int) ros.Length];
ros.WriteTo(payloads, 0);
}
else
{
payloads = null;
}
int[] postings = scratch.Get();
ent = new LowFreqTerm(postings, payloads, docFreq, (int) totalTermFreq);
}
else
{
var docs = new int[docFreq];
int[] freqs;
int[][] positions;
byte[][][] payloads;
if (hasFreq)
{
freqs = new int[docFreq];
if (hasPos)
{
positions = new int[docFreq][];
if (hasPayloads_Renamed)
{
payloads = new byte[docFreq][][];
}
else
{
payloads = null;
}
}
else
{
positions = null;
payloads = null;
}
}
else
{
freqs = null;
positions = null;
payloads = null;
}
// Use separate int[] for the postings for high-freq
// terms:
int upto = 0;
while ((docID = docsEnum2.NextDoc()) != DocsEnum.NO_MORE_DOCS)
{
docs[upto] = docID;
if (hasFreq)
{
int freq = docsEnum2.Freq();
freqs[upto] = freq;
if (hasPos)
{
int mult;
if (hasOffsets_Renamed)
{
mult = 3;
}
else
{
mult = 1;
}
if (hasPayloads_Renamed)
{
payloads[upto] = new byte[freq][];
}
positions[upto] = new int[mult*freq];
int posUpto = 0;
for (int pos = 0; pos < freq; pos++)
{
positions[upto][posUpto] = docsAndPositionsEnum.NextPosition();
if (hasPayloads_Renamed)
{
BytesRef payload = docsAndPositionsEnum.Payload;
if (payload != null)
{
var payloadBytes = new byte[payload.Length];
Array.Copy(payload.Bytes, payload.Offset, payloadBytes, 0,
payload.Length);
payloads[upto][pos] = payloadBytes;
}
}
posUpto++;
if (hasOffsets_Renamed)
{
positions[upto][posUpto++] = docsAndPositionsEnum.StartOffset();
positions[upto][posUpto++] = docsAndPositionsEnum.EndOffset();
}
}
}
}
upto++;
}
Debug.Assert(upto == docFreq);
ent = new HighFreqTerm(docs, freqs, positions, payloads, totalTermFreq);
}
terms[count] = ent;
SetSkips(count, termBytes);
count++;
}
// End sentinel:
termOffsets[count] = termOffset;
FinishSkips();
//System.out.println(skipCount + " skips: " + field);
this.termBytes = new byte[termOffset];
Array.Copy(termBytes, 0, this.termBytes, 0, termOffset);
// Pack skips:
this.skips = new int[skipCount];
this.skipOffsets = new int[1 + numTerms];
int skipOffset = 0;
for (int i = 0; i < numTerms; i++)
{
int[] termSkips = terms[i].skips;
skipOffsets[i] = skipOffset;
if (termSkips != null)
{
Array.Copy(termSkips, 0, skips, skipOffset, termSkips.Length);
skipOffset += termSkips.Length;
terms[i].skips = null;
}
}
this.skipOffsets[numTerms] = skipOffset;
Debug.Assert(skipOffset == skipCount);
}
private void FlushBlock()
{
// First pass: compute common prefix for all terms
// in the block, against term before first term in
// this block:
int commonPrefix = SharedPrefix(_lastPrevTerm, _pendingTerms[0].Term);
for (int termCount = 1; termCount < _pendingCount; termCount++)
{
commonPrefix = Math.Min(commonPrefix,
SharedPrefix(_lastPrevTerm,
_pendingTerms[termCount].Term));
}
_btw._output.WriteVInt(_pendingCount);
_btw._output.WriteVInt(commonPrefix);
// 2nd pass: write suffixes, as separate byte[] blob
for (var termCount = 0; termCount < _pendingCount; termCount++)
{
var suffix = _pendingTerms[termCount].Term.Length - commonPrefix;
// TODO: cutover to better intblock codec, instead
// of interleaving here:
_bytesWriter.WriteVInt(suffix);
_bytesWriter.WriteBytes(_pendingTerms[termCount].Term.Bytes, commonPrefix, suffix);
}
_btw._output.WriteVInt((int)_bytesWriter.FilePointer);
_bytesWriter.WriteTo(_btw._output);
_bytesWriter.Reset();
// 3rd pass: write the freqs as byte[] blob
// TODO: cutover to better intblock codec. simple64?
// write prefix, suffix first:
for (int termCount = 0; termCount < _pendingCount; termCount++)
{
BlockTermState state = _pendingTerms[termCount].State;
Debug.Assert(state != null);
_bytesWriter.WriteVInt(state.DocFreq);
if (_fieldInfo.FieldIndexOptions != FieldInfo.IndexOptions.DOCS_ONLY)
{
_bytesWriter.WriteVLong(state.TotalTermFreq - state.DocFreq);
}
}
_btw._output.WriteVInt((int)_bytesWriter.FilePointer);
_bytesWriter.WriteTo(_btw._output);
_bytesWriter.Reset();
// 4th pass: write the metadata
var longs = new long[_longsSize];
bool absolute = true;
for (int termCount = 0; termCount < _pendingCount; termCount++)
{
BlockTermState state = _pendingTerms[termCount].State;
_postingsWriter.EncodeTerm(longs, _bufferWriter, _fieldInfo, state, absolute);
for (int i = 0; i < _longsSize; i++)
{
_bytesWriter.WriteVLong(longs[i]);
}
_bufferWriter.WriteTo(_bytesWriter);
_bufferWriter.Reset();
absolute = false;
}
_btw._output.WriteVInt((int)_bytesWriter.FilePointer);
_bytesWriter.WriteTo(_btw._output);
_bytesWriter.Reset();
_lastPrevTerm.CopyBytes(_pendingTerms[_pendingCount - 1].Term);
_pendingCount = 0;
}
