use of org.apache.lucene.document.FloatPoint in project lucene-solr by apache.
the class JoinUtil method createJoinQuery.
/**
* Method for query time joining for numeric fields. It supports multi- and single- values longs, ints, floats and longs.
* All considerations from {@link JoinUtil#createJoinQuery(String, boolean, String, Query, IndexSearcher, ScoreMode)} are applicable here too,
* though memory consumption might be higher.
* <p>
*
* @param fromField The from field to join from
* @param multipleValuesPerDocument Whether the from field has multiple terms per document
* when true fromField might be {@link DocValuesType#SORTED_NUMERIC},
* otherwise fromField should be {@link DocValuesType#NUMERIC}
* @param toField The to field to join to, should be {@link IntPoint}, {@link LongPoint}, {@link FloatPoint}
* or {@link DoublePoint}.
* @param numericType either {@link java.lang.Integer}, {@link java.lang.Long}, {@link java.lang.Float}
* or {@link java.lang.Double} it should correspond to toField types
* @param fromQuery The query to match documents on the from side
* @param fromSearcher The searcher that executed the specified fromQuery
* @param scoreMode Instructs how scores from the fromQuery are mapped to the returned query
* @return a {@link Query} instance that can be used to join documents based on the
* terms in the from and to field
* @throws IOException If I/O related errors occur
*/
public static Query createJoinQuery(String fromField, boolean multipleValuesPerDocument, String toField, Class<? extends Number> numericType, Query fromQuery, IndexSearcher fromSearcher, ScoreMode scoreMode) throws IOException {
TreeSet<Long> joinValues = new TreeSet<>();
Map<Long, Float> aggregatedScores = new HashMap<>();
Map<Long, Integer> occurrences = new HashMap<>();
boolean needsScore = scoreMode != ScoreMode.None;
BiConsumer<Long, Float> scoreAggregator;
if (scoreMode == ScoreMode.Max) {
scoreAggregator = (key, score) -> {
Float currentValue = aggregatedScores.putIfAbsent(key, score);
if (currentValue != null) {
aggregatedScores.put(key, Math.max(currentValue, score));
}
};
} else if (scoreMode == ScoreMode.Min) {
scoreAggregator = (key, score) -> {
Float currentValue = aggregatedScores.putIfAbsent(key, score);
if (currentValue != null) {
aggregatedScores.put(key, Math.min(currentValue, score));
}
};
} else if (scoreMode == ScoreMode.Total) {
scoreAggregator = (key, score) -> {
Float currentValue = aggregatedScores.putIfAbsent(key, score);
if (currentValue != null) {
aggregatedScores.put(key, currentValue + score);
}
};
} else if (scoreMode == ScoreMode.Avg) {
scoreAggregator = (key, score) -> {
Float currentSore = aggregatedScores.putIfAbsent(key, score);
if (currentSore != null) {
aggregatedScores.put(key, currentSore + score);
}
Integer currentOccurrence = occurrences.putIfAbsent(key, 1);
if (currentOccurrence != null) {
occurrences.put(key, ++currentOccurrence);
}
};
} else {
scoreAggregator = (key, score) -> {
throw new UnsupportedOperationException();
};
}
LongFunction<Float> joinScorer;
if (scoreMode == ScoreMode.Avg) {
joinScorer = (joinValue) -> {
Float aggregatedScore = aggregatedScores.get(joinValue);
Integer occurrence = occurrences.get(joinValue);
return aggregatedScore / occurrence;
};
} else {
joinScorer = aggregatedScores::get;
}
Collector collector;
if (multipleValuesPerDocument) {
collector = new SimpleCollector() {
SortedNumericDocValues sortedNumericDocValues;
Scorer scorer;
@Override
public void collect(int doc) throws IOException {
if (doc > sortedNumericDocValues.docID()) {
sortedNumericDocValues.advance(doc);
}
if (doc == sortedNumericDocValues.docID()) {
for (int i = 0; i < sortedNumericDocValues.docValueCount(); i++) {
long value = sortedNumericDocValues.nextValue();
joinValues.add(value);
if (needsScore) {
scoreAggregator.accept(value, scorer.score());
}
}
}
}
@Override
protected void doSetNextReader(LeafReaderContext context) throws IOException {
sortedNumericDocValues = DocValues.getSortedNumeric(context.reader(), fromField);
}
@Override
public void setScorer(Scorer scorer) throws IOException {
this.scorer = scorer;
}
@Override
public boolean needsScores() {
return needsScore;
}
};
} else {
collector = new SimpleCollector() {
NumericDocValues numericDocValues;
Scorer scorer;
private int lastDocID = -1;
private boolean docsInOrder(int docID) {
if (docID < lastDocID) {
throw new AssertionError("docs out of order: lastDocID=" + lastDocID + " vs docID=" + docID);
}
lastDocID = docID;
return true;
}
@Override
public void collect(int doc) throws IOException {
assert docsInOrder(doc);
int dvDocID = numericDocValues.docID();
if (dvDocID < doc) {
dvDocID = numericDocValues.advance(doc);
}
long value;
if (dvDocID == doc) {
value = numericDocValues.longValue();
} else {
value = 0;
}
joinValues.add(value);
if (needsScore) {
scoreAggregator.accept(value, scorer.score());
}
}
@Override
protected void doSetNextReader(LeafReaderContext context) throws IOException {
numericDocValues = DocValues.getNumeric(context.reader(), fromField);
lastDocID = -1;
}
@Override
public void setScorer(Scorer scorer) throws IOException {
this.scorer = scorer;
}
@Override
public boolean needsScores() {
return needsScore;
}
};
}
fromSearcher.search(fromQuery, collector);
Iterator<Long> iterator = joinValues.iterator();
final int bytesPerDim;
final BytesRef encoded = new BytesRef();
final PointInSetIncludingScoreQuery.Stream stream;
if (Integer.class.equals(numericType)) {
bytesPerDim = Integer.BYTES;
stream = new PointInSetIncludingScoreQuery.Stream() {
@Override
public BytesRef next() {
if (iterator.hasNext()) {
long value = iterator.next();
IntPoint.encodeDimension((int) value, encoded.bytes, 0);
if (needsScore) {
score = joinScorer.apply(value);
}
return encoded;
} else {
return null;
}
}
};
} else if (Long.class.equals(numericType)) {
bytesPerDim = Long.BYTES;
stream = new PointInSetIncludingScoreQuery.Stream() {
@Override
public BytesRef next() {
if (iterator.hasNext()) {
long value = iterator.next();
LongPoint.encodeDimension(value, encoded.bytes, 0);
if (needsScore) {
score = joinScorer.apply(value);
}
return encoded;
} else {
return null;
}
}
};
} else if (Float.class.equals(numericType)) {
bytesPerDim = Float.BYTES;
stream = new PointInSetIncludingScoreQuery.Stream() {
@Override
public BytesRef next() {
if (iterator.hasNext()) {
long value = iterator.next();
FloatPoint.encodeDimension(Float.intBitsToFloat((int) value), encoded.bytes, 0);
if (needsScore) {
score = joinScorer.apply(value);
}
return encoded;
} else {
return null;
}
}
};
} else if (Double.class.equals(numericType)) {
bytesPerDim = Double.BYTES;
stream = new PointInSetIncludingScoreQuery.Stream() {
@Override
public BytesRef next() {
if (iterator.hasNext()) {
long value = iterator.next();
DoublePoint.encodeDimension(Double.longBitsToDouble(value), encoded.bytes, 0);
if (needsScore) {
score = joinScorer.apply(value);
}
return encoded;
} else {
return null;
}
}
};
} else {
throw new IllegalArgumentException("unsupported numeric type, only Integer, Long, Float and Double are supported");
}
encoded.bytes = new byte[bytesPerDim];
encoded.length = bytesPerDim;
if (needsScore) {
return new PointInSetIncludingScoreQuery(scoreMode, fromQuery, multipleValuesPerDocument, toField, bytesPerDim, stream) {
@Override
protected String toString(byte[] value) {
return toString.apply(value, numericType);
}
};
} else {
return new PointInSetQuery(toField, 1, bytesPerDim, stream) {
@Override
protected String toString(byte[] value) {
return PointInSetIncludingScoreQuery.toString.apply(value, numericType);
}
};
}
}
use of org.apache.lucene.document.FloatPoint in project lucene-solr by apache.
the class TestJoinUtil method addLinkFields.
private void addLinkFields(final Random random, Document document, final String fieldName, String linkValue, boolean multipleValuesPerDocument, boolean globalOrdinalJoin) {
document.add(newTextField(random, fieldName, linkValue, Field.Store.NO));
final int linkInt = Integer.parseUnsignedInt(linkValue, 16);
document.add(new IntPoint(fieldName + "INT", linkInt));
document.add(new FloatPoint(fieldName + "FLOAT", linkInt));
final long linkLong = linkInt << 32 | linkInt;
document.add(new LongPoint(fieldName + "LONG", linkLong));
document.add(new DoublePoint(fieldName + "DOUBLE", linkLong));
if (multipleValuesPerDocument) {
document.add(new SortedSetDocValuesField(fieldName, new BytesRef(linkValue)));
document.add(new SortedNumericDocValuesField(fieldName + "INT", linkInt));
document.add(new SortedNumericDocValuesField(fieldName + "FLOAT", Float.floatToRawIntBits(linkInt)));
document.add(new SortedNumericDocValuesField(fieldName + "LONG", linkLong));
document.add(new SortedNumericDocValuesField(fieldName + "DOUBLE", Double.doubleToRawLongBits(linkLong)));
} else {
document.add(new SortedDocValuesField(fieldName, new BytesRef(linkValue)));
document.add(new NumericDocValuesField(fieldName + "INT", linkInt));
document.add(new FloatDocValuesField(fieldName + "FLOAT", linkInt));
document.add(new NumericDocValuesField(fieldName + "LONG", linkLong));
document.add(new DoubleDocValuesField(fieldName + "DOUBLE", linkLong));
}
if (globalOrdinalJoin) {
document.add(new SortedDocValuesField("join_field", new BytesRef(linkValue)));
}
}
use of org.apache.lucene.document.FloatPoint in project lucene-solr by apache.
the class TestPointValues method testInvalidFloatPointUsage.
public void testInvalidFloatPointUsage() throws Exception {
FloatPoint field = new FloatPoint("field", 17, 42);
expectThrows(IllegalArgumentException.class, () -> {
field.setFloatValue(14);
});
expectThrows(IllegalStateException.class, () -> {
field.numericValue();
});
}
use of org.apache.lucene.document.FloatPoint in project lucene-solr by apache.
the class TestPointQueries method testBasicFloats.
public void testBasicFloats() throws Exception {
Directory dir = newDirectory();
IndexWriter w = new IndexWriter(dir, new IndexWriterConfig(new MockAnalyzer(random())));
Document doc = new Document();
doc.add(new FloatPoint("point", -7.0f));
w.addDocument(doc);
doc = new Document();
doc.add(new FloatPoint("point", 0.0f));
w.addDocument(doc);
doc = new Document();
doc.add(new FloatPoint("point", 3.0f));
w.addDocument(doc);
DirectoryReader r = DirectoryReader.open(w);
IndexSearcher s = new IndexSearcher(r);
assertEquals(2, s.count(FloatPoint.newRangeQuery("point", -8.0f, 1.0f)));
assertEquals(3, s.count(FloatPoint.newRangeQuery("point", -7.0f, 3.0f)));
assertEquals(1, s.count(FloatPoint.newExactQuery("point", -7.0f)));
assertEquals(0, s.count(FloatPoint.newExactQuery("point", -6.0f)));
w.close();
r.close();
dir.close();
}
use of org.apache.lucene.document.FloatPoint in project lucene-solr by apache.
the class TestPointQueries method testEmptyPointInSetQuery.
public void testEmptyPointInSetQuery() throws Exception {
Directory dir = newDirectory();
IndexWriterConfig iwc = newIndexWriterConfig();
iwc.setCodec(getCodec());
IndexWriter w = new IndexWriter(dir, iwc);
Document doc = new Document();
doc.add(new IntPoint("int", 17));
doc.add(new LongPoint("long", 17L));
doc.add(new FloatPoint("float", 17.0f));
doc.add(new DoublePoint("double", 17.0));
doc.add(new BinaryPoint("bytes", new byte[] { 0, 17 }));
w.addDocument(doc);
IndexReader r = DirectoryReader.open(w);
IndexSearcher s = newSearcher(r, false);
assertEquals(0, s.count(IntPoint.newSetQuery("int")));
assertEquals(0, s.count(LongPoint.newSetQuery("long")));
assertEquals(0, s.count(FloatPoint.newSetQuery("float")));
assertEquals(0, s.count(DoublePoint.newSetQuery("double")));
assertEquals(0, s.count(BinaryPoint.newSetQuery("bytes")));
w.close();
r.close();
dir.close();
}
Aggregations