package org.apache.lucene.codecs.block;
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
import java.io.IOException;
import org.apache.lucene.codecs.BlockTreeTermsReader;
import org.apache.lucene.codecs.BlockTreeTermsWriter;
import org.apache.lucene.codecs.CodecUtil;
import org.apache.lucene.codecs.FieldsConsumer;
import org.apache.lucene.codecs.FieldsProducer;
import org.apache.lucene.codecs.MultiLevelSkipListWriter;
import org.apache.lucene.codecs.PostingsFormat;
import org.apache.lucene.codecs.PostingsReaderBase;
import org.apache.lucene.codecs.PostingsWriterBase;
import org.apache.lucene.index.DocsEnum;
import org.apache.lucene.index.FieldInfo.IndexOptions;
import org.apache.lucene.index.SegmentReadState;
import org.apache.lucene.index.SegmentWriteState;
import org.apache.lucene.store.DataOutput;
import org.apache.lucene.util.IOUtils;
import org.apache.lucene.util.packed.PackedInts;
/**
* Block postings format, which encodes postings in packed int blocks
* for faster decode.
*
* <p><b>NOTE</b>: this format is still experimental and
* subject to change without backwards compatibility.
*
* <p>
* Basic idea:
* <ul>
* <li>
* <b>Packed Block and VInt Block</b>:
* <p>In packed block, integers are encoded with the same bit width ({@link PackedInts packed format}),
* the block size (i.e. number of integers inside block) is fixed. </p>
* <p>In VInt block, integers are encoded as {@link DataOutput#writeVInt VInt},
* the block size is variable.</p>
* </li>
*
* <li>
* <b>Block structure</b>:
* <p>When the postings is long enough, BlockPostingsFormat will try to encode most integer data
* as packed block.</p>
* <p>Take a term with 259 documents as example, the first 256 document ids are encoded as two packed
* blocks, while the remaining 3 as one VInt block. </p>
* <p>Different kinds of data are always encoded separately into different packed blocks, but may
* possible be encoded into a same VInt block. </p>
* <p>This strategy is applied to pairs:
* <document number, frequency>,
* <position, payload length>,
* <position, offset start, offset length>, and
* <position, payload length, offsetstart, offset length>.</p>
* </li>
*
* <li>
* <b>Skipper setting</b>:
* <p>The structure of skip table is quite similar to Lucene40PostingsFormat. Skip interval is the
* same as block size, and each skip entry points to the beginning of each block. However, for
* the first block, skip data is omitted.</p>
* </li>
*
* <li>
* <b>Positions, Payloads, and Offsets</b>:
* <p>A position is an integer indicating where the term occurs at within one document.
* A payload is a blob of metadata associated with current position.
* An offset is a pair of integers indicating the tokenized start/end offsets for given term
* in current position. </p>
* <p>When payloads and offsets are not omitted, numPositions==numPayloads==numOffsets (assuming a
* null payload contributes one count). As mentioned in block structure, it is possible to encode
* these three either combined or separately.
* <p>For all the cases, payloads and offsets are stored together. When encoded as packed block,
* position data is separated out as .pos, while payloads and offsets are encoded in .pay (payload
* metadata will also be stored directly in .pay). When encoded as VInt block, all these three are
* stored in .pos (so as payload metadata).</p>
* <p>With this strategy, the majority of payload and offset data will be outside .pos file.
* So for queries that require only position data, running on a full index with payloads and offsets,
* this reduces disk pre-fetches.</p>
* </li>
* </ul>
* </p>
*
* <p>
* Files and detailed format:
* <ul>
* <li><tt>.tim</tt>: <a href="#Termdictionary">Term Dictionary</a></li>
* <li><tt>.tip</tt>: <a href="#Termindex">Term Index</a></li>
* <li><tt>.doc</tt>: <a href="#Frequencies">Frequencies and Skip Data</a></li>
* <li><tt>.pos</tt>: <a href="#Positions">Positions</a></li>
* <li><tt>.pay</tt>: <a href="#Payloads">Payloads and Offsets</a></li>
* </ul>
* </p>
*
* <a name="Termdictionary" id="Termdictionary"></a>
* <dl>
* <dd>
* <b>Term Dictionary</b>
*
* <p>The .tim file format is quite similar to Lucene40PostingsFormat,
* with minor difference in MetadataBlock</p>
*
* <ul>
* <!-- TODO: expand on this, its not really correct and doesnt explain sub-blocks etc -->
* <li>TermDictionary(.tim) --> Header, DirOffset, PostingsHeader, PackedBlockSize,
* <Block><sup>NumBlocks</sup>, FieldSummary</li>
* <li>Block --> SuffixBlock, StatsBlock, MetadataBlock</li>
* <li>SuffixBlock --> EntryCount, SuffixLength, {@link DataOutput#writeByte byte}<sup>SuffixLength</sup></li>
* <li>StatsBlock --> StatsLength, <DocFreq, TotalTermFreq><sup>EntryCount</sup></li>
* <li>MetadataBlock --> MetaLength, <DocFPDelta,
* <PosFPDelta, PosVIntBlockFPDelta?, PayFPDelta?>?,
* SkipFPDelta?><sup>EntryCount</sup></li>
* <li>FieldSummary --> NumFields, <FieldNumber, NumTerms, RootCodeLength,
* {@link DataOutput#writeByte byte}<sup>RootCodeLength</sup>, SumDocFreq, DocCount>
* <sup>NumFields</sup></li>
* <li>Header, PostingsHeader --> {@link CodecUtil#writeHeader CodecHeader}</li>
* <li>DirOffset --> {@link DataOutput#writeLong Uint64}</li>
* <li>PackedBlockSize, EntryCount, SuffixLength, StatsLength, DocFreq, MetaLength,
* PosVIntBlockFPDelta, SkipFPDelta, NumFields, FieldNumber, RootCodeLength, DocCount -->
* {@link DataOutput#writeVInt VInt}</li>
* <li>TotalTermFreq, DocFPDelta, PosFPDelta, PayFPDelta, NumTerms, SumTotalTermFreq, SumDocFreq -->
* {@link DataOutput#writeVLong VLong}</li>
* </ul>
* <p>Notes:</p>
* <ul>
* <li>Here explains MetadataBlock only, other fields are mentioned in
* <a href="{@docRoot}/../core/org/apache/lucene/codecs/lucene40/Lucene40PostingsFormat.html#Termdictionary">Lucene40PostingsFormat:TermDictionary</a>
* </li>
* <li>PackedBlockSize is the fixed block size for packed blocks. In packed block, bit width is
* determined by the largest integer. Smaller block size result in smaller variance among width
* of integers hence smaller indexes. Larger block size result in more efficient bulk i/o hence
* better acceleration. This value should always be a multiple of 64, currently fixed as 128 as
* a tradeoff. It is also the skip interval used to accelerate {@link DocsEnum#advance(int)}.
* <li>DocFPDelta determines the position of this term's TermFreqs within the .doc file.
* In particular, it is the difference of file offset between this term's
* data and previous term's data (or zero, for the first term in the block).On disk it is
* stored as the difference from previous value in sequence. </li>
* <li>PosFPDelta determines the position of this term's TermPositions within the .pos file.
* While PayFPDelta determines the position of this term's <TermPayloads, TermOffsets?> within
* the .pay file. Similar to DocFPDelta, it is the difference between two file positions (or
* neglected, for fields that omit payloads and offsets).</li>
* <li>PosVIntBlockFPDelta determines the position of this term's last TermPosition in last pos packed
* block within the .pos file. It is synonym for PayVIntBlockFPDelta or OffsetVIntBlockFPDelta.
* This is actually used to indicate whether it is necessary to load following
* payloads and offsets from .pos instead of .pay. Every time a new block of positions are to be
* loaded, the PostingsReader will use this value to check whether current block is packed format
* or VInt. When packed format, payloads and offsets are fetched from .pay, otherwise from .pos.
* (this value is neglected when total number of positions i.e. totalTermFreq is less or equal
* to PackedBlockSize).
* <li>SkipFPDelta determines the position of this term's SkipData within the .doc
* file. In particular, it is the length of the TermFreq data.
* SkipDelta is only stored if DocFreq is not smaller than SkipMinimum
* (i.e. 8 in BlockPostingsFormat).</li>
* </ul>
* </dd>
* </dl>
*
* <a name="Termindex" id="Termindex"></a>
* <dl>
* <dd>
* <b>Term Index</b>
* <p>The .tim file format is mentioned in
* <a href="{@docRoot}/../core/org/apache/lucene/codecs/lucene40/Lucene40PostingsFormat.html#Termindex">Lucene40PostingsFormat:TermIndex</a>
* </dd>
* </dl>
*
*
* <a name="Frequencies" id="Frequencies"></a>
* <dl>
* <dd>
* <b>Frequencies and Skip Data</b>
*
* <p>The .doc file contains the lists of documents which contain each term, along
* with the frequency of the term in that document (except when frequencies are
* omitted: {@link IndexOptions#DOCS_ONLY}). It also saves skip data to the beginning of
* each packed or VInt block, when the length of document list is larger than packed block size.</p>
*
* <ul>
* <li>docFile(.doc) --> Header, <TermFreqs, SkipData?><sup>TermCount</sup></li>
* <li>Header --> {@link CodecUtil#writeHeader CodecHeader}</li>
* <li>TermFreqs --> <PackedBlock> <sup>PackedDocBlockNum</sup>,
* VIntBlock? </li>
* <li>PackedBlock --> PackedDocDeltaBlock, PackedFreqBlock?
* <li>VIntBlock --> <DocDelta[, Freq?]><sup>DocFreq-PackedBlockSize*PackedDocBlockNum</sup>
* <li>SkipData --> <<SkipLevelLength, SkipLevel>
* <sup>NumSkipLevels-1</sup>, SkipLevel>, SkipDatum?</li>
* <li>SkipLevel --> <SkipDatum> <sup>TrimmedDocFreq/(PackedBlockSize^(Level + 1))</sup></li>
* <li>SkipDatum --> DocSkip, DocFPSkip, <PosFPSkip, PosBlockOffset, PayLength?,
* OffsetStart?, PayFPSkip?>?, SkipChildLevelPointer?</li>
* <li>PackedDocDeltaBlock, PackedFreqBlock --> {@link PackedInts PackedInts}</li>
* <li>DocDelta, Freq, DocSkip, DocFPSkip, PosFPSkip, PosBlockOffset, PayLength, OffsetStart, PayFPSkip
* -->
* {@link DataOutput#writeVInt VInt}</li>
* <li>SkipChildLevelPointer --> {@link DataOutput#writeVLong VLong}</li>
* </ul>
* <p>Notes:</p>
* <ul>
* <li>PackedDocDeltaBlock is theoretically generated from two steps:
* <ol>
* <li>Calculate the difference between each document number and previous one,
* and get a d-gaps list (for the first document, use absolute value); </li>
* <li>For those d-gaps from first one to PackedDocBlockNum*PackedBlockSize<sup>th</sup>,
* separately encode as packed blocks.</li>
* </ol>
* If frequencies are not omitted, PackedFreqBlock will be generated without d-gap step.
* </li>
* <li>VIntBlock stores remaining d-gaps (along with frequencies when possible) with a format
* mentioned in
* <a href="{@docRoot}/../core/org/apache/lucene/codecs/lucene40/Lucene40PostingsFormat.html#Frequencies">Lucene40PostingsFormat:Frequencies</a>
* </li>
* <li>PackedDocBlockNum is the number of packed blocks for current term's docids or frequencies.
* In particular, PackedDocBlockNum = floor(DocFreq/PackedBlockSize) </li>
* <li>TrimmedDocFreq = DocFreq % PackedBlockSize == 0 ? DocFreq - 1 : DocFreq.
* We use this trick since the definition of skip entry is a little different from base interface.
* In {@link MultiLevelSkipListWriter}, skip data is assumed to be saved for
* skipInterval<sup>th</sup>, 2*skipInterval<sup>th</sup> ... posting in the list. However,
* in BlockPostingsFormat, the skip data is saved for skipInterval+1<sup>th</sup>,
* 2*skipInterval+1<sup>th</sup> ... posting (skipInterval==PackedBlockSize in this case).
* When DocFreq is multiple of PackedBlockSize, MultiLevelSkipListWriter will expect one
* more skip data than BlockSkipWriter. </li>
* <li>SkipDatum is the metadata of one skip entry.
* For the first block (no matter packed or VInt), it is omitted.</li>
* <li>DocSkip records the document number of every PackedBlockSize<sup>th</sup> document number in
* the postings (i.e. last document number in each packed block). On disk it is stored as the
* difference from previous value in the sequence. </li>
* <li>DocFPSkip records the file offsets of each block (excluding )posting at
* PackedBlockSize+1<sup>th</sup>, 2*PackedBlockSize+1<sup>th</sup> ... , in DocFile.
* The file offsets are relative to the start of current term's TermFreqs.
* On disk it is also stored as the difference from previous SkipDatum in the sequence.</li>
* <li>Since positions and payloads are also block encoded, the skip should skip to related block first,
* then fetch the values according to in-block offset. PosFPSkip and PayFPSkip record the file
* offsets of related block in .pos and .pay, respectively. While PosBlockOffset indicates
* which value to fetch inside the related block (PayBlockOffset is unnecessary since it is always
* equal to PosBlockOffset). Same as DocFPSkip, the file offsets are relative to the start of
* current term's TermFreqs, and stored as a difference sequence.</li>
* <li>PayLength indicates the length of last payload.</li>
* <li>OffsetStart indicates the first value of last offset pair.</li>
* </ul>
* </dd>
* </dl>
*
* <a name="Positions" id="Positions"></a>
* <dl>
* <dd>
* <b>Positions</b>
* <p>The .pos file contains the lists of positions that each term occurs at within documents. It also
* sometimes stores part of payloads and offsets for speedup.</p>
* <ul>
* <li>PosFile(.pos) --> Header, <TermPositions> <sup>TermCount</sup></li>
* <li>Header --> {@link CodecUtil#writeHeader CodecHeader}</li>
* <li>TermPositions --> <PackedPosDeltaBlock> <sup>PackedPosBlockNum</sup>,
* VIntBlock? </li>
* <li>VIntBlock --> PosVIntCount, <PosDelta[, PayLength?], PayData?,
* OffsetStartDelta?, OffsetLength?><sup>PosVIntCount</sup>
* <li>PackedPosDeltaBlock --> {@link PackedInts PackedInts}</li>
* <li>PosVIntCount, PosDelta, OffsetStartDelta, OffsetLength -->
* {@link DataOutput#writeVInt VInt}</li>
* <li>PayData --> {@link DataOutput#writeByte byte}<sup>PayLength</sup></li>
* </ul>
* <p>Notes:</p>
* <ul>
* <li>TermPositions are order by term (terms are implicit, from the term dictionary), and position
* values for each term document pair are incremental, and ordered by document number.</li>
* <li>PackedPosBlockNum is the number of packed blocks for current term's positions, payloads or offsets.
* In particular, PackedPosBlockNum = floor(totalTermFreq/PackedBlockSize) </li>
* <li>PosVIntCount is the number of positions encoded as VInt format. In particular,
* PosVIntCount = totalTermFreq - PackedPosBlockNum*PackedBlockSize</li>
* <li>The procedure how PackedPosDeltaBlock is generated is the same as PackedDocDeltaBlock
* in chapter <a href="#Frequencies">Frequencies and Skip Data</a>.</li>
* <li>PosDelta is the same as the format mentioned in
* <a href="{@docRoot}/../core/org/apache/lucene/codecs/lucene40/Lucene40PostingsFormat.html#Positions">Lucene40PostingsFormat:Positions</a>
* </li>
* <li>OffsetStartDelta is the difference between this position's startOffset from the previous
* occurrence (or zero, if this is the first occurrence in this document).</li>
* <li>OffsetLength indicates the length of the current offset (endOffset-startOffset).</li>
* <li>PayloadData is the blob of metadata associated with current position.</li>
* </ul>
* </dd>
* </dl>
*
* <a name="Payloads" id="Payloads"></a>
* <dl>
* <dd>
* <b>Payloads and Offsets</b>
* <p>The .pay file will store payloads and offsets associated with certain term-document positions.
* Some payloads and offsets will be separated out into .pos file, for speedup reason.</p>
* <ul>
* <li>PayFile(.pay): --> Header, <TermPayloads, TermOffsets?> <sup>TermCount</sup></li>
* <li>Header --> {@link CodecUtil#writeHeader CodecHeader}</li>
* <li>TermPayloads --> <PackedPayLengthBlock, SumPayLength, PayData> <sup>PackedPayBlockNum</sup>
* <li>TermOffsets --> <PackedOffsetStartDeltaBlock, PackedOffsetLengthBlock> <sup>PackedPayBlockNum</sup>
* <li>PackedPayLengthBlock, PackedOffsetStartDeltaBlock, PackedOffsetLengthBlock --> {@link PackedInts PackedInts}</li>
* <li>SumPayLength --> {@link DataOutput#writeVInt VInt}</li>
* <li>PayData --> {@link DataOutput#writeByte byte}<sup>SumPayLength</sup></li>
* </ul>
* <p>Notes:</p>
* <ul>
* <li>The order of TermPayloads/TermOffsets will be the same as TermPositions, note that part of
* payload/offsets are stored in .pos.</li>
* <li>The procedure how PackedPayLengthBlock and PackedOffsetLengthBlock are generated is the
* same as PackedFreqBlock in chapter <a href="#Frequencies">Frequencies and Skip Data</a>.
* While PackedStartDeltaBlock follows a same procedure as PackedDocDeltaBlock.</li>
* <li>PackedPayBlockNum is always equal to PackedPosBlockNum, for the same term. It is also synonym
* for PackedOffsetBlockNum.</li>
* <li>SumPayLength is the total length of payloads written within one block, should be the sum
* of PayLengths in one packed block.</li>
* <li>PayLength in PackedPayLengthBlock is the length of each payload, associated with current
* position.</li>
* </ul>
* </dd>
* </dl>
* </p>
*
* @lucene.experimental
*/
public final class BlockPostingsFormat extends PostingsFormat {
/**
* Filename extension for document number, frequencies, and skip data.
* See chapter: <a href="#Frequencies">Frequencies and Skip Data</a>
*/
public static final String DOC_EXTENSION = "doc";
/**
* Filename extension for positions.
* See chapter: <a href="#Positions">Positions</a>
*/
public static final String POS_EXTENSION = "pos";
/**
* Filename extension for payloads and offsets.
* See chapter: <a href="#Payloads">Payloads and Offsets</a>
*/
public static final String PAY_EXTENSION = "pay";
private final int minTermBlockSize;
private final int maxTermBlockSize;
/**
* Fixed packed block size, number of integers encoded in
* a single packed block.
*/
// NOTE: must be multiple of 64 because of PackedInts long-aligned encoding/decoding
public final static int BLOCK_SIZE = 128;
public BlockPostingsFormat() {
this(BlockTreeTermsWriter.DEFAULT_MIN_BLOCK_SIZE, BlockTreeTermsWriter.DEFAULT_MAX_BLOCK_SIZE);
}
public BlockPostingsFormat(int minTermBlockSize, int maxTermBlockSize) {
super("Block");
this.minTermBlockSize = minTermBlockSize;
assert minTermBlockSize > 1;
this.maxTermBlockSize = maxTermBlockSize;
assert minTermBlockSize <= maxTermBlockSize;
}
@Override
public String toString() {
return getName() + "(blocksize=" + BLOCK_SIZE + ")";
}
@Override
public FieldsConsumer fieldsConsumer(SegmentWriteState state) throws IOException {
PostingsWriterBase postingsWriter = new BlockPostingsWriter(state);
boolean success = false;
try {
FieldsConsumer ret = new BlockTreeTermsWriter(state,
postingsWriter,
minTermBlockSize,
maxTermBlockSize);
success = true;
return ret;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(postingsWriter);
}
}
}
@Override
public FieldsProducer fieldsProducer(SegmentReadState state) throws IOException {
PostingsReaderBase postingsReader = new BlockPostingsReader(state.dir,
state.fieldInfos,
state.segmentInfo,
state.context,
state.segmentSuffix);
boolean success = false;
try {
FieldsProducer ret = new BlockTreeTermsReader(state.dir,
state.fieldInfos,
state.segmentInfo,
postingsReader,
state.context,
state.segmentSuffix,
state.termsIndexDivisor);
success = true;
return ret;
} finally {
if (!success) {
IOUtils.closeWhileHandlingException(postingsReader);
}
}
}
}