All Downloads are FREE. Search and download functionalities are using the official Maven repository.
Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance.
Project price only 1 $
You can buy this project and download/modify it how often you want.
org.apache.lucene.codecs.blocktreeords.OrdsIntersectTermsEnum Maven / Gradle / Ivy
/*
* 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.
*/
package org.apache.lucene.codecs.blocktreeords;
import java.io.IOException;
import org.apache.lucene.codecs.blocktreeords.FSTOrdsOutputs.Output;
import org.apache.lucene.index.BaseTermsEnum;
import org.apache.lucene.index.ImpactsEnum;
import org.apache.lucene.index.PostingsEnum;
import org.apache.lucene.index.TermState;
import org.apache.lucene.store.IndexInput;
import org.apache.lucene.util.ArrayUtil;
import org.apache.lucene.util.BytesRef;
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.fst.FST;
// NOTE: cannot seek!
final class OrdsIntersectTermsEnum extends BaseTermsEnum {
final IndexInput in;
private OrdsIntersectTermsEnumFrame[] stack;
@SuppressWarnings({"rawtypes","unchecked"}) private FST.Arc[] arcs = new FST.Arc[5];
final RunAutomaton runAutomaton;
final CompiledAutomaton compiledAutomaton;
private OrdsIntersectTermsEnumFrame currentFrame;
private final BytesRef term = new BytesRef();
private final FST.BytesReader fstReader;
final OrdsFieldReader fr;
private BytesRef savedStartTerm;
// TODO: in some cases we can filter by length? eg
// regexp foo*bar must be at least length 6 bytes
public OrdsIntersectTermsEnum(OrdsFieldReader fr, CompiledAutomaton compiled, BytesRef startTerm) throws IOException {
// if (DEBUG) {
// System.out.println("\nintEnum.init seg=" + segment + " commonSuffix=" + brToString(compiled.commonSuffixRef));
// }
this.fr = fr;
runAutomaton = compiled.runAutomaton;
compiledAutomaton = compiled;
in = fr.parent.in.clone();
stack = new OrdsIntersectTermsEnumFrame[5];
for(int idx=0;idx();
}
if (fr.index == null) {
fstReader = null;
} else {
fstReader = fr.index.getBytesReader();
}
// TODO: if the automaton is "smallish" we really
// should use the terms index to seek at least to
// the initial term and likely to subsequent terms
// (or, maybe just fallback to ATE for such cases).
// Else the seek cost of loading the frames will be
// too costly.
final FST.Arc arc = fr.index.getFirstArc(arcs[0]);
// Empty string prefix must have an output in the index!
assert arc.isFinal();
// Special pushFrame since it's the first one:
final OrdsIntersectTermsEnumFrame f = stack[0];
f.fp = f.fpOrig = fr.rootBlockFP;
f.prefix = 0;
f.setState(0);
f.arc = arc;
f.outputPrefix = arc.output;
f.load(fr.rootCode);
// for assert:
assert setSavedStartTerm(startTerm);
currentFrame = f;
if (startTerm != null) {
seekToStartTerm(startTerm);
}
}
// only for assert:
private boolean setSavedStartTerm(BytesRef startTerm) {
savedStartTerm = startTerm == null ? null : BytesRef.deepCopyOf(startTerm);
return true;
}
@Override
public TermState termState() throws IOException {
currentFrame.decodeMetaData();
return currentFrame.termState.clone();
}
private OrdsIntersectTermsEnumFrame getFrame(int ord) throws IOException {
if (ord >= stack.length) {
final OrdsIntersectTermsEnumFrame[] next = new OrdsIntersectTermsEnumFrame[ArrayUtil.oversize(1+ord, RamUsageEstimator.NUM_BYTES_OBJECT_REF)];
System.arraycopy(stack, 0, next, 0, stack.length);
for(int stackOrd=stack.length;stackOrd getArc(int ord) {
if (ord >= arcs.length) {
@SuppressWarnings({"rawtypes","unchecked"}) final FST.Arc[] 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();
}
arcs = next;
}
return arcs[ord];
}
private OrdsIntersectTermsEnumFrame pushFrame(int state) throws IOException {
final OrdsIntersectTermsEnumFrame f = getFrame(currentFrame == null ? 0 : 1+currentFrame.ord);
f.fp = f.fpOrig = currentFrame.lastSubFP;
f.prefix = currentFrame.prefix + currentFrame.suffix;
// if (DEBUG) System.out.println(" pushFrame state=" + state + " prefix=" + f.prefix);
f.setState(state);
// Walk the arc through the index -- we only
// "bother" with this so we can get the floor data
// from the index and skip floor blocks when
// possible:
FST.Arc arc = currentFrame.arc;
int idx = currentFrame.prefix;
assert currentFrame.suffix > 0;
Output output = currentFrame.outputPrefix;
while (idx < f.prefix) {
final int target = term.bytes[idx] & 0xff;
// TODO: we could be more efficient for the next()
// case by using current arc as starting point,
// passed to findTargetArc
arc = fr.index.findTargetArc(target, arc, getArc(1+idx), fstReader);
assert arc != null;
output = OrdsBlockTreeTermsWriter.FST_OUTPUTS.add(output, arc.output);
idx++;
}
f.arc = arc;
f.outputPrefix = output;
assert arc.isFinal();
f.load(OrdsBlockTreeTermsWriter.FST_OUTPUTS.add(output, arc.nextFinalOutput));
return f;
}
@Override
public BytesRef term() {
return term;
}
// TODO: do we need ord() here? OrdsIntersectTermsEnumFrame tracks termOrd but it may be buggy!
@Override
public int docFreq() throws IOException {
//if (DEBUG) System.out.println("BTIR.docFreq");
currentFrame.decodeMetaData();
//if (DEBUG) System.out.println(" return " + currentFrame.termState.docFreq);
return currentFrame.termState.docFreq;
}
@Override
public long totalTermFreq() throws IOException {
currentFrame.decodeMetaData();
return currentFrame.termState.totalTermFreq;
}
@Override
public PostingsEnum postings(PostingsEnum reuse, int flags) throws IOException {
currentFrame.decodeMetaData();
return fr.parent.postingsReader.postings(fr.fieldInfo, currentFrame.termState, reuse, flags);
}
@Override
public ImpactsEnum impacts(int flags) throws IOException {
currentFrame.decodeMetaData();
return fr.parent.postingsReader.impacts(fr.fieldInfo, currentFrame.termState, flags);
}
private int getState() {
int state = currentFrame.state;
for(int idx=0;idx 0) {
// A prefix of the common suffix overlaps with
// the suffix of the block prefix so we first
// test whether the prefix part matches:
final byte[] termBytes = term.bytes;
int termBytesPos = currentFrame.prefix - lenInPrefix;
assert termBytesPos >= 0;
final int termBytesPosEnd = currentFrame.prefix;
while (termBytesPos < termBytesPosEnd) {
if (termBytes[termBytesPos++] != commonSuffixBytes[commonSuffixBytesPos++]) {
// if (DEBUG) {
// System.out.println(" skip: common suffix mismatch (in prefix)");
// }
continue nextTerm;
}
}
suffixBytesPos = currentFrame.startBytePos;
} else {
suffixBytesPos = currentFrame.startBytePos + currentFrame.suffix - compiledAutomaton.commonSuffixRef.length;
}
// Test overlapping suffix part:
final int commonSuffixBytesPosEnd = compiledAutomaton.commonSuffixRef.length;
while (commonSuffixBytesPos < commonSuffixBytesPosEnd) {
if (suffixBytes[suffixBytesPos++] != commonSuffixBytes[commonSuffixBytesPos++]) {
// if (DEBUG) {
// System.out.println(" skip: common suffix mismatch");
// }
continue nextTerm;
}
}
}
// TODO: maybe we should do the same linear test
// that AutomatonTermsEnum does, so that if we
// reach a part of the automaton where .* is
// "temporarily" accepted, we just blindly .next()
// until the limit
// See if the term prefix matches the automaton:
int state = currentFrame.state;
for (int idx=0;idx-->
