Apache
/
lucene
mirror of https://github.com/apache/lucene.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

cut over these collectors 

git-svn-id: https://svn.apache.org/repos/asf/lucene/dev/branches/lucene4765@1444887 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Robert Muir 2013-02-11 17:16:49 +00:00
parent d74fda14f2
commit f2798476cd
1 changed files with 39 additions and 79 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

118
lucene/join/src/java/org/apache/lucene/search/join/TermsWithScoreCollector.java
View File

@ -21,8 +21,7 @@ import java.io.IOException;
import org.apache.lucene.index.AtomicReaderContext; import org.apache.lucene.index.AtomicReaderContext;
import org.apache.lucene.index.BinaryDocValues; import org.apache.lucene.index.BinaryDocValues;
import org.apache.lucene.index.DocTermOrds; import org.apache.lucene.index.SortedSetDocValues;
import org.apache.lucene.index.TermsEnum;
import org.apache.lucene.search.Collector; import org.apache.lucene.search.Collector;
import org.apache.lucene.search.FieldCache; import org.apache.lucene.search.FieldCache;
import org.apache.lucene.search.Scorer; import org.apache.lucene.search.Scorer;
@ -181,9 +180,8 @@ abstract class TermsWithScoreCollector extends Collector {
// impl that works with multiple values per document // impl that works with multiple values per document
static class MV extends TermsWithScoreCollector { static class MV extends TermsWithScoreCollector {
DocTermOrds fromDocTermOrds; SortedSetDocValues fromDocTermOrds;
TermsEnum docTermsEnum; final BytesRef scratch = new BytesRef();
DocTermOrds.TermOrdsIterator reuse;
MV(String field, ScoreMode scoreMode) { MV(String field, ScoreMode scoreMode) {
super(field, scoreMode); super(field, scoreMode);
@ -191,54 +189,33 @@ abstract class TermsWithScoreCollector extends Collector {
@Override @Override
public void collect(int doc) throws IOException { public void collect(int doc) throws IOException {
reuse = fromDocTermOrds.lookup(doc, reuse); fromDocTermOrds.setDocument(doc);
int[] buffer = new int[5]; long ord;
while ((ord = fromDocTermOrds.nextOrd()) != SortedSetDocValues.NO_MORE_ORDS) {
int chunk; fromDocTermOrds.lookupOrd(ord, scratch);
do {
chunk = reuse.read(buffer); int termID = collectedTerms.add(scratch);
if (chunk == 0) { if (termID < 0) {
return; termID = -termID - 1;
} } else {
if (termID >= scoreSums.length) {
for (int idx = 0; idx < chunk; idx++) { scoreSums = ArrayUtil.grow(scoreSums);
int key = buffer[idx];
docTermsEnum.seekExact((long) key);
int ord = collectedTerms.add(docTermsEnum.term());
if (ord < 0) {
ord = -ord - 1;
} else {
if (ord >= scoreSums.length) {
scoreSums = ArrayUtil.grow(scoreSums);
}
}
final float current = scorer.score();
final float existing = scoreSums[ord];
if (Float.compare(existing, 0.0f) == 0) {
scoreSums[ord] = current;
} else {
switch (scoreMode) {
case Total:
scoreSums[ord] = existing + current;
break;
case Max:
if (current > existing) {
scoreSums[ord] = current;
}
}
} }
} }
} while (chunk >= buffer.length);
switch (scoreMode) {
case Total:
scoreSums[termID] += scorer.score();
break;
case Max:
scoreSums[termID] = Math.max(scoreSums[termID], scorer.score());
}
}
} }
@Override @Override
public void setNextReader(AtomicReaderContext context) throws IOException { public void setNextReader(AtomicReaderContext context) throws IOException {
// nocommit: cut over fromDocTermOrds = FieldCache.DEFAULT.getDocTermOrds(context.reader(), field);
DocTermOrds.Iterator iterator = (DocTermOrds.Iterator) FieldCache.DEFAULT.getDocTermOrds(context.reader(), field);
fromDocTermOrds = iterator.getParent();
docTermsEnum = fromDocTermOrds.getOrdTermsEnum(context.reader());
reuse = null; // LUCENE-3377 needs to be fixed first then this statement can be removed...
} }
static class Avg extends MV { static class Avg extends MV {
@ -251,40 +228,23 @@ abstract class TermsWithScoreCollector extends Collector {
@Override @Override
public void collect(int doc) throws IOException { public void collect(int doc) throws IOException {
reuse = fromDocTermOrds.lookup(doc, reuse); fromDocTermOrds.setDocument(doc);
int[] buffer = new int[5]; long ord;
while ((ord = fromDocTermOrds.nextOrd()) != SortedSetDocValues.NO_MORE_ORDS) {
int chunk; fromDocTermOrds.lookupOrd(ord, scratch);
do {
chunk = reuse.read(buffer); int termID = collectedTerms.add(scratch);
if (chunk == 0) { if (termID < 0) {
return; termID = -termID - 1;
} } else {
if (termID >= scoreSums.length) {
for (int idx = 0; idx < chunk; idx++) { scoreSums = ArrayUtil.grow(scoreSums);
int key = buffer[idx];
docTermsEnum.seekExact((long) key);
int ord = collectedTerms.add(docTermsEnum.term());
if (ord < 0) {
ord = -ord - 1;
} else {
if (ord >= scoreSums.length) {
scoreSums = ArrayUtil.grow(scoreSums);
scoreCounts = ArrayUtil.grow(scoreCounts);
}
}
float current = scorer.score();
float existing = scoreSums[ord];
if (Float.compare(existing, 0.0f) == 0) {
scoreSums[ord] = current;
scoreCounts[ord] = 1;
} else {
scoreSums[ord] = scoreSums[ord] + current;
scoreCounts[ord]++;
} }
} }
} while (chunk >= buffer.length);
scoreSums[termID] += scorer.score();
scoreCounts[termID]++;
}
} }
@Override @Override