use of org.apache.lucene.document.Document in project elasticsearch by elastic.
the class CollapsingTopDocsCollectorTests method testCollapseString.
public void testCollapseString() throws Exception {
CollapsingDocValuesProducer producer = new CollapsingDocValuesProducer<BytesRef>() {
@Override
public BytesRef randomGroup(int maxGroup) {
return new BytesRef(Integer.toString(randomIntBetween(0, maxGroup - 1)));
}
@Override
public void add(Document doc, BytesRef value, boolean multivalued) {
if (multivalued) {
doc.add(new SortedSetDocValuesField("field", value));
} else {
doc.add(new SortedDocValuesField("field", value));
}
}
@Override
public SortField sortField(boolean multivalued) {
if (multivalued) {
return new SortedSetSortField("field", false);
} else {
return new SortField("field", SortField.Type.STRING_VAL);
}
}
};
assertSearchCollapse(producer, false);
}
use of org.apache.lucene.document.Document in project elasticsearch by elastic.
the class CollapsingTopDocsCollectorTests method testCollapseDouble.
public void testCollapseDouble() throws Exception {
CollapsingDocValuesProducer producer = new CollapsingDocValuesProducer<Double>() {
@Override
public Double randomGroup(int maxGroup) {
return new Double(randomIntBetween(0, maxGroup - 1));
}
@Override
public void add(Document doc, Double value, boolean multivalued) {
if (multivalued) {
doc.add(new SortedNumericDocValuesField("field", NumericUtils.doubleToSortableLong(value)));
} else {
doc.add(new NumericDocValuesField("field", Double.doubleToLongBits(value)));
}
}
@Override
public SortField sortField(boolean multivalued) {
if (multivalued) {
return new SortedNumericSortField("field", SortField.Type.DOUBLE);
} else {
return new SortField("field", SortField.Type.DOUBLE);
}
}
};
assertSearchCollapse(producer, true);
}
use of org.apache.lucene.document.Document in project elasticsearch by elastic.
the class CollapsingTopDocsCollectorTests method assertSearchCollapse.
private <T extends Comparable> void assertSearchCollapse(CollapsingDocValuesProducer<T> dvProducers, boolean numeric, boolean multivalued) throws IOException {
final int numDocs = randomIntBetween(1000, 2000);
int maxGroup = randomIntBetween(2, 500);
final Directory dir = newDirectory();
final RandomIndexWriter w = new RandomIndexWriter(random(), dir);
Set<T> values = new HashSet<>();
int totalHits = 0;
for (int i = 0; i < numDocs; i++) {
final T value = dvProducers.randomGroup(maxGroup);
values.add(value);
Document doc = new Document();
dvProducers.add(doc, value, multivalued);
doc.add(new NumericDocValuesField("sort1", randomIntBetween(0, 10)));
doc.add(new NumericDocValuesField("sort2", randomLong()));
w.addDocument(doc);
totalHits++;
}
List<T> valueList = new ArrayList<>(values);
Collections.sort(valueList);
final IndexReader reader = w.getReader();
final IndexSearcher searcher = newSearcher(reader);
final SortField collapseField = dvProducers.sortField(multivalued);
final SortField sort1 = new SortField("sort1", SortField.Type.INT);
final SortField sort2 = new SortField("sort2", SortField.Type.LONG);
Sort sort = new Sort(sort1, sort2, collapseField);
int expectedNumGroups = values.size();
final CollapsingTopDocsCollector collapsingCollector;
if (numeric) {
collapsingCollector = CollapsingTopDocsCollector.createNumeric(collapseField.getField(), sort, expectedNumGroups, false);
} else {
collapsingCollector = CollapsingTopDocsCollector.createKeyword(collapseField.getField(), sort, expectedNumGroups, false);
}
TopFieldCollector topFieldCollector = TopFieldCollector.create(sort, totalHits, true, false, false);
searcher.search(new MatchAllDocsQuery(), collapsingCollector);
searcher.search(new MatchAllDocsQuery(), topFieldCollector);
CollapseTopFieldDocs collapseTopFieldDocs = collapsingCollector.getTopDocs();
TopFieldDocs topDocs = topFieldCollector.topDocs();
assertEquals(collapseField.getField(), collapseTopFieldDocs.field);
assertEquals(expectedNumGroups, collapseTopFieldDocs.scoreDocs.length);
assertEquals(totalHits, collapseTopFieldDocs.totalHits);
assertEquals(totalHits, topDocs.scoreDocs.length);
assertEquals(totalHits, topDocs.totalHits);
Set<Object> seen = new HashSet<>();
// collapse field is the last sort
int collapseIndex = sort.getSort().length - 1;
int topDocsIndex = 0;
for (int i = 0; i < expectedNumGroups; i++) {
FieldDoc fieldDoc = null;
for (; topDocsIndex < totalHits; topDocsIndex++) {
fieldDoc = (FieldDoc) topDocs.scoreDocs[topDocsIndex];
if (seen.contains(fieldDoc.fields[collapseIndex]) == false) {
break;
}
}
FieldDoc collapseFieldDoc = (FieldDoc) collapseTopFieldDocs.scoreDocs[i];
assertNotNull(fieldDoc);
assertEquals(collapseFieldDoc.doc, fieldDoc.doc);
assertArrayEquals(collapseFieldDoc.fields, fieldDoc.fields);
seen.add(fieldDoc.fields[fieldDoc.fields.length - 1]);
}
for (; topDocsIndex < totalHits; topDocsIndex++) {
FieldDoc fieldDoc = (FieldDoc) topDocs.scoreDocs[topDocsIndex];
assertTrue(seen.contains(fieldDoc.fields[collapseIndex]));
}
// check merge
final IndexReaderContext ctx = searcher.getTopReaderContext();
final SegmentSearcher[] subSearchers;
final int[] docStarts;
if (ctx instanceof LeafReaderContext) {
subSearchers = new SegmentSearcher[1];
docStarts = new int[1];
subSearchers[0] = new SegmentSearcher((LeafReaderContext) ctx, ctx);
docStarts[0] = 0;
} else {
final CompositeReaderContext compCTX = (CompositeReaderContext) ctx;
final int size = compCTX.leaves().size();
subSearchers = new SegmentSearcher[size];
docStarts = new int[size];
int docBase = 0;
for (int searcherIDX = 0; searcherIDX < subSearchers.length; searcherIDX++) {
final LeafReaderContext leave = compCTX.leaves().get(searcherIDX);
subSearchers[searcherIDX] = new SegmentSearcher(leave, compCTX);
docStarts[searcherIDX] = docBase;
docBase += leave.reader().maxDoc();
}
}
final CollapseTopFieldDocs[] shardHits = new CollapseTopFieldDocs[subSearchers.length];
final Weight weight = searcher.createNormalizedWeight(new MatchAllDocsQuery(), false);
for (int shardIDX = 0; shardIDX < subSearchers.length; shardIDX++) {
final SegmentSearcher subSearcher = subSearchers[shardIDX];
final CollapsingTopDocsCollector c;
if (numeric) {
c = CollapsingTopDocsCollector.createNumeric(collapseField.getField(), sort, expectedNumGroups, false);
} else {
c = CollapsingTopDocsCollector.createKeyword(collapseField.getField(), sort, expectedNumGroups, false);
}
subSearcher.search(weight, c);
shardHits[shardIDX] = c.getTopDocs();
}
CollapseTopFieldDocs mergedFieldDocs = CollapseTopFieldDocs.merge(sort, 0, expectedNumGroups, shardHits);
assertTopDocsEquals(mergedFieldDocs, collapseTopFieldDocs);
w.close();
reader.close();
dir.close();
}
use of org.apache.lucene.document.Document in project elasticsearch by elastic.
the class CollapsingTopDocsCollectorTests method testCollapseInt.
public void testCollapseInt() throws Exception {
CollapsingDocValuesProducer producer = new CollapsingDocValuesProducer<Integer>() {
@Override
public Integer randomGroup(int maxGroup) {
return randomIntBetween(0, maxGroup - 1);
}
@Override
public void add(Document doc, Integer value, boolean multivalued) {
if (multivalued) {
doc.add(new SortedNumericDocValuesField("field", value));
} else {
doc.add(new NumericDocValuesField("field", value));
}
}
@Override
public SortField sortField(boolean multivalued) {
if (multivalued) {
return new SortedNumericSortField("field", SortField.Type.INT);
} else {
return new SortField("field", SortField.Type.INT);
}
}
};
assertSearchCollapse(producer, true);
}
use of org.apache.lucene.document.Document in project elasticsearch by elastic.
the class CustomUnifiedHighlighterTests method assertHighlightOneDoc.
private void assertHighlightOneDoc(String fieldName, String[] inputs, Analyzer analyzer, Query query, Locale locale, BreakIterator breakIterator, int noMatchSize, String[] expectedPassages) throws Exception {
Directory dir = newDirectory();
IndexWriterConfig iwc = newIndexWriterConfig(analyzer);
iwc.setMergePolicy(newTieredMergePolicy(random()));
RandomIndexWriter iw = new RandomIndexWriter(random(), dir, iwc);
FieldType ft = new FieldType(TextField.TYPE_STORED);
ft.setIndexOptions(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS_AND_OFFSETS);
ft.freeze();
Document doc = new Document();
for (String input : inputs) {
Field field = new Field(fieldName, "", ft);
field.setStringValue(input);
doc.add(field);
}
iw.addDocument(doc);
DirectoryReader reader = iw.getReader();
IndexSearcher searcher = newSearcher(reader);
iw.close();
TopDocs topDocs = searcher.search(new MatchAllDocsQuery(), 1, Sort.INDEXORDER);
assertThat(topDocs.totalHits, equalTo(1));
String rawValue = Strings.arrayToDelimitedString(inputs, String.valueOf(MULTIVAL_SEP_CHAR));
CustomUnifiedHighlighter highlighter = new CustomUnifiedHighlighter(searcher, analyzer, new CustomPassageFormatter("<b>", "</b>", new DefaultEncoder()), locale, breakIterator, rawValue, noMatchSize);
highlighter.setFieldMatcher((name) -> "text".equals(name));
final Snippet[] snippets = highlighter.highlightField("text", query, topDocs.scoreDocs[0].doc, expectedPassages.length);
assertEquals(snippets.length, expectedPassages.length);
for (int i = 0; i < snippets.length; i++) {
assertEquals(snippets[i].getText(), expectedPassages[i]);
}
reader.close();
dir.close();
}
Aggregations