private void executeRandomJoin(
boolean multipleValuesPerDocument,
int maxIndexIter,
int maxSearchIter,
int numberOfDocumentsToIndex)
throws Exception {
for (int indexIter = 1; indexIter <= maxIndexIter; indexIter++) {
if (VERBOSE) {
System.out.println("indexIter=" + indexIter);
}
Directory dir = newDirectory();
RandomIndexWriter w =
new RandomIndexWriter(
random(),
dir,
newIndexWriterConfig(
TEST_VERSION_CURRENT,
new MockAnalyzer(random(), MockTokenizer.KEYWORD, false))
.setMergePolicy(newLogMergePolicy()));
final boolean scoreDocsInOrder = TestJoinUtil.random().nextBoolean();
IndexIterationContext context =
createContext(numberOfDocumentsToIndex, w, multipleValuesPerDocument, scoreDocsInOrder);
IndexReader topLevelReader = w.getReader();
w.close();
for (int searchIter = 1; searchIter <= maxSearchIter; searchIter++) {
if (VERBOSE) {
System.out.println("searchIter=" + searchIter);
}
IndexSearcher indexSearcher = newSearcher(topLevelReader);
int r = random().nextInt(context.randomUniqueValues.length);
boolean from = context.randomFrom[r];
String randomValue = context.randomUniqueValues[r];
FixedBitSet expectedResult =
createExpectedResult(randomValue, from, indexSearcher.getIndexReader(), context);
final Query actualQuery = new TermQuery(new Term("value", randomValue));
if (VERBOSE) {
System.out.println("actualQuery=" + actualQuery);
}
final ScoreMode scoreMode = ScoreMode.values()[random().nextInt(ScoreMode.values().length)];
if (VERBOSE) {
System.out.println("scoreMode=" + scoreMode);
}
final Query joinQuery;
if (from) {
joinQuery =
JoinUtil.createJoinQuery(
"from", multipleValuesPerDocument, "to", actualQuery, indexSearcher, scoreMode);
} else {
joinQuery =
JoinUtil.createJoinQuery(
"to", multipleValuesPerDocument, "from", actualQuery, indexSearcher, scoreMode);
}
if (VERBOSE) {
System.out.println("joinQuery=" + joinQuery);
}
// Need to know all documents that have matches. TopDocs doesn't give me that and then I'd
// be also testing TopDocsCollector...
final FixedBitSet actualResult = new FixedBitSet(indexSearcher.getIndexReader().maxDoc());
final TopScoreDocCollector topScoreDocCollector = TopScoreDocCollector.create(10, false);
indexSearcher.search(
joinQuery,
new Collector() {
int docBase;
@Override
public void collect(int doc) throws IOException {
actualResult.set(doc + docBase);
topScoreDocCollector.collect(doc);
}
@Override
public void setNextReader(AtomicReaderContext context) {
docBase = context.docBase;
topScoreDocCollector.setNextReader(context);
}
@Override
public void setScorer(Scorer scorer) throws IOException {
topScoreDocCollector.setScorer(scorer);
}
@Override
public boolean acceptsDocsOutOfOrder() {
return scoreDocsInOrder;
}
});
// Asserting bit set...
if (VERBOSE) {
System.out.println("expected cardinality:" + expectedResult.cardinality());
DocIdSetIterator iterator = expectedResult.iterator();
for (int doc = iterator.nextDoc();
doc != DocIdSetIterator.NO_MORE_DOCS;
doc = iterator.nextDoc()) {
System.out.println(
String.format(
Locale.ROOT,
"Expected doc[%d] with id value %s",
doc,
indexSearcher.doc(doc).get("id")));
}
System.out.println("actual cardinality:" + actualResult.cardinality());
iterator = actualResult.iterator();
for (int doc = iterator.nextDoc();
doc != DocIdSetIterator.NO_MORE_DOCS;
doc = iterator.nextDoc()) {
System.out.println(
String.format(
Locale.ROOT,
"Actual doc[%d] with id value %s",
doc,
indexSearcher.doc(doc).get("id")));
}
}
assertEquals(expectedResult, actualResult);
// Asserting TopDocs...
TopDocs expectedTopDocs = createExpectedTopDocs(randomValue, from, scoreMode, context);
TopDocs actualTopDocs = topScoreDocCollector.topDocs();
assertEquals(expectedTopDocs.totalHits, actualTopDocs.totalHits);
assertEquals(expectedTopDocs.scoreDocs.length, actualTopDocs.scoreDocs.length);
if (scoreMode == ScoreMode.None) {
continue;
}
assertEquals(expectedTopDocs.getMaxScore(), actualTopDocs.getMaxScore(), 0.0f);
for (int i = 0; i < expectedTopDocs.scoreDocs.length; i++) {
if (VERBOSE) {
System.out.printf(
Locale.ENGLISH,
"Expected doc: %d | Actual doc: %d\n",
expectedTopDocs.scoreDocs[i].doc,
actualTopDocs.scoreDocs[i].doc);
System.out.printf(
Locale.ENGLISH,
"Expected score: %f | Actual score: %f\n",
expectedTopDocs.scoreDocs[i].score,
actualTopDocs.scoreDocs[i].score);
}
assertEquals(expectedTopDocs.scoreDocs[i].doc, actualTopDocs.scoreDocs[i].doc);
assertEquals(expectedTopDocs.scoreDocs[i].score, actualTopDocs.scoreDocs[i].score, 0.0f);
Explanation explanation =
indexSearcher.explain(joinQuery, expectedTopDocs.scoreDocs[i].doc);
assertEquals(expectedTopDocs.scoreDocs[i].score, explanation.getValue(), 0.0f);
}
}
topLevelReader.close();
dir.close();
}
}
public void testSimple() throws Exception {
final String idField = "id";
final String toField = "productId";
Directory dir = newDirectory();
RandomIndexWriter w =
new RandomIndexWriter(
random(),
dir,
newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random()))
.setMergePolicy(newLogMergePolicy()));
// 0
Document doc = new Document();
doc.add(new TextField("description", "random text", Field.Store.NO));
doc.add(new TextField("name", "name1", Field.Store.NO));
doc.add(new TextField(idField, "1", Field.Store.NO));
w.addDocument(doc);
// 1
doc = new Document();
doc.add(new TextField("price", "10.0", Field.Store.NO));
doc.add(new TextField(idField, "2", Field.Store.NO));
doc.add(new TextField(toField, "1", Field.Store.NO));
w.addDocument(doc);
// 2
doc = new Document();
doc.add(new TextField("price", "20.0", Field.Store.NO));
doc.add(new TextField(idField, "3", Field.Store.NO));
doc.add(new TextField(toField, "1", Field.Store.NO));
w.addDocument(doc);
// 3
doc = new Document();
doc.add(new TextField("description", "more random text", Field.Store.NO));
doc.add(new TextField("name", "name2", Field.Store.NO));
doc.add(new TextField(idField, "4", Field.Store.NO));
w.addDocument(doc);
w.commit();
// 4
doc = new Document();
doc.add(new TextField("price", "10.0", Field.Store.NO));
doc.add(new TextField(idField, "5", Field.Store.NO));
doc.add(new TextField(toField, "4", Field.Store.NO));
w.addDocument(doc);
// 5
doc = new Document();
doc.add(new TextField("price", "20.0", Field.Store.NO));
doc.add(new TextField(idField, "6", Field.Store.NO));
doc.add(new TextField(toField, "4", Field.Store.NO));
w.addDocument(doc);
IndexSearcher indexSearcher = new IndexSearcher(w.getReader());
w.close();
// Search for product
Query joinQuery =
JoinUtil.createJoinQuery(
idField,
false,
toField,
new TermQuery(new Term("name", "name2")),
indexSearcher,
ScoreMode.None);
TopDocs result = indexSearcher.search(joinQuery, 10);
assertEquals(2, result.totalHits);
assertEquals(4, result.scoreDocs[0].doc);
assertEquals(5, result.scoreDocs[1].doc);
joinQuery =
JoinUtil.createJoinQuery(
idField,
false,
toField,
new TermQuery(new Term("name", "name1")),
indexSearcher,
ScoreMode.None);
result = indexSearcher.search(joinQuery, 10);
assertEquals(2, result.totalHits);
assertEquals(1, result.scoreDocs[0].doc);
assertEquals(2, result.scoreDocs[1].doc);
// Search for offer
joinQuery =
JoinUtil.createJoinQuery(
toField,
false,
idField,
new TermQuery(new Term("id", "5")),
indexSearcher,
ScoreMode.None);
result = indexSearcher.search(joinQuery, 10);
assertEquals(1, result.totalHits);
assertEquals(3, result.scoreDocs[0].doc);
indexSearcher.getIndexReader().close();
dir.close();
}
public void testSimpleWithScoring() throws Exception {
final String idField = "id";
final String toField = "movieId";
Directory dir = newDirectory();
RandomIndexWriter w =
new RandomIndexWriter(
random(),
dir,
newIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random()))
.setMergePolicy(newLogMergePolicy()));
// 0
Document doc = new Document();
doc.add(new TextField("description", "A random movie", Field.Store.NO));
doc.add(new TextField("name", "Movie 1", Field.Store.NO));
doc.add(new TextField(idField, "1", Field.Store.NO));
w.addDocument(doc);
// 1
doc = new Document();
doc.add(new TextField("subtitle", "The first subtitle of this movie", Field.Store.NO));
doc.add(new TextField(idField, "2", Field.Store.NO));
doc.add(new TextField(toField, "1", Field.Store.NO));
w.addDocument(doc);
// 2
doc = new Document();
doc.add(new TextField("subtitle", "random subtitle; random event movie", Field.Store.NO));
doc.add(new TextField(idField, "3", Field.Store.NO));
doc.add(new TextField(toField, "1", Field.Store.NO));
w.addDocument(doc);
// 3
doc = new Document();
doc.add(new TextField("description", "A second random movie", Field.Store.NO));
doc.add(new TextField("name", "Movie 2", Field.Store.NO));
doc.add(new TextField(idField, "4", Field.Store.NO));
w.addDocument(doc);
w.commit();
// 4
doc = new Document();
doc.add(
new TextField(
"subtitle", "a very random event happened during christmas night", Field.Store.NO));
doc.add(new TextField(idField, "5", Field.Store.NO));
doc.add(new TextField(toField, "4", Field.Store.NO));
w.addDocument(doc);
// 5
doc = new Document();
doc.add(new TextField("subtitle", "movie end movie test 123 test 123 random", Field.Store.NO));
doc.add(new TextField(idField, "6", Field.Store.NO));
doc.add(new TextField(toField, "4", Field.Store.NO));
w.addDocument(doc);
IndexSearcher indexSearcher = new IndexSearcher(w.getReader());
w.close();
// Search for movie via subtitle
Query joinQuery =
JoinUtil.createJoinQuery(
toField,
false,
idField,
new TermQuery(new Term("subtitle", "random")),
indexSearcher,
ScoreMode.Max);
TopDocs result = indexSearcher.search(joinQuery, 10);
assertEquals(2, result.totalHits);
assertEquals(0, result.scoreDocs[0].doc);
assertEquals(3, result.scoreDocs[1].doc);
// Score mode max.
joinQuery =
JoinUtil.createJoinQuery(
toField,
false,
idField,
new TermQuery(new Term("subtitle", "movie")),
indexSearcher,
ScoreMode.Max);
result = indexSearcher.search(joinQuery, 10);
assertEquals(2, result.totalHits);
assertEquals(3, result.scoreDocs[0].doc);
assertEquals(0, result.scoreDocs[1].doc);
// Score mode total
joinQuery =
JoinUtil.createJoinQuery(
toField,
false,
idField,
new TermQuery(new Term("subtitle", "movie")),
indexSearcher,
ScoreMode.Total);
result = indexSearcher.search(joinQuery, 10);
assertEquals(2, result.totalHits);
assertEquals(0, result.scoreDocs[0].doc);
assertEquals(3, result.scoreDocs[1].doc);
// Score mode avg
joinQuery =
JoinUtil.createJoinQuery(
toField,
false,
idField,
new TermQuery(new Term("subtitle", "movie")),
indexSearcher,
ScoreMode.Avg);
result = indexSearcher.search(joinQuery, 10);
assertEquals(2, result.totalHits);
assertEquals(3, result.scoreDocs[0].doc);
assertEquals(0, result.scoreDocs[1].doc);
indexSearcher.getIndexReader().close();
dir.close();
}