mirror of https://github.com/apache/lucene.git
LUCENE-5675: initial fork of BT with versioning added
git-svn-id: https://svn.apache.org/repos/asf/lucene/dev/branches/lucene5675@1595064 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Michael McCandless
parent
1ea0ce4062
commit
491dc6a716
|
|
@ -17,8 +17,6 @@ package org.apache.lucene.codecs.idversion;
|
|||
* limitations under the License.
|
||||
*/
|
||||
|
||||
import static org.apache.lucene.codecs.idversion.IDVersionPostingsWriter.IDVersionTermState;
|
||||
|
||||
import java.io.IOException;
|
||||
|
||||
import org.apache.lucene.codecs.BlockTermState;
|
||||
|
|
|
|||
|
|
@ -43,26 +43,6 @@ public final class IDVersionPostingsWriter extends PushPostingsWriterBase {
|
|||
private int lastPosition;
|
||||
private long lastVersion;
|
||||
|
||||
final static class IDVersionTermState extends BlockTermState {
|
||||
long idVersion;
|
||||
int docID;
|
||||
|
||||
@Override
|
||||
public IDVersionTermState clone() {
|
||||
IDVersionTermState other = new IDVersionTermState();
|
||||
other.copyFrom(this);
|
||||
return other;
|
||||
}
|
||||
|
||||
@Override
|
||||
public void copyFrom(TermState _other) {
|
||||
super.copyFrom(_other);
|
||||
IDVersionTermState other = (IDVersionTermState) _other;
|
||||
idVersion = other.idVersion;
|
||||
docID = other.docID;
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public IDVersionTermState newTermState() {
|
||||
return new IDVersionTermState();
|
||||
|
|
@ -144,8 +124,8 @@ public final class IDVersionPostingsWriter extends PushPostingsWriterBase {
|
|||
@Override
|
||||
public void encodeTerm(long[] longs, DataOutput out, FieldInfo fieldInfo, BlockTermState _state, boolean absolute) throws IOException {
|
||||
IDVersionTermState state = (IDVersionTermState) _state;
|
||||
// nocommit must send version up to FST somehow ...
|
||||
out.writeVInt(state.docID);
|
||||
out.writeVLong(state.idVersion);
|
||||
}
|
||||
|
||||
@Override
|
||||
|
|
|
|||
|
|
@ -0,0 +1,971 @@
|
|||
package org.apache.lucene.codecs.idversion;
|
||||
|
||||
/*
|
||||
* 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 java.io.PrintStream;
|
||||
|
||||
import org.apache.lucene.codecs.BlockTermState;
|
||||
import org.apache.lucene.index.DocsAndPositionsEnum;
|
||||
import org.apache.lucene.index.DocsEnum;
|
||||
import org.apache.lucene.index.FieldInfo.IndexOptions;
|
||||
import org.apache.lucene.index.TermState;
|
||||
import org.apache.lucene.index.TermsEnum;
|
||||
import org.apache.lucene.store.ByteArrayDataInput;
|
||||
import org.apache.lucene.store.IndexInput;
|
||||
import org.apache.lucene.util.ArrayUtil;
|
||||
import org.apache.lucene.util.Bits;
|
||||
import org.apache.lucene.util.BytesRef;
|
||||
import org.apache.lucene.util.RamUsageEstimator;
|
||||
import org.apache.lucene.util.fst.ByteSequenceOutputs;
|
||||
import org.apache.lucene.util.fst.FST;
|
||||
import org.apache.lucene.util.fst.Outputs;
|
||||
import org.apache.lucene.util.fst.PairOutputs.Pair;
|
||||
import org.apache.lucene.util.fst.PairOutputs;
|
||||
import org.apache.lucene.util.fst.Util;
|
||||
|
||||
/** Iterates through terms in this field */
|
||||
final class IDVersionSegmentTermsEnum extends TermsEnum {
|
||||
|
||||
final static Outputs<Pair<BytesRef,Long>> fstOutputs = VersionBlockTreeTermsWriter.getFSTOutputs();
|
||||
final static Pair<BytesRef,Long> NO_OUTPUT = fstOutputs.getNoOutput();
|
||||
|
||||
// Lazy init:
|
||||
IndexInput in;
|
||||
|
||||
private IDVersionSegmentTermsEnumFrame[] stack;
|
||||
private final IDVersionSegmentTermsEnumFrame staticFrame;
|
||||
IDVersionSegmentTermsEnumFrame currentFrame;
|
||||
boolean termExists;
|
||||
final VersionFieldReader fr;
|
||||
|
||||
// nocommit make this public "for casting" and add a getVersion method?
|
||||
|
||||
private int targetBeforeCurrentLength;
|
||||
|
||||
private final ByteArrayDataInput scratchReader = new ByteArrayDataInput();
|
||||
|
||||
// What prefix of the current term was present in the index:
|
||||
private int validIndexPrefix;
|
||||
|
||||
// assert only:
|
||||
private boolean eof;
|
||||
|
||||
final BytesRef term = new BytesRef();
|
||||
private final FST.BytesReader fstReader;
|
||||
|
||||
@SuppressWarnings({"rawtypes","unchecked"}) private FST.Arc<Pair<BytesRef,Long>>[] arcs =
|
||||
new FST.Arc[1];
|
||||
|
||||
public IDVersionSegmentTermsEnum(VersionFieldReader fr) throws IOException {
|
||||
this.fr = fr;
|
||||
|
||||
//if (DEBUG) System.out.println("BTTR.init seg=" + segment);
|
||||
stack = new IDVersionSegmentTermsEnumFrame[0];
|
||||
|
||||
// Used to hold seek by TermState, or cached seek
|
||||
staticFrame = new IDVersionSegmentTermsEnumFrame(this, -1);
|
||||
|
||||
if (fr.index == null) {
|
||||
fstReader = null;
|
||||
} else {
|
||||
fstReader = fr.index.getBytesReader();
|
||||
}
|
||||
|
||||
// Init w/ root block; don't use index since it may
|
||||
// not (and need not) have been loaded
|
||||
for(int arcIdx=0;arcIdx<arcs.length;arcIdx++) {
|
||||
arcs[arcIdx] = new FST.Arc<>();
|
||||
}
|
||||
|
||||
currentFrame = staticFrame;
|
||||
final FST.Arc<Pair<BytesRef,Long>> arc;
|
||||
if (fr.index != null) {
|
||||
arc = fr.index.getFirstArc(arcs[0]);
|
||||
// Empty string prefix must have an output in the index!
|
||||
assert arc.isFinal();
|
||||
} else {
|
||||
arc = null;
|
||||
}
|
||||
currentFrame = staticFrame;
|
||||
//currentFrame = pushFrame(arc, rootCode, 0);
|
||||
//currentFrame.loadBlock();
|
||||
validIndexPrefix = 0;
|
||||
// if (DEBUG) {
|
||||
// System.out.println("init frame state " + currentFrame.ord);
|
||||
// printSeekState();
|
||||
// }
|
||||
|
||||
//System.out.println();
|
||||
// computeBlockStats().print(System.out);
|
||||
}
|
||||
|
||||
// Not private to avoid synthetic access$NNN methods
|
||||
void initIndexInput() {
|
||||
if (this.in == null) {
|
||||
this.in = fr.parent.in.clone();
|
||||
}
|
||||
}
|
||||
|
||||
private IDVersionSegmentTermsEnumFrame getFrame(int ord) throws IOException {
|
||||
if (ord >= stack.length) {
|
||||
final IDVersionSegmentTermsEnumFrame[] next = new IDVersionSegmentTermsEnumFrame[ArrayUtil.oversize(1+ord, RamUsageEstimator.NUM_BYTES_OBJECT_REF)];
|
||||
System.arraycopy(stack, 0, next, 0, stack.length);
|
||||
for(int stackOrd=stack.length;stackOrd<next.length;stackOrd++) {
|
||||
next[stackOrd] = new IDVersionSegmentTermsEnumFrame(this, stackOrd);
|
||||
}
|
||||
stack = next;
|
||||
}
|
||||
assert stack[ord].ord == ord;
|
||||
return stack[ord];
|
||||
}
|
||||
|
||||
private FST.Arc<Pair<BytesRef,Long>> getArc(int ord) {
|
||||
if (ord >= arcs.length) {
|
||||
@SuppressWarnings({"rawtypes","unchecked"}) final FST.Arc<Pair<BytesRef,Long>>[] next =
|
||||
new FST.Arc[ArrayUtil.oversize(1+ord, RamUsageEstimator.NUM_BYTES_OBJECT_REF)];
|
||||
System.arraycopy(arcs, 0, next, 0, arcs.length);
|
||||
for(int arcOrd=arcs.length;arcOrd<next.length;arcOrd++) {
|
||||
next[arcOrd] = new FST.Arc<>();
|
||||
}
|
||||
arcs = next;
|
||||
}
|
||||
return arcs[ord];
|
||||
}
|
||||
|
||||
// Pushes a frame we seek'd to
|
||||
IDVersionSegmentTermsEnumFrame pushFrame(FST.Arc<Pair<BytesRef,Long>> arc, Pair<BytesRef,Long> frameData, int length) throws IOException {
|
||||
scratchReader.reset(frameData.output1.bytes, frameData.output1.offset, frameData.output1.length);
|
||||
final long code = scratchReader.readVLong();
|
||||
final long fpSeek = code >>> VersionBlockTreeTermsWriter.OUTPUT_FLAGS_NUM_BITS;
|
||||
final IDVersionSegmentTermsEnumFrame f = getFrame(1+currentFrame.ord);
|
||||
f.hasTerms = (code & VersionBlockTreeTermsWriter.OUTPUT_FLAG_HAS_TERMS) != 0;
|
||||
f.hasTermsOrig = f.hasTerms;
|
||||
f.isFloor = (code & VersionBlockTreeTermsWriter.OUTPUT_FLAG_IS_FLOOR) != 0;
|
||||
if (f.isFloor) {
|
||||
f.setFloorData(scratchReader, frameData.output1);
|
||||
}
|
||||
pushFrame(arc, fpSeek, length);
|
||||
|
||||
return f;
|
||||
}
|
||||
|
||||
// Pushes next'd frame or seek'd frame; we later
|
||||
// lazy-load the frame only when needed
|
||||
IDVersionSegmentTermsEnumFrame pushFrame(FST.Arc<Pair<BytesRef,Long>> arc, long fp, int length) throws IOException {
|
||||
final IDVersionSegmentTermsEnumFrame f = getFrame(1+currentFrame.ord);
|
||||
f.arc = arc;
|
||||
if (f.fpOrig == fp && f.nextEnt != -1) {
|
||||
//if (DEBUG) System.out.println(" push reused frame ord=" + f.ord + " fp=" + f.fp + " isFloor?=" + f.isFloor + " hasTerms=" + f.hasTerms + " pref=" + term + " nextEnt=" + f.nextEnt + " targetBeforeCurrentLength=" + targetBeforeCurrentLength + " term.length=" + term.length + " vs prefix=" + f.prefix);
|
||||
if (f.prefix > targetBeforeCurrentLength) {
|
||||
f.rewind();
|
||||
} else {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" skip rewind!");
|
||||
// }
|
||||
}
|
||||
assert length == f.prefix;
|
||||
} else {
|
||||
f.nextEnt = -1;
|
||||
f.prefix = length;
|
||||
f.state.termBlockOrd = 0;
|
||||
f.fpOrig = f.fp = fp;
|
||||
f.lastSubFP = -1;
|
||||
// if (DEBUG) {
|
||||
// final int sav = term.length;
|
||||
// term.length = length;
|
||||
// System.out.println(" push new frame ord=" + f.ord + " fp=" + f.fp + " hasTerms=" + f.hasTerms + " isFloor=" + f.isFloor + " pref=" + brToString(term));
|
||||
// term.length = sav;
|
||||
// }
|
||||
}
|
||||
|
||||
return f;
|
||||
}
|
||||
|
||||
// asserts only
|
||||
private boolean clearEOF() {
|
||||
eof = false;
|
||||
return true;
|
||||
}
|
||||
|
||||
// asserts only
|
||||
private boolean setEOF() {
|
||||
eof = true;
|
||||
return true;
|
||||
}
|
||||
|
||||
// nocommit we need a seekExact(BytesRef target, long minVersion) API?
|
||||
|
||||
@Override
|
||||
public boolean seekExact(final BytesRef target) throws IOException {
|
||||
return seekExact(target, 0);
|
||||
}
|
||||
|
||||
/** Returns false if the term deos not exist, or it exists but its version is < minIDVersion. */
|
||||
public boolean seekExact(final BytesRef target, long minIDVersion) throws IOException {
|
||||
|
||||
if (fr.index == null) {
|
||||
throw new IllegalStateException("terms index was not loaded");
|
||||
}
|
||||
|
||||
if (term.bytes.length <= target.length) {
|
||||
term.bytes = ArrayUtil.grow(term.bytes, 1+target.length);
|
||||
}
|
||||
|
||||
assert clearEOF();
|
||||
|
||||
// if (DEBUG) {
|
||||
// System.out.println("\nBTTR.seekExact seg=" + segment + " target=" + fieldInfo.name + ":" + brToString(target) + " current=" + brToString(term) + " (exists?=" + termExists + ") validIndexPrefix=" + validIndexPrefix);
|
||||
// printSeekState();
|
||||
// }
|
||||
|
||||
FST.Arc<Pair<BytesRef,Long>> arc;
|
||||
int targetUpto;
|
||||
Pair<BytesRef,Long> output;
|
||||
|
||||
targetBeforeCurrentLength = currentFrame.ord;
|
||||
|
||||
if (currentFrame != staticFrame) {
|
||||
|
||||
// We are already seek'd; find the common
|
||||
// prefix of new seek term vs current term and
|
||||
// re-use the corresponding seek state. For
|
||||
// example, if app first seeks to foobar, then
|
||||
// seeks to foobaz, we can re-use the seek state
|
||||
// for the first 5 bytes.
|
||||
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" re-use current seek state validIndexPrefix=" + validIndexPrefix);
|
||||
// }
|
||||
|
||||
arc = arcs[0];
|
||||
assert arc.isFinal();
|
||||
output = arc.output;
|
||||
targetUpto = 0;
|
||||
|
||||
IDVersionSegmentTermsEnumFrame lastFrame = stack[0];
|
||||
assert validIndexPrefix <= term.length;
|
||||
|
||||
final int targetLimit = Math.min(target.length, validIndexPrefix);
|
||||
|
||||
int cmp = 0;
|
||||
|
||||
// TODO: reverse vLong byte order for better FST
|
||||
// prefix output sharing
|
||||
|
||||
// First compare up to valid seek frames:
|
||||
while (targetUpto < targetLimit) {
|
||||
cmp = (term.bytes[targetUpto]&0xFF) - (target.bytes[target.offset + targetUpto]&0xFF);
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" cycle targetUpto=" + targetUpto + " (vs limit=" + targetLimit + ") cmp=" + cmp + " (targetLabel=" + (char) (target.bytes[target.offset + targetUpto]) + " vs termLabel=" + (char) (term.bytes[targetUpto]) + ")" + " arc.output=" + arc.output + " output=" + output);
|
||||
// }
|
||||
if (cmp != 0) {
|
||||
break;
|
||||
}
|
||||
arc = arcs[1+targetUpto];
|
||||
//if (arc.label != (target.bytes[target.offset + targetUpto] & 0xFF)) {
|
||||
//System.out.println("FAIL: arc.label=" + (char) arc.label + " targetLabel=" + (char) (target.bytes[target.offset + targetUpto] & 0xFF));
|
||||
//}
|
||||
assert arc.label == (target.bytes[target.offset + targetUpto] & 0xFF): "arc.label=" + (char) arc.label + " targetLabel=" + (char) (target.bytes[target.offset + targetUpto] & 0xFF);
|
||||
if (arc.output != NO_OUTPUT) {
|
||||
output = fstOutputs.add(output, arc.output);
|
||||
}
|
||||
if (arc.isFinal()) {
|
||||
lastFrame = stack[1+lastFrame.ord];
|
||||
}
|
||||
targetUpto++;
|
||||
}
|
||||
|
||||
if (cmp == 0) {
|
||||
final int targetUptoMid = targetUpto;
|
||||
|
||||
// Second compare the rest of the term, but
|
||||
// don't save arc/output/frame; we only do this
|
||||
// to find out if the target term is before,
|
||||
// equal or after the current term
|
||||
final int targetLimit2 = Math.min(target.length, term.length);
|
||||
while (targetUpto < targetLimit2) {
|
||||
cmp = (term.bytes[targetUpto]&0xFF) - (target.bytes[target.offset + targetUpto]&0xFF);
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" cycle2 targetUpto=" + targetUpto + " (vs limit=" + targetLimit + ") cmp=" + cmp + " (targetLabel=" + (char) (target.bytes[target.offset + targetUpto]) + " vs termLabel=" + (char) (term.bytes[targetUpto]) + ")");
|
||||
// }
|
||||
if (cmp != 0) {
|
||||
break;
|
||||
}
|
||||
targetUpto++;
|
||||
}
|
||||
|
||||
if (cmp == 0) {
|
||||
cmp = term.length - target.length;
|
||||
}
|
||||
targetUpto = targetUptoMid;
|
||||
}
|
||||
|
||||
if (cmp < 0) {
|
||||
// Common case: target term is after current
|
||||
// term, ie, app is seeking multiple terms
|
||||
// in sorted order
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" target is after current (shares prefixLen=" + targetUpto + "); frame.ord=" + lastFrame.ord);
|
||||
// }
|
||||
currentFrame = lastFrame;
|
||||
|
||||
} else if (cmp > 0) {
|
||||
// Uncommon case: target term
|
||||
// is before current term; this means we can
|
||||
// keep the currentFrame but we must rewind it
|
||||
// (so we scan from the start)
|
||||
targetBeforeCurrentLength = 0;
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" target is before current (shares prefixLen=" + targetUpto + "); rewind frame ord=" + lastFrame.ord);
|
||||
// }
|
||||
currentFrame = lastFrame;
|
||||
currentFrame.rewind();
|
||||
} else {
|
||||
// Target is exactly the same as current term
|
||||
assert term.length == target.length;
|
||||
if (termExists) {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" target is same as current; return true");
|
||||
// }
|
||||
return true;
|
||||
} else {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" target is same as current but term doesn't exist");
|
||||
// }
|
||||
}
|
||||
//validIndexPrefix = currentFrame.depth;
|
||||
//term.length = target.length;
|
||||
//return termExists;
|
||||
}
|
||||
|
||||
} else {
|
||||
|
||||
targetBeforeCurrentLength = -1;
|
||||
arc = fr.index.getFirstArc(arcs[0]);
|
||||
|
||||
// Empty string prefix must have an output (block) in the index!
|
||||
assert arc.isFinal();
|
||||
assert arc.output != null;
|
||||
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" no seek state; push root frame");
|
||||
// }
|
||||
|
||||
output = arc.output;
|
||||
|
||||
currentFrame = staticFrame;
|
||||
|
||||
//term.length = 0;
|
||||
targetUpto = 0;
|
||||
currentFrame = pushFrame(arc, fstOutputs.add(output, arc.nextFinalOutput), 0);
|
||||
}
|
||||
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" start index loop targetUpto=" + targetUpto + " output=" + output + " currentFrame.ord=" + currentFrame.ord + " targetBeforeCurrentLength=" + targetBeforeCurrentLength);
|
||||
// }
|
||||
|
||||
while (targetUpto < target.length) {
|
||||
|
||||
final int targetLabel = target.bytes[target.offset + targetUpto] & 0xFF;
|
||||
|
||||
final FST.Arc<Pair<BytesRef,Long>> nextArc = fr.index.findTargetArc(targetLabel, arc, getArc(1+targetUpto), fstReader);
|
||||
|
||||
if (nextArc == null) {
|
||||
|
||||
// Index is exhausted
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" index: index exhausted label=" + ((char) targetLabel) + " " + toHex(targetLabel));
|
||||
// }
|
||||
|
||||
validIndexPrefix = currentFrame.prefix;
|
||||
//validIndexPrefix = targetUpto;
|
||||
|
||||
currentFrame.scanToFloorFrame(target);
|
||||
|
||||
if (!currentFrame.hasTerms) {
|
||||
termExists = false;
|
||||
term.bytes[targetUpto] = (byte) targetLabel;
|
||||
term.length = 1+targetUpto;
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" FAST NOT_FOUND term=" + brToString(term));
|
||||
// }
|
||||
return false;
|
||||
}
|
||||
|
||||
if ((Long.MAX_VALUE-output.output2) < minIDVersion) {
|
||||
// The max version for all terms in this block is lower than the minVersion
|
||||
return false;
|
||||
}
|
||||
|
||||
currentFrame.loadBlock();
|
||||
|
||||
final SeekStatus result = currentFrame.scanToTerm(target, true);
|
||||
if (result == SeekStatus.FOUND) {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" return FOUND term=" + term.utf8ToString() + " " + term);
|
||||
// }
|
||||
|
||||
currentFrame.decodeMetaData();
|
||||
if (((IDVersionTermState) currentFrame.state).idVersion < minIDVersion) {
|
||||
// The max version for this term is lower than the minVersion
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
} else {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" got " + result + "; return NOT_FOUND term=" + brToString(term));
|
||||
// }
|
||||
return false;
|
||||
}
|
||||
} else {
|
||||
// Follow this arc
|
||||
arc = nextArc;
|
||||
term.bytes[targetUpto] = (byte) targetLabel;
|
||||
// Aggregate output as we go:
|
||||
assert arc.output != null;
|
||||
if (arc.output != NO_OUTPUT) {
|
||||
output = fstOutputs.add(output, arc.output);
|
||||
}
|
||||
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" index: follow label=" + toHex(target.bytes[target.offset + targetUpto]&0xff) + " arc.output=" + arc.output + " arc.nfo=" + arc.nextFinalOutput);
|
||||
// }
|
||||
targetUpto++;
|
||||
|
||||
if (arc.isFinal()) {
|
||||
//if (DEBUG) System.out.println(" arc is final!");
|
||||
currentFrame = pushFrame(arc, fstOutputs.add(output, arc.nextFinalOutput), targetUpto);
|
||||
//if (DEBUG) System.out.println(" curFrame.ord=" + currentFrame.ord + " hasTerms=" + currentFrame.hasTerms);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
//validIndexPrefix = targetUpto;
|
||||
validIndexPrefix = currentFrame.prefix;
|
||||
|
||||
currentFrame.scanToFloorFrame(target);
|
||||
|
||||
// Target term is entirely contained in the index:
|
||||
if (!currentFrame.hasTerms) {
|
||||
termExists = false;
|
||||
term.length = targetUpto;
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" FAST NOT_FOUND term=" + brToString(term));
|
||||
// }
|
||||
return false;
|
||||
}
|
||||
|
||||
currentFrame.loadBlock();
|
||||
|
||||
final SeekStatus result = currentFrame.scanToTerm(target, true);
|
||||
if (result == SeekStatus.FOUND) {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" return FOUND term=" + term.utf8ToString() + " " + term);
|
||||
// }
|
||||
return true;
|
||||
} else {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" got result " + result + "; return NOT_FOUND term=" + term.utf8ToString());
|
||||
// }
|
||||
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public SeekStatus seekCeil(final BytesRef target) throws IOException {
|
||||
if (fr.index == null) {
|
||||
throw new IllegalStateException("terms index was not loaded");
|
||||
}
|
||||
|
||||
if (term.bytes.length <= target.length) {
|
||||
term.bytes = ArrayUtil.grow(term.bytes, 1+target.length);
|
||||
}
|
||||
|
||||
assert clearEOF();
|
||||
|
||||
//if (DEBUG) {
|
||||
//System.out.println("\nBTTR.seekCeil seg=" + segment + " target=" + fieldInfo.name + ":" + target.utf8ToString() + " " + target + " current=" + brToString(term) + " (exists?=" + termExists + ") validIndexPrefix= " + validIndexPrefix);
|
||||
//printSeekState();
|
||||
//}
|
||||
|
||||
FST.Arc<Pair<BytesRef,Long>> arc;
|
||||
int targetUpto;
|
||||
Pair<BytesRef,Long> output;
|
||||
|
||||
targetBeforeCurrentLength = currentFrame.ord;
|
||||
|
||||
if (currentFrame != staticFrame) {
|
||||
|
||||
// We are already seek'd; find the common
|
||||
// prefix of new seek term vs current term and
|
||||
// re-use the corresponding seek state. For
|
||||
// example, if app first seeks to foobar, then
|
||||
// seeks to foobaz, we can re-use the seek state
|
||||
// for the first 5 bytes.
|
||||
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" re-use current seek state validIndexPrefix=" + validIndexPrefix);
|
||||
//}
|
||||
|
||||
arc = arcs[0];
|
||||
assert arc.isFinal();
|
||||
output = arc.output;
|
||||
targetUpto = 0;
|
||||
|
||||
IDVersionSegmentTermsEnumFrame lastFrame = stack[0];
|
||||
assert validIndexPrefix <= term.length;
|
||||
|
||||
final int targetLimit = Math.min(target.length, validIndexPrefix);
|
||||
|
||||
int cmp = 0;
|
||||
|
||||
// TOOD: we should write our vLong backwards (MSB
|
||||
// first) to get better sharing from the FST
|
||||
|
||||
// First compare up to valid seek frames:
|
||||
while (targetUpto < targetLimit) {
|
||||
cmp = (term.bytes[targetUpto]&0xFF) - (target.bytes[target.offset + targetUpto]&0xFF);
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" cycle targetUpto=" + targetUpto + " (vs limit=" + targetLimit + ") cmp=" + cmp + " (targetLabel=" + (char) (target.bytes[target.offset + targetUpto]) + " vs termLabel=" + (char) (term.bytes[targetUpto]) + ")" + " arc.output=" + arc.output + " output=" + output);
|
||||
//}
|
||||
if (cmp != 0) {
|
||||
break;
|
||||
}
|
||||
arc = arcs[1+targetUpto];
|
||||
assert arc.label == (target.bytes[target.offset + targetUpto] & 0xFF): "arc.label=" + (char) arc.label + " targetLabel=" + (char) (target.bytes[target.offset + targetUpto] & 0xFF);
|
||||
// TOOD: we could save the outputs in local
|
||||
// byte[][] instead of making new objs ever
|
||||
// seek; but, often the FST doesn't have any
|
||||
// shared bytes (but this could change if we
|
||||
// reverse vLong byte order)
|
||||
if (arc.output != NO_OUTPUT) {
|
||||
output = fstOutputs.add(output, arc.output);
|
||||
}
|
||||
if (arc.isFinal()) {
|
||||
lastFrame = stack[1+lastFrame.ord];
|
||||
}
|
||||
targetUpto++;
|
||||
}
|
||||
|
||||
|
||||
if (cmp == 0) {
|
||||
final int targetUptoMid = targetUpto;
|
||||
// Second compare the rest of the term, but
|
||||
// don't save arc/output/frame:
|
||||
final int targetLimit2 = Math.min(target.length, term.length);
|
||||
while (targetUpto < targetLimit2) {
|
||||
cmp = (term.bytes[targetUpto]&0xFF) - (target.bytes[target.offset + targetUpto]&0xFF);
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" cycle2 targetUpto=" + targetUpto + " (vs limit=" + targetLimit + ") cmp=" + cmp + " (targetLabel=" + (char) (target.bytes[target.offset + targetUpto]) + " vs termLabel=" + (char) (term.bytes[targetUpto]) + ")");
|
||||
//}
|
||||
if (cmp != 0) {
|
||||
break;
|
||||
}
|
||||
targetUpto++;
|
||||
}
|
||||
|
||||
if (cmp == 0) {
|
||||
cmp = term.length - target.length;
|
||||
}
|
||||
targetUpto = targetUptoMid;
|
||||
}
|
||||
|
||||
if (cmp < 0) {
|
||||
// Common case: target term is after current
|
||||
// term, ie, app is seeking multiple terms
|
||||
// in sorted order
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" target is after current (shares prefixLen=" + targetUpto + "); clear frame.scanned ord=" + lastFrame.ord);
|
||||
//}
|
||||
currentFrame = lastFrame;
|
||||
|
||||
} else if (cmp > 0) {
|
||||
// Uncommon case: target term
|
||||
// is before current term; this means we can
|
||||
// keep the currentFrame but we must rewind it
|
||||
// (so we scan from the start)
|
||||
targetBeforeCurrentLength = 0;
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" target is before current (shares prefixLen=" + targetUpto + "); rewind frame ord=" + lastFrame.ord);
|
||||
//}
|
||||
currentFrame = lastFrame;
|
||||
currentFrame.rewind();
|
||||
} else {
|
||||
// Target is exactly the same as current term
|
||||
assert term.length == target.length;
|
||||
if (termExists) {
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" target is same as current; return FOUND");
|
||||
//}
|
||||
return SeekStatus.FOUND;
|
||||
} else {
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" target is same as current but term doesn't exist");
|
||||
//}
|
||||
}
|
||||
}
|
||||
|
||||
} else {
|
||||
|
||||
targetBeforeCurrentLength = -1;
|
||||
arc = fr.index.getFirstArc(arcs[0]);
|
||||
|
||||
// Empty string prefix must have an output (block) in the index!
|
||||
assert arc.isFinal();
|
||||
assert arc.output != null;
|
||||
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" no seek state; push root frame");
|
||||
//}
|
||||
|
||||
output = arc.output;
|
||||
|
||||
currentFrame = staticFrame;
|
||||
|
||||
//term.length = 0;
|
||||
targetUpto = 0;
|
||||
currentFrame = pushFrame(arc, fstOutputs.add(output, arc.nextFinalOutput), 0);
|
||||
}
|
||||
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" start index loop targetUpto=" + targetUpto + " output=" + output + " currentFrame.ord+1=" + currentFrame.ord + " targetBeforeCurrentLength=" + targetBeforeCurrentLength);
|
||||
//}
|
||||
|
||||
while (targetUpto < target.length) {
|
||||
|
||||
final int targetLabel = target.bytes[target.offset + targetUpto] & 0xFF;
|
||||
|
||||
final FST.Arc<Pair<BytesRef,Long>> nextArc = fr.index.findTargetArc(targetLabel, arc, getArc(1+targetUpto), fstReader);
|
||||
|
||||
if (nextArc == null) {
|
||||
|
||||
// Index is exhausted
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" index: index exhausted label=" + ((char) targetLabel) + " " + toHex(targetLabel));
|
||||
// }
|
||||
|
||||
validIndexPrefix = currentFrame.prefix;
|
||||
//validIndexPrefix = targetUpto;
|
||||
|
||||
currentFrame.scanToFloorFrame(target);
|
||||
|
||||
currentFrame.loadBlock();
|
||||
|
||||
final SeekStatus result = currentFrame.scanToTerm(target, false);
|
||||
if (result == SeekStatus.END) {
|
||||
term.copyBytes(target);
|
||||
termExists = false;
|
||||
|
||||
if (next() != null) {
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" return NOT_FOUND term=" + brToString(term) + " " + term);
|
||||
//}
|
||||
return SeekStatus.NOT_FOUND;
|
||||
} else {
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" return END");
|
||||
//}
|
||||
return SeekStatus.END;
|
||||
}
|
||||
} else {
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" return " + result + " term=" + brToString(term) + " " + term);
|
||||
//}
|
||||
return result;
|
||||
}
|
||||
} else {
|
||||
// Follow this arc
|
||||
term.bytes[targetUpto] = (byte) targetLabel;
|
||||
arc = nextArc;
|
||||
// Aggregate output as we go:
|
||||
assert arc.output != null;
|
||||
if (arc.output != NO_OUTPUT) {
|
||||
output = fstOutputs.add(output, arc.output);
|
||||
}
|
||||
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" index: follow label=" + toHex(target.bytes[target.offset + targetUpto]&0xff) + " arc.output=" + arc.output + " arc.nfo=" + arc.nextFinalOutput);
|
||||
//}
|
||||
targetUpto++;
|
||||
|
||||
if (arc.isFinal()) {
|
||||
//if (DEBUG) System.out.println(" arc is final!");
|
||||
currentFrame = pushFrame(arc, fstOutputs.add(output, arc.nextFinalOutput), targetUpto);
|
||||
//if (DEBUG) System.out.println(" curFrame.ord=" + currentFrame.ord + " hasTerms=" + currentFrame.hasTerms);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
//validIndexPrefix = targetUpto;
|
||||
validIndexPrefix = currentFrame.prefix;
|
||||
|
||||
currentFrame.scanToFloorFrame(target);
|
||||
|
||||
currentFrame.loadBlock();
|
||||
|
||||
final SeekStatus result = currentFrame.scanToTerm(target, false);
|
||||
|
||||
if (result == SeekStatus.END) {
|
||||
term.copyBytes(target);
|
||||
termExists = false;
|
||||
if (next() != null) {
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" return NOT_FOUND term=" + term.utf8ToString() + " " + term);
|
||||
//}
|
||||
return SeekStatus.NOT_FOUND;
|
||||
} else {
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" return END");
|
||||
//}
|
||||
return SeekStatus.END;
|
||||
}
|
||||
} else {
|
||||
return result;
|
||||
}
|
||||
}
|
||||
|
||||
@SuppressWarnings("unused")
|
||||
private void printSeekState(PrintStream out) throws IOException {
|
||||
if (currentFrame == staticFrame) {
|
||||
out.println(" no prior seek");
|
||||
} else {
|
||||
out.println(" prior seek state:");
|
||||
int ord = 0;
|
||||
boolean isSeekFrame = true;
|
||||
while(true) {
|
||||
IDVersionSegmentTermsEnumFrame f = getFrame(ord);
|
||||
assert f != null;
|
||||
final BytesRef prefix = new BytesRef(term.bytes, 0, f.prefix);
|
||||
if (f.nextEnt == -1) {
|
||||
out.println(" frame " + (isSeekFrame ? "(seek)" : "(next)") + " ord=" + ord + " fp=" + f.fp + (f.isFloor ? (" (fpOrig=" + f.fpOrig + ")") : "") + " prefixLen=" + f.prefix + " prefix=" + prefix + (f.nextEnt == -1 ? "" : (" (of " + f.entCount + ")")) + " hasTerms=" + f.hasTerms + " isFloor=" + f.isFloor + " code=" + ((f.fp<<VersionBlockTreeTermsWriter.OUTPUT_FLAGS_NUM_BITS) + (f.hasTerms ? VersionBlockTreeTermsWriter.OUTPUT_FLAG_HAS_TERMS:0) + (f.isFloor ? VersionBlockTreeTermsWriter.OUTPUT_FLAG_IS_FLOOR:0)) + " isLastInFloor=" + f.isLastInFloor + " mdUpto=" + f.metaDataUpto + " tbOrd=" + f.getTermBlockOrd());
|
||||
} else {
|
||||
out.println(" frame " + (isSeekFrame ? "(seek, loaded)" : "(next, loaded)") + " ord=" + ord + " fp=" + f.fp + (f.isFloor ? (" (fpOrig=" + f.fpOrig + ")") : "") + " prefixLen=" + f.prefix + " prefix=" + prefix + " nextEnt=" + f.nextEnt + (f.nextEnt == -1 ? "" : (" (of " + f.entCount + ")")) + " hasTerms=" + f.hasTerms + " isFloor=" + f.isFloor + " code=" + ((f.fp<<VersionBlockTreeTermsWriter.OUTPUT_FLAGS_NUM_BITS) + (f.hasTerms ? VersionBlockTreeTermsWriter.OUTPUT_FLAG_HAS_TERMS:0) + (f.isFloor ? VersionBlockTreeTermsWriter.OUTPUT_FLAG_IS_FLOOR:0)) + " lastSubFP=" + f.lastSubFP + " isLastInFloor=" + f.isLastInFloor + " mdUpto=" + f.metaDataUpto + " tbOrd=" + f.getTermBlockOrd());
|
||||
}
|
||||
if (fr.index != null) {
|
||||
assert !isSeekFrame || f.arc != null: "isSeekFrame=" + isSeekFrame + " f.arc=" + f.arc;
|
||||
if (f.prefix > 0 && isSeekFrame && f.arc.label != (term.bytes[f.prefix-1]&0xFF)) {
|
||||
out.println(" broken seek state: arc.label=" + (char) f.arc.label + " vs term byte=" + (char) (term.bytes[f.prefix-1]&0xFF));
|
||||
throw new RuntimeException("seek state is broken");
|
||||
}
|
||||
Pair<BytesRef,Long> output = Util.get(fr.index, prefix);
|
||||
if (output == null) {
|
||||
out.println(" broken seek state: prefix is not final in index");
|
||||
throw new RuntimeException("seek state is broken");
|
||||
} else if (isSeekFrame && !f.isFloor) {
|
||||
final ByteArrayDataInput reader = new ByteArrayDataInput(output.output1.bytes, output.output1.offset, output.output1.length);
|
||||
final long codeOrig = reader.readVLong();
|
||||
final long code = (f.fp << VersionBlockTreeTermsWriter.OUTPUT_FLAGS_NUM_BITS) | (f.hasTerms ? VersionBlockTreeTermsWriter.OUTPUT_FLAG_HAS_TERMS:0) | (f.isFloor ? VersionBlockTreeTermsWriter.OUTPUT_FLAG_IS_FLOOR:0);
|
||||
if (codeOrig != code) {
|
||||
out.println(" broken seek state: output code=" + codeOrig + " doesn't match frame code=" + code);
|
||||
throw new RuntimeException("seek state is broken");
|
||||
}
|
||||
}
|
||||
}
|
||||
if (f == currentFrame) {
|
||||
break;
|
||||
}
|
||||
if (f.prefix == validIndexPrefix) {
|
||||
isSeekFrame = false;
|
||||
}
|
||||
ord++;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/* Decodes only the term bytes of the next term. If caller then asks for
|
||||
metadata, ie docFreq, totalTermFreq or pulls a D/&PEnum, we then (lazily)
|
||||
decode all metadata up to the current term. */
|
||||
@Override
|
||||
public BytesRef next() throws IOException {
|
||||
|
||||
if (in == null) {
|
||||
// Fresh TermsEnum; seek to first term:
|
||||
final FST.Arc<Pair<BytesRef,Long>> arc;
|
||||
if (fr.index != null) {
|
||||
arc = fr.index.getFirstArc(arcs[0]);
|
||||
// Empty string prefix must have an output in the index!
|
||||
assert arc.isFinal();
|
||||
} else {
|
||||
arc = null;
|
||||
}
|
||||
currentFrame = pushFrame(arc, fr.rootCode, 0);
|
||||
currentFrame.loadBlock();
|
||||
}
|
||||
|
||||
targetBeforeCurrentLength = currentFrame.ord;
|
||||
|
||||
assert !eof;
|
||||
//if (DEBUG) {
|
||||
//System.out.println("\nBTTR.next seg=" + segment + " term=" + brToString(term) + " termExists?=" + termExists + " field=" + fieldInfo.name + " termBlockOrd=" + currentFrame.state.termBlockOrd + " validIndexPrefix=" + validIndexPrefix);
|
||||
//printSeekState();
|
||||
//}
|
||||
|
||||
if (currentFrame == staticFrame) {
|
||||
// If seek was previously called and the term was
|
||||
// cached, or seek(TermState) was called, usually
|
||||
// caller is just going to pull a D/&PEnum or get
|
||||
// docFreq, etc. But, if they then call next(),
|
||||
// this method catches up all internal state so next()
|
||||
// works properly:
|
||||
//if (DEBUG) System.out.println(" re-seek to pending term=" + term.utf8ToString() + " " + term);
|
||||
final boolean result = seekExact(term);
|
||||
assert result;
|
||||
}
|
||||
|
||||
// Pop finished blocks
|
||||
while (currentFrame.nextEnt == currentFrame.entCount) {
|
||||
if (!currentFrame.isLastInFloor) {
|
||||
currentFrame.loadNextFloorBlock();
|
||||
} else {
|
||||
//if (DEBUG) System.out.println(" pop frame");
|
||||
if (currentFrame.ord == 0) {
|
||||
//if (DEBUG) System.out.println(" return null");
|
||||
assert setEOF();
|
||||
term.length = 0;
|
||||
validIndexPrefix = 0;
|
||||
currentFrame.rewind();
|
||||
termExists = false;
|
||||
return null;
|
||||
}
|
||||
final long lastFP = currentFrame.fpOrig;
|
||||
currentFrame = stack[currentFrame.ord-1];
|
||||
|
||||
if (currentFrame.nextEnt == -1 || currentFrame.lastSubFP != lastFP) {
|
||||
// We popped into a frame that's not loaded
|
||||
// yet or not scan'd to the right entry
|
||||
currentFrame.scanToFloorFrame(term);
|
||||
currentFrame.loadBlock();
|
||||
currentFrame.scanToSubBlock(lastFP);
|
||||
}
|
||||
|
||||
// Note that the seek state (last seek) has been
|
||||
// invalidated beyond this depth
|
||||
validIndexPrefix = Math.min(validIndexPrefix, currentFrame.prefix);
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" reset validIndexPrefix=" + validIndexPrefix);
|
||||
//}
|
||||
}
|
||||
}
|
||||
|
||||
while(true) {
|
||||
if (currentFrame.next()) {
|
||||
// Push to new block:
|
||||
//if (DEBUG) System.out.println(" push frame");
|
||||
currentFrame = pushFrame(null, currentFrame.lastSubFP, term.length);
|
||||
// This is a "next" frame -- even if it's
|
||||
// floor'd we must pretend it isn't so we don't
|
||||
// try to scan to the right floor frame:
|
||||
currentFrame.isFloor = false;
|
||||
//currentFrame.hasTerms = true;
|
||||
currentFrame.loadBlock();
|
||||
} else {
|
||||
//if (DEBUG) System.out.println(" return term=" + term.utf8ToString() + " " + term + " currentFrame.ord=" + currentFrame.ord);
|
||||
return term;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public BytesRef term() {
|
||||
assert !eof;
|
||||
return term;
|
||||
}
|
||||
|
||||
@Override
|
||||
public int docFreq() throws IOException {
|
||||
assert !eof;
|
||||
//if (DEBUG) System.out.println("BTR.docFreq");
|
||||
currentFrame.decodeMetaData();
|
||||
//if (DEBUG) System.out.println(" return " + currentFrame.state.docFreq);
|
||||
return currentFrame.state.docFreq;
|
||||
}
|
||||
|
||||
@Override
|
||||
public long totalTermFreq() throws IOException {
|
||||
assert !eof;
|
||||
currentFrame.decodeMetaData();
|
||||
return currentFrame.state.totalTermFreq;
|
||||
}
|
||||
|
||||
@Override
|
||||
public DocsEnum docs(Bits skipDocs, DocsEnum reuse, int flags) throws IOException {
|
||||
assert !eof;
|
||||
//if (DEBUG) {
|
||||
//System.out.println("BTTR.docs seg=" + segment);
|
||||
//}
|
||||
currentFrame.decodeMetaData();
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" state=" + currentFrame.state);
|
||||
//}
|
||||
return fr.parent.postingsReader.docs(fr.fieldInfo, currentFrame.state, skipDocs, reuse, flags);
|
||||
}
|
||||
|
||||
@Override
|
||||
public DocsAndPositionsEnum docsAndPositions(Bits skipDocs, DocsAndPositionsEnum reuse, int flags) throws IOException {
|
||||
if (fr.fieldInfo.getIndexOptions().compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) < 0) {
|
||||
// Positions were not indexed:
|
||||
return null;
|
||||
}
|
||||
|
||||
assert !eof;
|
||||
currentFrame.decodeMetaData();
|
||||
return fr.parent.postingsReader.docsAndPositions(fr.fieldInfo, currentFrame.state, skipDocs, reuse, flags);
|
||||
}
|
||||
|
||||
@Override
|
||||
public void seekExact(BytesRef target, TermState otherState) {
|
||||
// if (DEBUG) {
|
||||
// System.out.println("BTTR.seekExact termState seg=" + segment + " target=" + target.utf8ToString() + " " + target + " state=" + otherState);
|
||||
// }
|
||||
assert clearEOF();
|
||||
if (target.compareTo(term) != 0 || !termExists) {
|
||||
assert otherState != null && otherState instanceof BlockTermState;
|
||||
currentFrame = staticFrame;
|
||||
currentFrame.state.copyFrom(otherState);
|
||||
term.copyBytes(target);
|
||||
currentFrame.metaDataUpto = currentFrame.getTermBlockOrd();
|
||||
assert currentFrame.metaDataUpto > 0;
|
||||
validIndexPrefix = 0;
|
||||
} else {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" skip seek: already on target state=" + currentFrame.state);
|
||||
// }
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public TermState termState() throws IOException {
|
||||
assert !eof;
|
||||
currentFrame.decodeMetaData();
|
||||
TermState ts = currentFrame.state.clone();
|
||||
//if (DEBUG) System.out.println("BTTR.termState seg=" + segment + " state=" + ts);
|
||||
return ts;
|
||||
}
|
||||
|
||||
@Override
|
||||
public void seekExact(long ord) {
|
||||
throw new UnsupportedOperationException();
|
||||
}
|
||||
|
||||
@Override
|
||||
public long ord() {
|
||||
throw new UnsupportedOperationException();
|
||||
}
|
||||
}
|
||||
|
|
@ -0,0 +1,746 @@
|
|||
package org.apache.lucene.codecs.idversion;
|
||||
|
||||
/*
|
||||
* 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.BlockTermState;
|
||||
import org.apache.lucene.index.FieldInfo.IndexOptions;
|
||||
import org.apache.lucene.index.TermsEnum.SeekStatus;
|
||||
import org.apache.lucene.store.ByteArrayDataInput;
|
||||
import org.apache.lucene.util.ArrayUtil;
|
||||
import org.apache.lucene.util.BytesRef;
|
||||
import org.apache.lucene.util.fst.FST;
|
||||
import org.apache.lucene.util.fst.PairOutputs.Pair;
|
||||
|
||||
final class IDVersionSegmentTermsEnumFrame {
|
||||
// Our index in stack[]:
|
||||
final int ord;
|
||||
|
||||
boolean hasTerms;
|
||||
boolean hasTermsOrig;
|
||||
boolean isFloor;
|
||||
|
||||
FST.Arc<Pair<BytesRef,Long>> arc;
|
||||
|
||||
// File pointer where this block was loaded from
|
||||
long fp;
|
||||
long fpOrig;
|
||||
long fpEnd;
|
||||
|
||||
byte[] suffixBytes = new byte[128];
|
||||
final ByteArrayDataInput suffixesReader = new ByteArrayDataInput();
|
||||
|
||||
byte[] statBytes = new byte[64];
|
||||
final ByteArrayDataInput statsReader = new ByteArrayDataInput();
|
||||
|
||||
byte[] floorData = new byte[32];
|
||||
final ByteArrayDataInput floorDataReader = new ByteArrayDataInput();
|
||||
|
||||
// Length of prefix shared by all terms in this block
|
||||
int prefix;
|
||||
|
||||
// Number of entries (term or sub-block) in this block
|
||||
int entCount;
|
||||
|
||||
// Which term we will next read, or -1 if the block
|
||||
// isn't loaded yet
|
||||
int nextEnt;
|
||||
|
||||
// True if this block is either not a floor block,
|
||||
// or, it's the last sub-block of a floor block
|
||||
boolean isLastInFloor;
|
||||
|
||||
// True if all entries are terms
|
||||
boolean isLeafBlock;
|
||||
|
||||
long lastSubFP;
|
||||
|
||||
int nextFloorLabel;
|
||||
int numFollowFloorBlocks;
|
||||
|
||||
// Next term to decode metaData; we decode metaData
|
||||
// lazily so that scanning to find the matching term is
|
||||
// fast and only if you find a match and app wants the
|
||||
// stats or docs/positions enums, will we decode the
|
||||
// metaData
|
||||
int metaDataUpto;
|
||||
|
||||
final BlockTermState state;
|
||||
|
||||
// metadata buffer, holding monotonic values
|
||||
public long[] longs;
|
||||
// metadata buffer, holding general values
|
||||
public byte[] bytes;
|
||||
ByteArrayDataInput bytesReader;
|
||||
|
||||
private final IDVersionSegmentTermsEnum ste;
|
||||
|
||||
public IDVersionSegmentTermsEnumFrame(IDVersionSegmentTermsEnum ste, int ord) throws IOException {
|
||||
this.ste = ste;
|
||||
this.ord = ord;
|
||||
this.state = ste.fr.parent.postingsReader.newTermState();
|
||||
this.state.totalTermFreq = -1;
|
||||
this.longs = new long[ste.fr.longsSize];
|
||||
}
|
||||
|
||||
public void setFloorData(ByteArrayDataInput in, BytesRef source) {
|
||||
final int numBytes = source.length - (in.getPosition() - source.offset);
|
||||
if (numBytes > floorData.length) {
|
||||
floorData = new byte[ArrayUtil.oversize(numBytes, 1)];
|
||||
}
|
||||
System.arraycopy(source.bytes, source.offset+in.getPosition(), floorData, 0, numBytes);
|
||||
floorDataReader.reset(floorData, 0, numBytes);
|
||||
numFollowFloorBlocks = floorDataReader.readVInt();
|
||||
nextFloorLabel = floorDataReader.readByte() & 0xff;
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" setFloorData fpOrig=" + fpOrig + " bytes=" + new BytesRef(source.bytes, source.offset + in.getPosition(), numBytes) + " numFollowFloorBlocks=" + numFollowFloorBlocks + " nextFloorLabel=" + toHex(nextFloorLabel));
|
||||
//}
|
||||
}
|
||||
|
||||
public int getTermBlockOrd() {
|
||||
return isLeafBlock ? nextEnt : state.termBlockOrd;
|
||||
}
|
||||
|
||||
void loadNextFloorBlock() throws IOException {
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" loadNextFloorBlock fp=" + fp + " fpEnd=" + fpEnd);
|
||||
//}
|
||||
assert arc == null || isFloor: "arc=" + arc + " isFloor=" + isFloor;
|
||||
fp = fpEnd;
|
||||
nextEnt = -1;
|
||||
loadBlock();
|
||||
}
|
||||
|
||||
/* Does initial decode of next block of terms; this
|
||||
doesn't actually decode the docFreq, totalTermFreq,
|
||||
postings details (frq/prx offset, etc.) metadata;
|
||||
it just loads them as byte[] blobs which are then
|
||||
decoded on-demand if the metadata is ever requested
|
||||
for any term in this block. This enables terms-only
|
||||
intensive consumes (eg certain MTQs, respelling) to
|
||||
not pay the price of decoding metadata they won't
|
||||
use. */
|
||||
void loadBlock() throws IOException {
|
||||
|
||||
// Clone the IndexInput lazily, so that consumers
|
||||
// that just pull a TermsEnum to
|
||||
// seekExact(TermState) don't pay this cost:
|
||||
ste.initIndexInput();
|
||||
|
||||
if (nextEnt != -1) {
|
||||
// Already loaded
|
||||
return;
|
||||
}
|
||||
//System.out.println("blc=" + blockLoadCount);
|
||||
|
||||
ste.in.seek(fp);
|
||||
int code = ste.in.readVInt();
|
||||
entCount = code >>> 1;
|
||||
assert entCount > 0;
|
||||
isLastInFloor = (code & 1) != 0;
|
||||
assert arc == null || (isLastInFloor || isFloor);
|
||||
|
||||
// TODO: if suffixes were stored in random-access
|
||||
// array structure, then we could do binary search
|
||||
// instead of linear scan to find target term; eg
|
||||
// we could have simple array of offsets
|
||||
|
||||
// term suffixes:
|
||||
code = ste.in.readVInt();
|
||||
isLeafBlock = (code & 1) != 0;
|
||||
int numBytes = code >>> 1;
|
||||
if (suffixBytes.length < numBytes) {
|
||||
suffixBytes = new byte[ArrayUtil.oversize(numBytes, 1)];
|
||||
}
|
||||
ste.in.readBytes(suffixBytes, 0, numBytes);
|
||||
suffixesReader.reset(suffixBytes, 0, numBytes);
|
||||
|
||||
/*if (DEBUG) {
|
||||
if (arc == null) {
|
||||
System.out.println(" loadBlock (next) fp=" + fp + " entCount=" + entCount + " prefixLen=" + prefix + " isLastInFloor=" + isLastInFloor + " leaf?=" + isLeafBlock);
|
||||
} else {
|
||||
System.out.println(" loadBlock (seek) fp=" + fp + " entCount=" + entCount + " prefixLen=" + prefix + " hasTerms?=" + hasTerms + " isFloor?=" + isFloor + " isLastInFloor=" + isLastInFloor + " leaf?=" + isLeafBlock);
|
||||
}
|
||||
}*/
|
||||
|
||||
// stats
|
||||
numBytes = ste.in.readVInt();
|
||||
if (statBytes.length < numBytes) {
|
||||
statBytes = new byte[ArrayUtil.oversize(numBytes, 1)];
|
||||
}
|
||||
ste.in.readBytes(statBytes, 0, numBytes);
|
||||
statsReader.reset(statBytes, 0, numBytes);
|
||||
metaDataUpto = 0;
|
||||
|
||||
state.termBlockOrd = 0;
|
||||
nextEnt = 0;
|
||||
lastSubFP = -1;
|
||||
|
||||
// TODO: we could skip this if !hasTerms; but
|
||||
// that's rare so won't help much
|
||||
// metadata
|
||||
numBytes = ste.in.readVInt();
|
||||
if (bytes == null) {
|
||||
bytes = new byte[ArrayUtil.oversize(numBytes, 1)];
|
||||
bytesReader = new ByteArrayDataInput();
|
||||
} else if (bytes.length < numBytes) {
|
||||
bytes = new byte[ArrayUtil.oversize(numBytes, 1)];
|
||||
}
|
||||
ste.in.readBytes(bytes, 0, numBytes);
|
||||
bytesReader.reset(bytes, 0, numBytes);
|
||||
|
||||
|
||||
// Sub-blocks of a single floor block are always
|
||||
// written one after another -- tail recurse:
|
||||
fpEnd = ste.in.getFilePointer();
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" fpEnd=" + fpEnd);
|
||||
// }
|
||||
}
|
||||
|
||||
void rewind() {
|
||||
|
||||
// Force reload:
|
||||
fp = fpOrig;
|
||||
nextEnt = -1;
|
||||
hasTerms = hasTermsOrig;
|
||||
if (isFloor) {
|
||||
floorDataReader.rewind();
|
||||
numFollowFloorBlocks = floorDataReader.readVInt();
|
||||
nextFloorLabel = floorDataReader.readByte() & 0xff;
|
||||
}
|
||||
|
||||
/*
|
||||
//System.out.println("rewind");
|
||||
// Keeps the block loaded, but rewinds its state:
|
||||
if (nextEnt > 0 || fp != fpOrig) {
|
||||
if (DEBUG) {
|
||||
System.out.println(" rewind frame ord=" + ord + " fpOrig=" + fpOrig + " fp=" + fp + " hasTerms?=" + hasTerms + " isFloor?=" + isFloor + " nextEnt=" + nextEnt + " prefixLen=" + prefix);
|
||||
}
|
||||
if (fp != fpOrig) {
|
||||
fp = fpOrig;
|
||||
nextEnt = -1;
|
||||
} else {
|
||||
nextEnt = 0;
|
||||
}
|
||||
hasTerms = hasTermsOrig;
|
||||
if (isFloor) {
|
||||
floorDataReader.rewind();
|
||||
numFollowFloorBlocks = floorDataReader.readVInt();
|
||||
nextFloorLabel = floorDataReader.readByte() & 0xff;
|
||||
}
|
||||
assert suffixBytes != null;
|
||||
suffixesReader.rewind();
|
||||
assert statBytes != null;
|
||||
statsReader.rewind();
|
||||
metaDataUpto = 0;
|
||||
state.termBlockOrd = 0;
|
||||
// TODO: skip this if !hasTerms? Then postings
|
||||
// impl wouldn't have to write useless 0 byte
|
||||
postingsReader.resetTermsBlock(fieldInfo, state);
|
||||
lastSubFP = -1;
|
||||
} else if (DEBUG) {
|
||||
System.out.println(" skip rewind fp=" + fp + " fpOrig=" + fpOrig + " nextEnt=" + nextEnt + " ord=" + ord);
|
||||
}
|
||||
*/
|
||||
}
|
||||
|
||||
public boolean next() {
|
||||
return isLeafBlock ? nextLeaf() : nextNonLeaf();
|
||||
}
|
||||
|
||||
// Decodes next entry; returns true if it's a sub-block
|
||||
public boolean nextLeaf() {
|
||||
//if (DEBUG) System.out.println(" frame.next ord=" + ord + " nextEnt=" + nextEnt + " entCount=" + entCount);
|
||||
assert nextEnt != -1 && nextEnt < entCount: "nextEnt=" + nextEnt + " entCount=" + entCount + " fp=" + fp;
|
||||
nextEnt++;
|
||||
suffix = suffixesReader.readVInt();
|
||||
startBytePos = suffixesReader.getPosition();
|
||||
ste.term.length = prefix + suffix;
|
||||
if (ste.term.bytes.length < ste.term.length) {
|
||||
ste.term.grow(ste.term.length);
|
||||
}
|
||||
suffixesReader.readBytes(ste.term.bytes, prefix, suffix);
|
||||
// A normal term
|
||||
ste.termExists = true;
|
||||
return false;
|
||||
}
|
||||
|
||||
public boolean nextNonLeaf() {
|
||||
//if (DEBUG) System.out.println(" frame.next ord=" + ord + " nextEnt=" + nextEnt + " entCount=" + entCount);
|
||||
assert nextEnt != -1 && nextEnt < entCount: "nextEnt=" + nextEnt + " entCount=" + entCount + " fp=" + fp;
|
||||
nextEnt++;
|
||||
final int code = suffixesReader.readVInt();
|
||||
suffix = code >>> 1;
|
||||
startBytePos = suffixesReader.getPosition();
|
||||
ste.term.length = prefix + suffix;
|
||||
if (ste.term.bytes.length < ste.term.length) {
|
||||
ste.term.grow(ste.term.length);
|
||||
}
|
||||
suffixesReader.readBytes(ste.term.bytes, prefix, suffix);
|
||||
if ((code & 1) == 0) {
|
||||
// A normal term
|
||||
ste.termExists = true;
|
||||
subCode = 0;
|
||||
state.termBlockOrd++;
|
||||
return false;
|
||||
} else {
|
||||
// A sub-block; make sub-FP absolute:
|
||||
ste.termExists = false;
|
||||
subCode = suffixesReader.readVLong();
|
||||
lastSubFP = fp - subCode;
|
||||
//if (DEBUG) {
|
||||
//System.out.println(" lastSubFP=" + lastSubFP);
|
||||
//}
|
||||
return true;
|
||||
}
|
||||
}
|
||||
|
||||
// TODO: make this array'd so we can do bin search?
|
||||
// likely not worth it? need to measure how many
|
||||
// floor blocks we "typically" get
|
||||
public void scanToFloorFrame(BytesRef target) {
|
||||
|
||||
if (!isFloor || target.length <= prefix) {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" scanToFloorFrame skip: isFloor=" + isFloor + " target.length=" + target.length + " vs prefix=" + prefix);
|
||||
// }
|
||||
return;
|
||||
}
|
||||
|
||||
final int targetLabel = target.bytes[target.offset + prefix] & 0xFF;
|
||||
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" scanToFloorFrame fpOrig=" + fpOrig + " targetLabel=" + toHex(targetLabel) + " vs nextFloorLabel=" + toHex(nextFloorLabel) + " numFollowFloorBlocks=" + numFollowFloorBlocks);
|
||||
// }
|
||||
|
||||
if (targetLabel < nextFloorLabel) {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" already on correct block");
|
||||
// }
|
||||
return;
|
||||
}
|
||||
|
||||
assert numFollowFloorBlocks != 0;
|
||||
|
||||
long newFP = fpOrig;
|
||||
while (true) {
|
||||
final long code = floorDataReader.readVLong();
|
||||
newFP = fpOrig + (code >>> 1);
|
||||
hasTerms = (code & 1) != 0;
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" label=" + toHex(nextFloorLabel) + " fp=" + newFP + " hasTerms?=" + hasTerms + " numFollowFloor=" + numFollowFloorBlocks);
|
||||
// }
|
||||
|
||||
isLastInFloor = numFollowFloorBlocks == 1;
|
||||
numFollowFloorBlocks--;
|
||||
|
||||
if (isLastInFloor) {
|
||||
nextFloorLabel = 256;
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" stop! last block nextFloorLabel=" + toHex(nextFloorLabel));
|
||||
// }
|
||||
break;
|
||||
} else {
|
||||
nextFloorLabel = floorDataReader.readByte() & 0xff;
|
||||
if (targetLabel < nextFloorLabel) {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" stop! nextFloorLabel=" + toHex(nextFloorLabel));
|
||||
// }
|
||||
break;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if (newFP != fp) {
|
||||
// Force re-load of the block:
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" force switch to fp=" + newFP + " oldFP=" + fp);
|
||||
// }
|
||||
nextEnt = -1;
|
||||
fp = newFP;
|
||||
} else {
|
||||
// if (DEBUG) {
|
||||
// System.out.println(" stay on same fp=" + newFP);
|
||||
// }
|
||||
}
|
||||
}
|
||||
|
||||
public void decodeMetaData() throws IOException {
|
||||
|
||||
//if (DEBUG) System.out.println("\nBTTR.decodeMetadata seg=" + segment + " mdUpto=" + metaDataUpto + " vs termBlockOrd=" + state.termBlockOrd);
|
||||
|
||||
// lazily catch up on metadata decode:
|
||||
final int limit = getTermBlockOrd();
|
||||
boolean absolute = metaDataUpto == 0;
|
||||
assert limit > 0;
|
||||
|
||||
// TODO: better API would be "jump straight to term=N"???
|
||||
while (metaDataUpto < limit) {
|
||||
|
||||
// TODO: we could make "tiers" of metadata, ie,
|
||||
// decode docFreq/totalTF but don't decode postings
|
||||
// metadata; this way caller could get
|
||||
// docFreq/totalTF w/o paying decode cost for
|
||||
// postings
|
||||
|
||||
// TODO: if docFreq were bulk decoded we could
|
||||
// just skipN here:
|
||||
|
||||
// stats
|
||||
state.docFreq = statsReader.readVInt();
|
||||
//if (DEBUG) System.out.println(" dF=" + state.docFreq);
|
||||
if (ste.fr.fieldInfo.getIndexOptions() != IndexOptions.DOCS_ONLY) {
|
||||
state.totalTermFreq = state.docFreq + statsReader.readVLong();
|
||||
//if (DEBUG) System.out.println(" totTF=" + state.totalTermFreq);
|
||||
}
|
||||
// metadata
|
||||
for (int i = 0; i < ste.fr.longsSize; i++) {
|
||||
longs[i] = bytesReader.readVLong();
|
||||
}
|
||||
ste.fr.parent.postingsReader.decodeTerm(longs, bytesReader, ste.fr.fieldInfo, state, absolute);
|
||||
|
||||
metaDataUpto++;
|
||||
absolute = false;
|
||||
}
|
||||
state.termBlockOrd = metaDataUpto;
|
||||
}
|
||||
|
||||
// Used only by assert
|
||||
private boolean prefixMatches(BytesRef target) {
|
||||
for(int bytePos=0;bytePos<prefix;bytePos++) {
|
||||
if (target.bytes[target.offset + bytePos] != ste.term.bytes[bytePos]) {
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
return true;
|
||||
}
|
||||
|
||||
// Scans to sub-block that has this target fp; only
|
||||
// called by next(); NOTE: does not set
|
||||
// startBytePos/suffix as a side effect
|
||||
public void scanToSubBlock(long subFP) {
|
||||
assert !isLeafBlock;
|
||||
//if (DEBUG) System.out.println(" scanToSubBlock fp=" + fp + " subFP=" + subFP + " entCount=" + entCount + " lastSubFP=" + lastSubFP);
|
||||
//assert nextEnt == 0;
|
||||
if (lastSubFP == subFP) {
|
||||
//if (DEBUG) System.out.println(" already positioned");
|
||||
return;
|
||||
}
|
||||
assert subFP < fp : "fp=" + fp + " subFP=" + subFP;
|
||||
final long targetSubCode = fp - subFP;
|
||||
//if (DEBUG) System.out.println(" targetSubCode=" + targetSubCode);
|
||||
while(true) {
|
||||
assert nextEnt < entCount;
|
||||
nextEnt++;
|
||||
final int code = suffixesReader.readVInt();
|
||||
suffixesReader.skipBytes(isLeafBlock ? code : code >>> 1);
|
||||
//if (DEBUG) System.out.println(" " + nextEnt + " (of " + entCount + ") ent isSubBlock=" + ((code&1)==1));
|
||||
if ((code & 1) != 0) {
|
||||
final long subCode = suffixesReader.readVLong();
|
||||
//if (DEBUG) System.out.println(" subCode=" + subCode);
|
||||
if (targetSubCode == subCode) {
|
||||
//if (DEBUG) System.out.println(" match!");
|
||||
lastSubFP = subFP;
|
||||
return;
|
||||
}
|
||||
} else {
|
||||
state.termBlockOrd++;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// NOTE: sets startBytePos/suffix as a side effect
|
||||
public SeekStatus scanToTerm(BytesRef target, boolean exactOnly) throws IOException {
|
||||
return isLeafBlock ? scanToTermLeaf(target, exactOnly) : scanToTermNonLeaf(target, exactOnly);
|
||||
}
|
||||
|
||||
private int startBytePos;
|
||||
private int suffix;
|
||||
private long subCode;
|
||||
|
||||
// Target's prefix matches this block's prefix; we
|
||||
// scan the entries check if the suffix matches.
|
||||
public SeekStatus scanToTermLeaf(BytesRef target, boolean exactOnly) throws IOException {
|
||||
|
||||
// if (DEBUG) System.out.println(" scanToTermLeaf: block fp=" + fp + " prefix=" + prefix + " nextEnt=" + nextEnt + " (of " + entCount + ") target=" + brToString(target) + " term=" + brToString(term));
|
||||
|
||||
assert nextEnt != -1;
|
||||
|
||||
ste.termExists = true;
|
||||
subCode = 0;
|
||||
|
||||
if (nextEnt == entCount) {
|
||||
if (exactOnly) {
|
||||
fillTerm();
|
||||
}
|
||||
return SeekStatus.END;
|
||||
}
|
||||
|
||||
assert prefixMatches(target);
|
||||
|
||||
// Loop over each entry (term or sub-block) in this block:
|
||||
//nextTerm: while(nextEnt < entCount) {
|
||||
nextTerm: while (true) {
|
||||
nextEnt++;
|
||||
|
||||
suffix = suffixesReader.readVInt();
|
||||
|
||||
// if (DEBUG) {
|
||||
// BytesRef suffixBytesRef = new BytesRef();
|
||||
// suffixBytesRef.bytes = suffixBytes;
|
||||
// suffixBytesRef.offset = suffixesReader.getPosition();
|
||||
// suffixBytesRef.length = suffix;
|
||||
// System.out.println(" cycle: term " + (nextEnt-1) + " (of " + entCount + ") suffix=" + brToString(suffixBytesRef));
|
||||
// }
|
||||
|
||||
final int termLen = prefix + suffix;
|
||||
startBytePos = suffixesReader.getPosition();
|
||||
suffixesReader.skipBytes(suffix);
|
||||
|
||||
final int targetLimit = target.offset + (target.length < termLen ? target.length : termLen);
|
||||
int targetPos = target.offset + prefix;
|
||||
|
||||
// Loop over bytes in the suffix, comparing to
|
||||
// the target
|
||||
int bytePos = startBytePos;
|
||||
while(true) {
|
||||
final int cmp;
|
||||
final boolean stop;
|
||||
if (targetPos < targetLimit) {
|
||||
cmp = (suffixBytes[bytePos++]&0xFF) - (target.bytes[targetPos++]&0xFF);
|
||||
stop = false;
|
||||
} else {
|
||||
assert targetPos == targetLimit;
|
||||
cmp = termLen - target.length;
|
||||
stop = true;
|
||||
}
|
||||
|
||||
if (cmp < 0) {
|
||||
// Current entry is still before the target;
|
||||
// keep scanning
|
||||
|
||||
if (nextEnt == entCount) {
|
||||
if (exactOnly) {
|
||||
fillTerm();
|
||||
}
|
||||
// We are done scanning this block
|
||||
break nextTerm;
|
||||
} else {
|
||||
continue nextTerm;
|
||||
}
|
||||
} else if (cmp > 0) {
|
||||
|
||||
// Done! Current entry is after target --
|
||||
// return NOT_FOUND:
|
||||
fillTerm();
|
||||
|
||||
if (!exactOnly && !ste.termExists) {
|
||||
// We are on a sub-block, and caller wants
|
||||
// us to position to the next term after
|
||||
// the target, so we must recurse into the
|
||||
// sub-frame(s):
|
||||
ste.currentFrame = ste.pushFrame(null, ste.currentFrame.lastSubFP, termLen);
|
||||
ste.currentFrame.loadBlock();
|
||||
while (ste.currentFrame.next()) {
|
||||
ste.currentFrame = ste.pushFrame(null, ste.currentFrame.lastSubFP, ste.term.length);
|
||||
ste.currentFrame.loadBlock();
|
||||
}
|
||||
}
|
||||
|
||||
//if (DEBUG) System.out.println(" not found");
|
||||
return SeekStatus.NOT_FOUND;
|
||||
} else if (stop) {
|
||||
// Exact match!
|
||||
|
||||
// This cannot be a sub-block because we
|
||||
// would have followed the index to this
|
||||
// sub-block from the start:
|
||||
|
||||
assert ste.termExists;
|
||||
fillTerm();
|
||||
//if (DEBUG) System.out.println(" found!");
|
||||
return SeekStatus.FOUND;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// It is possible (and OK) that terms index pointed us
|
||||
// at this block, but, we scanned the entire block and
|
||||
// did not find the term to position to. This happens
|
||||
// when the target is after the last term in the block
|
||||
// (but, before the next term in the index). EG
|
||||
// target could be foozzz, and terms index pointed us
|
||||
// to the foo* block, but the last term in this block
|
||||
// was fooz (and, eg, first term in the next block will
|
||||
// bee fop).
|
||||
//if (DEBUG) System.out.println(" block end");
|
||||
if (exactOnly) {
|
||||
fillTerm();
|
||||
}
|
||||
|
||||
// TODO: not consistent that in the
|
||||
// not-exact case we don't next() into the next
|
||||
// frame here
|
||||
return SeekStatus.END;
|
||||
}
|
||||
|
||||
// Target's prefix matches this block's prefix; we
|
||||
// scan the entries check if the suffix matches.
|
||||
public SeekStatus scanToTermNonLeaf(BytesRef target, boolean exactOnly) throws IOException {
|
||||
|
||||
//if (DEBUG) System.out.println(" scanToTermNonLeaf: block fp=" + fp + " prefix=" + prefix + " nextEnt=" + nextEnt + " (of " + entCount + ") target=" + brToString(target) + " term=" + brToString(term));
|
||||
|
||||
assert nextEnt != -1;
|
||||
|
||||
if (nextEnt == entCount) {
|
||||
if (exactOnly) {
|
||||
fillTerm();
|
||||
ste.termExists = subCode == 0;
|
||||
}
|
||||
return SeekStatus.END;
|
||||
}
|
||||
|
||||
assert prefixMatches(target);
|
||||
|
||||
// Loop over each entry (term or sub-block) in this block:
|
||||
//nextTerm: while(nextEnt < entCount) {
|
||||
nextTerm: while (true) {
|
||||
nextEnt++;
|
||||
|
||||
final int code = suffixesReader.readVInt();
|
||||
suffix = code >>> 1;
|
||||
// if (DEBUG) {
|
||||
// BytesRef suffixBytesRef = new BytesRef();
|
||||
// suffixBytesRef.bytes = suffixBytes;
|
||||
// suffixBytesRef.offset = suffixesReader.getPosition();
|
||||
// suffixBytesRef.length = suffix;
|
||||
// System.out.println(" cycle: " + ((code&1)==1 ? "sub-block" : "term") + " " + (nextEnt-1) + " (of " + entCount + ") suffix=" + brToString(suffixBytesRef));
|
||||
// }
|
||||
|
||||
ste.termExists = (code & 1) == 0;
|
||||
final int termLen = prefix + suffix;
|
||||
startBytePos = suffixesReader.getPosition();
|
||||
suffixesReader.skipBytes(suffix);
|
||||
if (ste.termExists) {
|
||||
state.termBlockOrd++;
|
||||
subCode = 0;
|
||||
} else {
|
||||
subCode = suffixesReader.readVLong();
|
||||
lastSubFP = fp - subCode;
|
||||
}
|
||||
|
||||
final int targetLimit = target.offset + (target.length < termLen ? target.length : termLen);
|
||||
int targetPos = target.offset + prefix;
|
||||
|
||||
// Loop over bytes in the suffix, comparing to
|
||||
// the target
|
||||
int bytePos = startBytePos;
|
||||
while(true) {
|
||||
final int cmp;
|
||||
final boolean stop;
|
||||
if (targetPos < targetLimit) {
|
||||
cmp = (suffixBytes[bytePos++]&0xFF) - (target.bytes[targetPos++]&0xFF);
|
||||
stop = false;
|
||||
} else {
|
||||
assert targetPos == targetLimit;
|
||||
cmp = termLen - target.length;
|
||||
stop = true;
|
||||
}
|
||||
|
||||
if (cmp < 0) {
|
||||
// Current entry is still before the target;
|
||||
// keep scanning
|
||||
|
||||
if (nextEnt == entCount) {
|
||||
if (exactOnly) {
|
||||
fillTerm();
|
||||
//termExists = true;
|
||||
}
|
||||
// We are done scanning this block
|
||||
break nextTerm;
|
||||
} else {
|
||||
continue nextTerm;
|
||||
}
|
||||
} else if (cmp > 0) {
|
||||
|
||||
// Done! Current entry is after target --
|
||||
// return NOT_FOUND:
|
||||
fillTerm();
|
||||
|
||||
if (!exactOnly && !ste.termExists) {
|
||||
// We are on a sub-block, and caller wants
|
||||
// us to position to the next term after
|
||||
// the target, so we must recurse into the
|
||||
// sub-frame(s):
|
||||
ste.currentFrame = ste.pushFrame(null, ste.currentFrame.lastSubFP, termLen);
|
||||
ste.currentFrame.loadBlock();
|
||||
while (ste.currentFrame.next()) {
|
||||
ste.currentFrame = ste.pushFrame(null, ste.currentFrame.lastSubFP, ste.term.length);
|
||||
ste.currentFrame.loadBlock();
|
||||
}
|
||||
}
|
||||
|
||||
//if (DEBUG) System.out.println(" not found");
|
||||
return SeekStatus.NOT_FOUND;
|
||||
} else if (stop) {
|
||||
// Exact match!
|
||||
|
||||
// This cannot be a sub-block because we
|
||||
// would have followed the index to this
|
||||
// sub-block from the start:
|
||||
|
||||
assert ste.termExists;
|
||||
fillTerm();
|
||||
//if (DEBUG) System.out.println(" found!");
|
||||
return SeekStatus.FOUND;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// It is possible (and OK) that terms index pointed us
|
||||
// at this block, but, we scanned the entire block and
|
||||
// did not find the term to position to. This happens
|
||||
// when the target is after the last term in the block
|
||||
// (but, before the next term in the index). EG
|
||||
// target could be foozzz, and terms index pointed us
|
||||
// to the foo* block, but the last term in this block
|
||||
// was fooz (and, eg, first term in the next block will
|
||||
// bee fop).
|
||||
//if (DEBUG) System.out.println(" block end");
|
||||
if (exactOnly) {
|
||||
fillTerm();
|
||||
}
|
||||
|
||||
// TODO: not consistent that in the
|
||||
// not-exact case we don't next() into the next
|
||||
// frame here
|
||||
return SeekStatus.END;
|
||||
}
|
||||
|
||||
private void fillTerm() {
|
||||
final int termLength = prefix + suffix;
|
||||
ste.term.length = prefix + suffix;
|
||||
if (ste.term.bytes.length < termLength) {
|
||||
ste.term.grow(termLength);
|
||||
}
|
||||
System.arraycopy(suffixBytes, startBytePos, ste.term.bytes, prefix, suffix);
|
||||
}
|
||||
}
|
||||
|
|
@ -0,0 +1,41 @@
|
|||
package org.apache.lucene.codecs.idversion;
|
||||
|
||||
/*
|
||||
* 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 org.apache.lucene.codecs.BlockTermState;
|
||||
import org.apache.lucene.index.TermState;
|
||||
|
||||
final class IDVersionTermState extends BlockTermState {
|
||||
long idVersion;
|
||||
int docID;
|
||||
|
||||
@Override
|
||||
public IDVersionTermState clone() {
|
||||
IDVersionTermState other = new IDVersionTermState();
|
||||
other.copyFrom(this);
|
||||
return other;
|
||||
}
|
||||
|
||||
@Override
|
||||
public void copyFrom(TermState _other) {
|
||||
super.copyFrom(_other);
|
||||
IDVersionTermState other = (IDVersionTermState) _other;
|
||||
idVersion = other.idVersion;
|
||||
docID = other.docID;
|
||||
}
|
||||
}
|
||||
|
|
@ -0,0 +1,319 @@
|
|||
package org.apache.lucene.codecs.idversion;
|
||||
|
||||
/*
|
||||
* 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 java.io.PrintStream;
|
||||
import java.util.Collections;
|
||||
import java.util.Iterator;
|
||||
import java.util.TreeMap;
|
||||
|
||||
import org.apache.lucene.codecs.BlockTermState;
|
||||
import org.apache.lucene.codecs.CodecUtil;
|
||||
import org.apache.lucene.codecs.FieldsProducer;
|
||||
import org.apache.lucene.codecs.PostingsReaderBase;
|
||||
import org.apache.lucene.index.CorruptIndexException;
|
||||
import org.apache.lucene.index.DocsAndPositionsEnum;
|
||||
import org.apache.lucene.index.DocsEnum;
|
||||
import org.apache.lucene.index.FieldInfo.IndexOptions;
|
||||
import org.apache.lucene.index.FieldInfo;
|
||||
import org.apache.lucene.index.FieldInfos;
|
||||
import org.apache.lucene.index.IndexFileNames;
|
||||
import org.apache.lucene.index.SegmentInfo;
|
||||
import org.apache.lucene.index.TermState;
|
||||
import org.apache.lucene.index.Terms;
|
||||
import org.apache.lucene.index.TermsEnum;
|
||||
import org.apache.lucene.store.ByteArrayDataInput;
|
||||
import org.apache.lucene.store.Directory;
|
||||
import org.apache.lucene.store.IOContext;
|
||||
import org.apache.lucene.store.IndexInput;
|
||||
import org.apache.lucene.util.ArrayUtil;
|
||||
import org.apache.lucene.util.Bits;
|
||||
import org.apache.lucene.util.BytesRef;
|
||||
import org.apache.lucene.util.IOUtils;
|
||||
import org.apache.lucene.util.RamUsageEstimator;
|
||||
import org.apache.lucene.util.StringHelper;
|
||||
import org.apache.lucene.util.automaton.CompiledAutomaton;
|
||||
import org.apache.lucene.util.automaton.RunAutomaton;
|
||||
import org.apache.lucene.util.automaton.Transition;
|
||||
import org.apache.lucene.util.fst.ByteSequenceOutputs;
|
||||
import org.apache.lucene.util.fst.FST;
|
||||
import org.apache.lucene.util.fst.Outputs;
|
||||
import org.apache.lucene.util.fst.PairOutputs.Pair;
|
||||
import org.apache.lucene.util.fst.PairOutputs;
|
||||
import org.apache.lucene.util.fst.Util;
|
||||
|
||||
/** A block-based terms index and dictionary that assigns
|
||||
* terms to variable length blocks according to how they
|
||||
* share prefixes. The terms index is a prefix trie
|
||||
* whose leaves are term blocks. The advantage of this
|
||||
* approach is that seekExact is often able to
|
||||
* determine a term cannot exist without doing any IO, and
|
||||
* intersection with Automata is very fast. Note that this
|
||||
* terms dictionary has it's own fixed terms index (ie, it
|
||||
* does not support a pluggable terms index
|
||||
* implementation).
|
||||
*
|
||||
* <p><b>NOTE</b>: this terms dictionary supports
|
||||
* min/maxItemsPerBlock during indexing to control how
|
||||
* much memory the terms index uses.</p>
|
||||
*
|
||||
* <p>The data structure used by this implementation is very
|
||||
* similar to a burst trie
|
||||
* (http://citeseer.ist.psu.edu/viewdoc/summary?doi=10.1.1.18.3499),
|
||||
* but with added logic to break up too-large blocks of all
|
||||
* terms sharing a given prefix into smaller ones.</p>
|
||||
*
|
||||
* <p>Use {@link org.apache.lucene.index.CheckIndex} with the <code>-verbose</code>
|
||||
* option to see summary statistics on the blocks in the
|
||||
* dictionary.
|
||||
*
|
||||
* See {@link BlockTreeTermsWriter}.
|
||||
*
|
||||
* @lucene.experimental
|
||||
*/
|
||||
|
||||
final class VersionBlockTreeTermsReader extends FieldsProducer {
|
||||
|
||||
// Open input to the main terms dict file (_X.tiv)
|
||||
final IndexInput in;
|
||||
|
||||
//private static final boolean DEBUG = BlockTreeTermsWriter.DEBUG;
|
||||
|
||||
// Reads the terms dict entries, to gather state to
|
||||
// produce DocsEnum on demand
|
||||
final PostingsReaderBase postingsReader;
|
||||
|
||||
private final TreeMap<String,VersionFieldReader> fields = new TreeMap<>();
|
||||
|
||||
/** File offset where the directory starts in the terms file. */
|
||||
private long dirOffset;
|
||||
|
||||
/** File offset where the directory starts in the index file. */
|
||||
private long indexDirOffset;
|
||||
|
||||
final String segment;
|
||||
|
||||
private final int version;
|
||||
|
||||
/** Sole constructor. */
|
||||
public VersionBlockTreeTermsReader(Directory dir, FieldInfos fieldInfos, SegmentInfo info,
|
||||
PostingsReaderBase postingsReader, IOContext ioContext,
|
||||
String segmentSuffix)
|
||||
throws IOException {
|
||||
|
||||
this.postingsReader = postingsReader;
|
||||
|
||||
this.segment = info.name;
|
||||
in = dir.openInput(IndexFileNames.segmentFileName(segment, segmentSuffix, VersionBlockTreeTermsWriter.TERMS_EXTENSION),
|
||||
ioContext);
|
||||
|
||||
boolean success = false;
|
||||
IndexInput indexIn = null;
|
||||
|
||||
try {
|
||||
version = readHeader(in);
|
||||
indexIn = dir.openInput(IndexFileNames.segmentFileName(segment, segmentSuffix, VersionBlockTreeTermsWriter.TERMS_INDEX_EXTENSION),
|
||||
ioContext);
|
||||
int indexVersion = readIndexHeader(indexIn);
|
||||
if (indexVersion != version) {
|
||||
throw new CorruptIndexException("mixmatched version files: " + in + "=" + version + "," + indexIn + "=" + indexVersion);
|
||||
}
|
||||
|
||||
// verify
|
||||
if (version >= VersionBlockTreeTermsWriter.VERSION_CHECKSUM) {
|
||||
CodecUtil.checksumEntireFile(indexIn);
|
||||
}
|
||||
|
||||
// Have PostingsReader init itself
|
||||
postingsReader.init(in);
|
||||
|
||||
// Read per-field details
|
||||
seekDir(in, dirOffset);
|
||||
seekDir(indexIn, indexDirOffset);
|
||||
|
||||
final int numFields = in.readVInt();
|
||||
if (numFields < 0) {
|
||||
throw new CorruptIndexException("invalid numFields: " + numFields + " (resource=" + in + ")");
|
||||
}
|
||||
|
||||
PairOutputs<BytesRef,Long> fstOutputs = VersionBlockTreeTermsWriter.getFSTOutputs();
|
||||
|
||||
for(int i=0;i<numFields;i++) {
|
||||
final int field = in.readVInt();
|
||||
final long numTerms = in.readVLong();
|
||||
assert numTerms >= 0;
|
||||
final int numBytes = in.readVInt();
|
||||
final BytesRef code = new BytesRef(new byte[numBytes]);
|
||||
in.readBytes(code.bytes, 0, numBytes);
|
||||
code.length = numBytes;
|
||||
final long version = in.readVLong();
|
||||
final Pair<BytesRef,Long> rootCode = fstOutputs.newPair(code, version);
|
||||
final FieldInfo fieldInfo = fieldInfos.fieldInfo(field);
|
||||
assert fieldInfo != null: "field=" + field;
|
||||
final long sumTotalTermFreq = fieldInfo.getIndexOptions() == IndexOptions.DOCS_ONLY ? -1 : in.readVLong();
|
||||
final long sumDocFreq = in.readVLong();
|
||||
final int docCount = in.readVInt();
|
||||
final int longsSize = version >= VersionBlockTreeTermsWriter.VERSION_META_ARRAY ? in.readVInt() : 0;
|
||||
|
||||
BytesRef minTerm, maxTerm;
|
||||
if (version >= VersionBlockTreeTermsWriter.VERSION_MIN_MAX_TERMS) {
|
||||
minTerm = readBytesRef(in);
|
||||
maxTerm = readBytesRef(in);
|
||||
} else {
|
||||
minTerm = maxTerm = null;
|
||||
}
|
||||
if (docCount < 0 || docCount > info.getDocCount()) { // #docs with field must be <= #docs
|
||||
throw new CorruptIndexException("invalid docCount: " + docCount + " maxDoc: " + info.getDocCount() + " (resource=" + in + ")");
|
||||
}
|
||||
if (sumDocFreq < docCount) { // #postings must be >= #docs with field
|
||||
throw new CorruptIndexException("invalid sumDocFreq: " + sumDocFreq + " docCount: " + docCount + " (resource=" + in + ")");
|
||||
}
|
||||
if (sumTotalTermFreq != -1 && sumTotalTermFreq < sumDocFreq) { // #positions must be >= #postings
|
||||
throw new CorruptIndexException("invalid sumTotalTermFreq: " + sumTotalTermFreq + " sumDocFreq: " + sumDocFreq + " (resource=" + in + ")");
|
||||
}
|
||||
final long indexStartFP = indexIn.readVLong();
|
||||
VersionFieldReader previous = fields.put(fieldInfo.name,
|
||||
new VersionFieldReader(this, fieldInfo, numTerms, rootCode, sumTotalTermFreq, sumDocFreq, docCount,
|
||||
indexStartFP, longsSize, indexIn, minTerm, maxTerm));
|
||||
if (previous != null) {
|
||||
throw new CorruptIndexException("duplicate field: " + fieldInfo.name + " (resource=" + in + ")");
|
||||
}
|
||||
}
|
||||
indexIn.close();
|
||||
|
||||
success = true;
|
||||
} finally {
|
||||
if (!success) {
|
||||
// this.close() will close in:
|
||||
IOUtils.closeWhileHandlingException(indexIn, this);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private static BytesRef readBytesRef(IndexInput in) throws IOException {
|
||||
BytesRef bytes = new BytesRef();
|
||||
bytes.length = in.readVInt();
|
||||
bytes.bytes = new byte[bytes.length];
|
||||
in.readBytes(bytes.bytes, 0, bytes.length);
|
||||
return bytes;
|
||||
}
|
||||
|
||||
/** Reads terms file header. */
|
||||
private int readHeader(IndexInput input) throws IOException {
|
||||
int version = CodecUtil.checkHeader(input, VersionBlockTreeTermsWriter.TERMS_CODEC_NAME,
|
||||
VersionBlockTreeTermsWriter.VERSION_START,
|
||||
VersionBlockTreeTermsWriter.VERSION_CURRENT);
|
||||
if (version < VersionBlockTreeTermsWriter.VERSION_APPEND_ONLY) {
|
||||
dirOffset = input.readLong();
|
||||
}
|
||||
return version;
|
||||
}
|
||||
|
||||
/** Reads index file header. */
|
||||
private int readIndexHeader(IndexInput input) throws IOException {
|
||||
int version = CodecUtil.checkHeader(input, VersionBlockTreeTermsWriter.TERMS_INDEX_CODEC_NAME,
|
||||
VersionBlockTreeTermsWriter.VERSION_START,
|
||||
VersionBlockTreeTermsWriter.VERSION_CURRENT);
|
||||
if (version < VersionBlockTreeTermsWriter.VERSION_APPEND_ONLY) {
|
||||
indexDirOffset = input.readLong();
|
||||
}
|
||||
return version;
|
||||
}
|
||||
|
||||
/** Seek {@code input} to the directory offset. */
|
||||
private void seekDir(IndexInput input, long dirOffset)
|
||||
throws IOException {
|
||||
if (version >= VersionBlockTreeTermsWriter.VERSION_CHECKSUM) {
|
||||
input.seek(input.length() - CodecUtil.footerLength() - 8);
|
||||
dirOffset = input.readLong();
|
||||
} else if (version >= VersionBlockTreeTermsWriter.VERSION_APPEND_ONLY) {
|
||||
input.seek(input.length() - 8);
|
||||
dirOffset = input.readLong();
|
||||
}
|
||||
input.seek(dirOffset);
|
||||
}
|
||||
|
||||
// for debugging
|
||||
// private static String toHex(int v) {
|
||||
// return "0x" + Integer.toHexString(v);
|
||||
// }
|
||||
|
||||
@Override
|
||||
public void close() throws IOException {
|
||||
try {
|
||||
IOUtils.close(in, postingsReader);
|
||||
} finally {
|
||||
// Clear so refs to terms index is GCable even if
|
||||
// app hangs onto us:
|
||||
fields.clear();
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public Iterator<String> iterator() {
|
||||
return Collections.unmodifiableSet(fields.keySet()).iterator();
|
||||
}
|
||||
|
||||
@Override
|
||||
public Terms terms(String field) throws IOException {
|
||||
assert field != null;
|
||||
return fields.get(field);
|
||||
}
|
||||
|
||||
@Override
|
||||
public int size() {
|
||||
return fields.size();
|
||||
}
|
||||
|
||||
// for debugging
|
||||
String brToString(BytesRef b) {
|
||||
if (b == null) {
|
||||
return "null";
|
||||
} else {
|
||||
try {
|
||||
return b.utf8ToString() + " " + b;
|
||||
} catch (Throwable t) {
|
||||
// If BytesRef isn't actually UTF8, or it's eg a
|
||||
// prefix of UTF8 that ends mid-unicode-char, we
|
||||
// fallback to hex:
|
||||
return b.toString();
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public long ramBytesUsed() {
|
||||
long sizeInByes = ((postingsReader!=null) ? postingsReader.ramBytesUsed() : 0);
|
||||
for(VersionFieldReader reader : fields.values()) {
|
||||
sizeInByes += reader.ramBytesUsed();
|
||||
}
|
||||
return sizeInByes;
|
||||
}
|
||||
|
||||
@Override
|
||||
public void checkIntegrity() throws IOException {
|
||||
if (version >= VersionBlockTreeTermsWriter.VERSION_CHECKSUM) {
|
||||
// term dictionary
|
||||
CodecUtil.checksumEntireFile(in);
|
||||
|
||||
// postings
|
||||
postingsReader.checkIntegrity();
|
||||
}
|
||||
}
|
||||
}
|
||||
File diff suppressed because it is too large
Load Diff
|
|
@ -0,0 +1,163 @@
|
|||
package org.apache.lucene.codecs.idversion;
|
||||
|
||||
/*
|
||||
* 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.index.FieldInfo.IndexOptions;
|
||||
import org.apache.lucene.index.FieldInfo;
|
||||
import org.apache.lucene.index.Terms;
|
||||
import org.apache.lucene.index.TermsEnum;
|
||||
import org.apache.lucene.store.ByteArrayDataInput;
|
||||
import org.apache.lucene.store.IndexInput;
|
||||
import org.apache.lucene.util.BytesRef;
|
||||
import org.apache.lucene.util.automaton.CompiledAutomaton;
|
||||
import org.apache.lucene.util.fst.ByteSequenceOutputs;
|
||||
import org.apache.lucene.util.fst.FST;
|
||||
import org.apache.lucene.util.fst.PairOutputs.Pair;
|
||||
|
||||
/** BlockTree's implementation of {@link Terms}. */
|
||||
// public for CheckIndex:
|
||||
final class VersionFieldReader extends Terms {
|
||||
final long numTerms;
|
||||
final FieldInfo fieldInfo;
|
||||
final long sumTotalTermFreq;
|
||||
final long sumDocFreq;
|
||||
final int docCount;
|
||||
final long indexStartFP;
|
||||
final long rootBlockFP;
|
||||
final Pair<BytesRef,Long> rootCode;
|
||||
final BytesRef minTerm;
|
||||
final BytesRef maxTerm;
|
||||
final int longsSize;
|
||||
final VersionBlockTreeTermsReader parent;
|
||||
|
||||
final FST<Pair<BytesRef,Long>> index;
|
||||
//private boolean DEBUG;
|
||||
|
||||
VersionFieldReader(VersionBlockTreeTermsReader parent, FieldInfo fieldInfo, long numTerms, Pair<BytesRef,Long> rootCode, long sumTotalTermFreq, long sumDocFreq, int docCount,
|
||||
long indexStartFP, int longsSize, IndexInput indexIn, BytesRef minTerm, BytesRef maxTerm) throws IOException {
|
||||
assert numTerms > 0;
|
||||
this.fieldInfo = fieldInfo;
|
||||
//DEBUG = BlockTreeTermsReader.DEBUG && fieldInfo.name.equals("id");
|
||||
this.parent = parent;
|
||||
this.numTerms = numTerms;
|
||||
this.sumTotalTermFreq = sumTotalTermFreq;
|
||||
this.sumDocFreq = sumDocFreq;
|
||||
this.docCount = docCount;
|
||||
this.indexStartFP = indexStartFP;
|
||||
this.rootCode = rootCode;
|
||||
this.longsSize = longsSize;
|
||||
this.minTerm = minTerm;
|
||||
this.maxTerm = maxTerm;
|
||||
// if (DEBUG) {
|
||||
// System.out.println("BTTR: seg=" + segment + " field=" + fieldInfo.name + " rootBlockCode=" + rootCode + " divisor=" + indexDivisor);
|
||||
// }
|
||||
|
||||
rootBlockFP = (new ByteArrayDataInput(rootCode.output1.bytes, rootCode.output1.offset, rootCode.output1.length)).readVLong() >>> VersionBlockTreeTermsWriter.OUTPUT_FLAGS_NUM_BITS;
|
||||
|
||||
if (indexIn != null) {
|
||||
final IndexInput clone = indexIn.clone();
|
||||
//System.out.println("start=" + indexStartFP + " field=" + fieldInfo.name);
|
||||
clone.seek(indexStartFP);
|
||||
index = new FST<>(clone, VersionBlockTreeTermsWriter.getFSTOutputs());
|
||||
|
||||
/*
|
||||
if (false) {
|
||||
final String dotFileName = segment + "_" + fieldInfo.name + ".dot";
|
||||
Writer w = new OutputStreamWriter(new FileOutputStream(dotFileName));
|
||||
Util.toDot(index, w, false, false);
|
||||
System.out.println("FST INDEX: SAVED to " + dotFileName);
|
||||
w.close();
|
||||
}
|
||||
*/
|
||||
} else {
|
||||
index = null;
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public BytesRef getMin() throws IOException {
|
||||
if (minTerm == null) {
|
||||
// Older index that didn't store min/maxTerm
|
||||
return super.getMin();
|
||||
} else {
|
||||
return minTerm;
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public BytesRef getMax() throws IOException {
|
||||
if (maxTerm == null) {
|
||||
// Older index that didn't store min/maxTerm
|
||||
return super.getMax();
|
||||
} else {
|
||||
return maxTerm;
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public boolean hasFreqs() {
|
||||
return fieldInfo.getIndexOptions().compareTo(IndexOptions.DOCS_AND_FREQS) >= 0;
|
||||
}
|
||||
|
||||
@Override
|
||||
public boolean hasOffsets() {
|
||||
return fieldInfo.getIndexOptions().compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS) >= 0;
|
||||
}
|
||||
|
||||
@Override
|
||||
public boolean hasPositions() {
|
||||
return fieldInfo.getIndexOptions().compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) >= 0;
|
||||
}
|
||||
|
||||
@Override
|
||||
public boolean hasPayloads() {
|
||||
return fieldInfo.hasPayloads();
|
||||
}
|
||||
|
||||
@Override
|
||||
public TermsEnum iterator(TermsEnum reuse) throws IOException {
|
||||
return new IDVersionSegmentTermsEnum(this);
|
||||
}
|
||||
|
||||
@Override
|
||||
public long size() {
|
||||
return numTerms;
|
||||
}
|
||||
|
||||
@Override
|
||||
public long getSumTotalTermFreq() {
|
||||
return sumTotalTermFreq;
|
||||
}
|
||||
|
||||
@Override
|
||||
public long getSumDocFreq() {
|
||||
return sumDocFreq;
|
||||
}
|
||||
|
||||
@Override
|
||||
public int getDocCount() {
|
||||
return docCount;
|
||||
}
|
||||
|
||||
/** Returns approximate RAM bytes used */
|
||||
public long ramBytesUsed() {
|
||||
return ((index!=null)? index.sizeInBytes() : 0);
|
||||
}
|
||||
}
|
||||
|
|
@ -86,7 +86,7 @@ import org.apache.lucene.util.fst.Util;
|
|||
* @lucene.experimental
|
||||
*/
|
||||
|
||||
public class BlockTreeTermsReader extends FieldsProducer {
|
||||
public final class BlockTreeTermsReader extends FieldsProducer {
|
||||
|
||||
// Open input to the main terms dict file (_X.tib)
|
||||
final IndexInput in;
|
||||
|
|
|
|||
|
|
@ -187,7 +187,7 @@ import org.apache.lucene.util.packed.PackedInts;
|
|||
* @see BlockTreeTermsReader
|
||||
* @lucene.experimental
|
||||
*/
|
||||
public class BlockTreeTermsWriter extends FieldsConsumer {
|
||||
public final class BlockTreeTermsWriter extends FieldsConsumer {
|
||||
|
||||
/** Suggested default value for the {@code
|
||||
* minItemsInBlock} parameter to {@link
|
||||
|
|
@ -1045,12 +1045,12 @@ public class BlockTreeTermsWriter extends FieldsConsumer {
|
|||
// terms into "good" blocks; we don't save the
|
||||
// resulting FST:
|
||||
blockBuilder = new Builder<>(FST.INPUT_TYPE.BYTE1,
|
||||
0, 0, true,
|
||||
true, Integer.MAX_VALUE,
|
||||
noOutputs,
|
||||
new FindBlocks(), false,
|
||||
PackedInts.COMPACT,
|
||||
true, 15);
|
||||
0, 0, true,
|
||||
true, Integer.MAX_VALUE,
|
||||
noOutputs,
|
||||
new FindBlocks(), false,
|
||||
PackedInts.COMPACT,
|
||||
true, 15);
|
||||
|
||||
this.longsSize = postingsWriter.setField(fieldInfo);
|
||||
}
|
||||
|
|
|
|||
Loading…
Reference in New Issue