@Override
public BytesRef indexedValueForSearch(Object value) {
long longValue = NumericUtils.doubleToSortableLong(parseDoubleValue(value));
BytesRefBuilder bytesRef = new BytesRefBuilder();
NumericUtils.longToPrefixCoded(longValue, 0, bytesRef); // 0 because of exact match
return bytesRef.get();
}
@Override
public void fillBytesRef() {
assert valueSize == 64 || valueSize == 32;
if (valueSize == 64) {
NumericUtils.longToPrefixCoded(value, shift, bytes);
} else {
NumericUtils.intToPrefixCoded((int) value, shift, bytes);
}
}
/**
* Converts a list of long value to their bytes ref representation as performed by {@link
* org.apache.lucene.analysis.NumericTokenStream}
*/
private BytesRef[] toBytesRefs(long[] values) {
BytesRef[] bytesRefs = new BytesRef[values.length];
for (int i = 0; i < values.length; i++) {
BytesRefBuilder b = new BytesRefBuilder();
NumericUtils.longToPrefixCoded(values[i], 0, b);
bytesRefs[i] = b.toBytesRef();
}
return bytesRefs;
}
@Override
public BytesRef getBytesRef() {
assert valueSize == 64 || valueSize == 32;
if (valueSize == 64) {
NumericUtils.longToPrefixCoded(value, shift, bytes);
} else {
NumericUtils.intToPrefixCoded((int) value, shift, bytes);
}
return bytes.get();
}
@Override
public String readableToIndexed(String val) {
switch (type) {
case INTEGER:
return NumericUtils.intToPrefixCoded(Integer.parseInt(val));
case FLOAT:
return NumericUtils.intToPrefixCoded(
NumericUtils.floatToSortableInt(Float.parseFloat(val)));
case LONG:
return NumericUtils.longToPrefixCoded(Long.parseLong(val));
case DOUBLE:
return NumericUtils.longToPrefixCoded(
NumericUtils.doubleToSortableLong(Double.parseDouble(val)));
case DATE:
return NumericUtils.longToPrefixCoded(dateField.parseMath(null, val).getTime());
default:
throw new SolrException(
SolrException.ErrorCode.SERVER_ERROR, "Unknown type for trie field: " + type);
}
}
/**
* Test that the we are able to index a 100 documents, that the job management works during that
* time and at least one of the submissions exists in the index afterwards.
*/
@Test
public void testCommit() throws CorruptIndexException, IOException, InterruptedException {
// Send the first 100 subissions to the current transaction.
Iterator<Submission> itr = subRepo.findAllSubmissions();
long lastSubId = 0;
for (int i = 0; i < 100; i++) {
Submission sub = itr.next();
assertNotNull(sub);
indexer.updated(sub);
lastSubId = sub.getId();
}
// Assert no job is running.
assertTrue(lastSubId > 0);
assertFalse(indexer.isJobRunning());
assertEquals("None", indexer.getCurrentJobLabel());
assertEquals(-1L, indexer.getCurrentJobProgress());
assertEquals(-1L, indexer.getCurrentJobTotal());
assertTrue(indexer.isUpdated(lastSubId));
// Commit the 100 jobs.
indexer.commit(false);
// Check that a job is running
assertTrue(indexer.isJobRunning());
assertEquals("Update Index", indexer.getCurrentJobLabel());
assertTrue(indexer.getCurrentJobProgress() >= 0 && indexer.getCurrentJobProgress() <= 100);
assertEquals(100L, indexer.getCurrentJobTotal());
// Wait for the current job to complete
for (int i = 0; i < 1000; i++) {
Thread.sleep(10);
if (!indexer.isJobRunning()) break;
}
assertFalse(indexer.isJobRunning());
// Check that we can find the documents we indexed.
IndexReader reader = IndexReader.open(indexer.index);
IndexSearcher searcher = new IndexSearcher(reader);
Query query = new TermQuery(new Term("subId", NumericUtils.longToPrefixCoded(lastSubId)));
TopDocs topDocs = searcher.search(query, 1);
assertNotNull(topDocs);
assertNotNull(topDocs.scoreDocs);
assertTrue(topDocs.scoreDocs.length > 0);
Document doc = searcher.doc(topDocs.scoreDocs[0].doc);
assertNotNull(doc);
assertEquals(String.valueOf(lastSubId), doc.get("subId"));
searcher.close();
}
private String convertNumber(Number number) {
if (Integer.class.isInstance(number)) {
return NumericUtils.intToPrefixCoded(number.intValue());
} else if (Double.class.isInstance(number)) {
return NumericUtils.doubleToPrefixCoded(number.doubleValue());
} else if (Long.class.isInstance(number)) {
return NumericUtils.longToPrefixCoded(number.longValue());
} else if (Float.class.isInstance(number)) {
return NumericUtils.floatToPrefixCoded(number.floatValue());
} else if (Byte.class.isInstance(number)) {
return NumericUtils.intToPrefixCoded(number.intValue());
} else if (Short.class.isInstance(number)) {
return NumericUtils.intToPrefixCoded(number.intValue());
} else if (BigDecimal.class.isInstance(number)) {
return NumericUtils.doubleToPrefixCoded(number.doubleValue());
} else if (BigInteger.class.isInstance(number)) {
return NumericUtils.longToPrefixCoded(number.longValue());
} else {
throw new IllegalArgumentException("Unsupported numeric type " + number.getClass().getName());
}
}
/** Test that the index can be rebuilt. */
@Test
public void testRebuild() throws CorruptIndexException, IOException, InterruptedException {
// Assert no job is running
assertFalse(indexer.isJobRunning());
assertEquals("None", indexer.getCurrentJobLabel());
assertEquals(-1L, indexer.getCurrentJobProgress());
assertEquals(-1L, indexer.getCurrentJobTotal());
// Kickoff rebuilding the index
indexer.rebuild(false);
// Check that a job is running
assertTrue(indexer.isJobRunning());
assertEquals("Rebuild Index", indexer.getCurrentJobLabel());
assertTrue(indexer.getCurrentJobProgress() >= 0);
assertTrue(indexer.getCurrentJobTotal() >= 100);
// Wait for the current job to complete
for (int i = 0; i < 1000; i++) {
Thread.sleep(10);
if (!indexer.isJobRunning()) break;
}
assertFalse(indexer.isJobRunning());
// Check that the first submission is present in the index.
Submission sub = subRepo.findAllSubmissions().next();
IndexReader reader = IndexReader.open(indexer.index);
IndexSearcher searcher = new IndexSearcher(reader);
Query query = new TermQuery(new Term("subId", NumericUtils.longToPrefixCoded(sub.getId())));
TopDocs topDocs = searcher.search(query, 1);
assertNotNull(topDocs);
assertNotNull(topDocs.scoreDocs);
assertTrue(topDocs.scoreDocs.length > 0);
Document doc = searcher.doc(topDocs.scoreDocs[0].doc);
assertNotNull(doc);
assertEquals(String.valueOf(sub.getId()), doc.get("subId"));
searcher.close();
}
public static void displayPointRanges(Scanner in) {
double[][] point = getPoints(in, 1);
long hash, hashUpper;
double lon, lat, lonUpper, latUpper;
for (int i = 63; i >= 45; i -= GeoPointField.PRECISION_STEP) {
BytesRefBuilder brb = new BytesRefBuilder();
NumericUtils.longToPrefixCoded(
GeoUtils.mortonHash(point[0][LON_INDEX], point[0][LAT_INDEX]), i, brb);
BytesRef br = brb.get();
hash = NumericUtils.prefixCodedToLong(br);
hashUpper = hash | ((1L << i) - 1);
lon = GeoUtils.mortonUnhashLon(hash);
lat = GeoUtils.mortonUnhashLat(hash);
lonUpper = GeoUtils.mortonUnhashLon(hashUpper);
latUpper = GeoUtils.mortonUnhashLat(hashUpper);
System.out.println(
i + ": " + br + " " + hash + " (" + lon + "," + lat + ")" + " : " + "(" + lonUpper + ","
+ latUpper + ")");
}
}
@Override
public BytesRef indexedValueForSearch(Object value) {
BytesRefBuilder bytesRef = new BytesRefBuilder();
NumericUtils.longToPrefixCoded(parseValue(value), 0, bytesRef); // 0 because of exact match
return bytesRef.get();
}
@Override
public BytesRef indexedValueForSearch(Object value) {
BytesRef bytesRef = new BytesRef();
NumericUtils.longToPrefixCoded(parseValue(value), precisionStep(), bytesRef);
return bytesRef;
}
/**
* Tests for LuceneSerializer
*
* @author vema
*/
public class LuceneSerializerTest {
private LuceneSerializer serializer;
private PathBuilder<Object> entityPath;
private StringPath title;
private StringPath author;
private StringPath text;
private StringPath rating;
private StringPath publisher;
private NumberPath<Integer> year;
private NumberPath<Double> gross;
private NumberPath<Long> longField;
private NumberPath<Short> shortField;
private NumberPath<Byte> byteField;
private NumberPath<Float> floatField;
private static final String YEAR_PREFIX_CODED = NumericUtils.intToPrefixCoded(1990);
private static final String GROSS_PREFIX_CODED = NumericUtils.doubleToPrefixCoded(900.00);
private static final String LONG_PREFIX_CODED = NumericUtils.longToPrefixCoded(1);
private static final String SHORT_PREFIX_CODED = NumericUtils.intToPrefixCoded(1);
private static final String BYTE_PREFIX_CODED = NumericUtils.intToPrefixCoded(1);
private static final String FLOAT_PREFIX_CODED = NumericUtils.floatToPrefixCoded((float) 1.0);
private IndexWriterConfig config;
private RAMDirectory idx;
private IndexWriter writer;
private IndexSearcher searcher;
private final QueryMetadata metadata = new DefaultQueryMetadata();
private Document createDocument() {
Document doc = new Document();
doc.add(new Field("title", new StringReader("Jurassic Park")));
doc.add(new Field("author", new StringReader("Michael Crichton")));
doc.add(new Field("text", new StringReader("It's a UNIX system! I know this!")));
doc.add(new Field("rating", new StringReader("Good")));
doc.add(new Field("publisher", "", Store.YES, Index.ANALYZED));
doc.add(new NumericField("year", Store.YES, true).setIntValue(1990));
doc.add(new NumericField("gross", Store.YES, true).setDoubleValue(900.00));
doc.add(new NumericField("longField", Store.YES, true).setLongValue(1));
doc.add(new NumericField("shortField", Store.YES, true).setIntValue(1));
doc.add(new NumericField("byteField", Store.YES, true).setIntValue(1));
doc.add(new NumericField("floatField", Store.YES, true).setFloatValue(1));
return doc;
}
@Before
public void setUp() throws Exception {
serializer = new LuceneSerializer(true, true);
entityPath = new PathBuilder<Object>(Object.class, "obj");
title = entityPath.getString("title");
author = entityPath.getString("author");
text = entityPath.getString("text");
publisher = entityPath.getString("publisher");
year = entityPath.getNumber("year", Integer.class);
rating = entityPath.getString("rating");
gross = entityPath.getNumber("gross", Double.class);
longField = entityPath.getNumber("longField", Long.class);
shortField = entityPath.getNumber("shortField", Short.class);
byteField = entityPath.getNumber("byteField", Byte.class);
floatField = entityPath.getNumber("floatField", Float.class);
idx = new RAMDirectory();
config =
new IndexWriterConfig(Version.LUCENE_31, new StandardAnalyzer(Version.LUCENE_30))
.setOpenMode(IndexWriterConfig.OpenMode.CREATE);
writer = new IndexWriter(idx, config);
writer.addDocument(createDocument());
writer.close();
IndexReader reader = IndexReader.open(idx);
searcher = new IndexSearcher(reader);
}
@After
public void tearDown() throws Exception {
searcher.close();
}
private void testQuery(Expression<?> expr, int expectedHits) throws Exception {
Query query = serializer.toQuery(expr, metadata);
TopDocs docs = searcher.search(query, 100);
assertEquals(expectedHits, docs.totalHits);
}
private void testQuery(Expression<?> expr, String expectedQuery, int expectedHits)
throws Exception {
Query query = serializer.toQuery(expr, metadata);
TopDocs docs = searcher.search(query, 100);
assertEquals(expectedHits, docs.totalHits);
assertEquals(expectedQuery, query.toString());
}
@Test
public void QueryElement() throws Exception {
Query query1 = serializer.toQuery(author.like("Michael"), metadata);
Query query2 = serializer.toQuery(text.like("Text"), metadata);
BooleanExpression query =
BooleanExpression.anyOf(new QueryElement(query1), new QueryElement(query2));
testQuery(query, "author:michael text:text", 1);
}
@Test
public void Like() throws Exception {
testQuery(author.like("*ichael*"), "author:*ichael*", 1);
}
@Test
public void Like_Custom_Wildcard_Single_Character() throws Exception {
testQuery(author.like("Mi?hael"), "author:mi?hael", 1);
}
@Test
public void Like_Custom_Wildcard_Multiple_Character() throws Exception {
testQuery(text.like("*U*X*"), "text:*u*x*", 1);
}
@Test
public void Like_Phrase() throws Exception {
testQuery(title.like("*rassic Par*"), "+title:**rassic* +title:*par**", 1);
}
@Test
public void Like_or_like() throws Exception {
testQuery(title.like("House").or(author.like("*ichae*")), "title:house author:*ichae*", 1);
}
@Test
public void Like_and_like() throws Exception {
testQuery(title.like("*assic*").and(rating.like("G?od")), "+title:*assic* +rating:g?od", 1);
}
@Test
public void Eq() throws Exception {
testQuery(rating.eq("good"), "rating:good", 1);
}
@Test
public void Eq_with_deep_path() throws Exception {
StringPath deepPath = entityPath.get("property1", Object.class).getString("property2");
testQuery(deepPath.eq("good"), "property1.property2:good", 0);
}
@Test
public void FuzzyLike() throws Exception {
testQuery(LuceneExpressions.fuzzyLike(rating, "Good"), "rating:Good~0.5", 1);
}
@Test
public void FuzzyLike_with_Similarity() throws Exception {
testQuery(LuceneExpressions.fuzzyLike(rating, "Good", 0.6f), "rating:Good~0.6", 1);
}
@Test
public void FuzzyLike_with_Similarity_and_prefix() throws Exception {
testQuery(LuceneExpressions.fuzzyLike(rating, "Good", 0.6f, 0), "rating:Good~0.6", 1);
}
@Test
public void Eq_Numeric_Integer() throws Exception {
testQuery(year.eq(1990), "year:" + YEAR_PREFIX_CODED, 1);
}
@Test
public void Eq_Numeric_Double() throws Exception {
testQuery(gross.eq(900.00), "gross:" + GROSS_PREFIX_CODED, 1);
}
@Test
public void Eq_Numeric() throws Exception {
testQuery(longField.eq(1l), "longField:" + LONG_PREFIX_CODED, 1);
testQuery(shortField.eq((short) 1), "shortField:" + SHORT_PREFIX_CODED, 1);
testQuery(byteField.eq((byte) 1), "byteField:" + BYTE_PREFIX_CODED, 1);
testQuery(floatField.eq((float) 1.0), "floatField:" + FLOAT_PREFIX_CODED, 1);
}
@Test
public void Equals_Ignores_Case() throws Exception {
testQuery(title.eq("Jurassic"), "title:jurassic", 1);
}
@Test(expected = UnsupportedOperationException.class)
public void Title_Equals_Ignore_Case_Or_Year_Equals() throws Exception {
testQuery(
title.equalsIgnoreCase("House").or(year.eq(1990)),
"title:house year:" + YEAR_PREFIX_CODED,
1);
}
@Test
public void Eq_and_eq() throws Exception {
testQuery(
title.eq("Jurassic Park").and(year.eq(1990)),
"+title:\"jurassic park\" +year:" + YEAR_PREFIX_CODED,
1);
}
@Test
public void Eq_and_Eq_and_eq() throws Exception {
testQuery(
title.eq("Jurassic Park").and(year.eq(1990)).and(author.eq("Michael Crichton")),
"+(+title:\"jurassic park\" +year:" + YEAR_PREFIX_CODED + ") +author:\"michael crichton\"",
1);
}
@Test(expected = UnsupportedOperationException.class)
public void Equals_Ignore_Case_And_Or() throws Exception {
testQuery(
title
.equalsIgnoreCase("Jurassic Park")
.and(rating.equalsIgnoreCase("Bad"))
.or(author.equalsIgnoreCase("Michael Crichton")),
"(+title:\"jurassic park\" +rating:bad) author:\"michael crichton\"",
1);
}
@Test
public void Eq_or_Eq_and_Eq_Does_Not_Find_Results() throws Exception {
testQuery(
title.eq("jeeves").or(rating.eq("superb")).and(author.eq("michael crichton")),
"+(title:jeeves rating:superb) +author:\"michael crichton\"",
0);
}
@Test
public void Eq_Phrase() throws Exception {
testQuery(title.eq("Jurassic Park"), "title:\"jurassic park\"", 1);
}
@Test
@Ignore("Not easily done in Lucene!")
public void Publisher_Equals_Empty_String() throws Exception {
testQuery(publisher.eq(""), "publisher:", 1);
}
@Test
public void Eq_Phrase_Should_Not_Find_Results_But_LuceNe_Semantics_Differs_From_Querydsls()
throws Exception {
testQuery(text.eq("UNIX System"), "text:\"unix system\"", 1);
}
@Test
public void Eq_Phrase_Does_Not_Find_Results_Because_Word_In_Middle() throws Exception {
testQuery(title.eq("Jurassic Amusement Park"), "title:\"jurassic amusement park\"", 0);
}
@Test
public void Like_not_Does_Not_Find_Results() throws Exception {
testQuery(title.like("*H*e*").not(), "-title:*h*e* +*:*", 1);
}
@Test(expected = UnsupportedOperationException.class)
public void Title_Equals_Ignore_Case_Negation_Or_Rating_Equals_Ignore_Case() throws Exception {
testQuery(
title.equalsIgnoreCase("House").not().or(rating.equalsIgnoreCase("Good")),
"-title:house rating:good",
1);
}
@Test
public void Eq_not_Does_Not_Find_Results() throws Exception {
testQuery(title.eq("Jurassic Park").not(), "-title:\"jurassic park\" +*:*", 0);
}
@Test
public void Title_Equals_Not_House() throws Exception {
testQuery(title.eq("house").not(), "-title:house +*:*", 1);
}
@Test
public void Eq_and_Eq_not_Does_Not_Find_Results_Because_Second_Expression_Finds_Nothing()
throws Exception {
testQuery(
rating.eq("superb").and(title.eq("house").not()), "+rating:superb +(-title:house +*:*)", 0);
}
@Test
public void Not_Equals_Finds_One() throws Exception {
testQuery(title.ne("house"), "-title:house +*:*", 1);
}
@Test
public void Not_Equals_Finds_None() throws Exception {
testQuery(title.ne("Jurassic Park"), "-title:\"jurassic park\" +*:*", 0);
}
@Test
public void Nothing_Found_With_Not_Equals_Or_Equals() throws Exception {
testQuery(
title.ne("jurassic park").or(rating.eq("lousy")),
"(-title:\"jurassic park\" +*:*) rating:lousy",
0);
}
@Test
public void Ne_and_eq() throws Exception {
testQuery(title.ne("house").and(rating.eq("good")), "+(-title:house +*:*) +rating:good", 1);
}
@Test
public void StartsWith() throws Exception {
testQuery(title.startsWith("Jurassi"), "title:jurassi*", 1);
}
@Test
public void StartsWith_Phrase() throws Exception {
testQuery(title.startsWith("jurassic par"), "+title:jurassic* +title:*par*", 1);
}
@Test(expected = UnsupportedOperationException.class)
public void Starts_With_Ignore_Case_Phrase_Does_Not_Find_Results() throws Exception {
testQuery(title.startsWithIgnoreCase("urassic Par"), "+title:urassic* +title:*par*", 0);
}
@Test
public void EndsWith() throws Exception {
testQuery(title.endsWith("ark"), "title:*ark", 1);
}
@Test(expected = UnsupportedOperationException.class)
public void Ends_With_Ignore_Case_Phrase() throws Exception {
testQuery(title.endsWithIgnoreCase("sic Park"), "+title:*sic* +title:*park", 1);
}
@Test(expected = UnsupportedOperationException.class)
public void Ends_With_Ignore_Case_Phrase_Does_Not_Find_Results() throws Exception {
testQuery(title.endsWithIgnoreCase("sic Par"), "+title:*sic* +title:*par", 0);
}
@Test
public void Contains() throws Exception {
testQuery(title.contains("rassi"), "title:*rassi*", 1);
}
@Test(expected = UnsupportedOperationException.class)
public void Contains_Ignore_Case_Phrase() throws Exception {
testQuery(title.containsIgnoreCase("rassi Pa"), "+title:*rassi* +title:*pa*", 1);
}
@Test
public void Contains_User_Inputted_Wildcards_Dont_Work() throws Exception {
testQuery(title.contains("r*i"), "title:*r\\*i*", 0);
}
@Test
public void Between() throws Exception {
testQuery(title.between("Indiana", "Kundun"), "title:[indiana TO kundun]", 1);
}
@Test
public void Between_Numeric_Integer() throws Exception {
testQuery(year.between(1980, 2000), "year:[1980 TO 2000]", 1);
}
@Test
public void Between_Numeric_Double() throws Exception {
testQuery(gross.between(10.00, 19030.00), "gross:[10.0 TO 19030.0]", 1);
}
@Test
public void Between_Numeric() throws Exception {
testQuery(longField.between(0l, 2l), "longField:[0 TO 2]", 1);
testQuery(shortField.between((short) 0, (short) 2), "shortField:[0 TO 2]", 1);
testQuery(byteField.between((byte) 0, (byte) 2), "byteField:[0 TO 2]", 1);
testQuery(floatField.between((float) 0.0, (float) 2.0), "floatField:[0.0 TO 2.0]", 1);
}
@Test
public void Between_Is_Inclusive_From_Start() throws Exception {
testQuery(title.between("Jurassic", "Kundun"), "title:[jurassic TO kundun]", 1);
}
@Test
public void Between_Is_Inclusive_To_End() throws Exception {
testQuery(title.between("Indiana", "Jurassic"), "title:[indiana TO jurassic]", 1);
}
@Test
public void Between_Does_Not_Find_Results() throws Exception {
testQuery(title.between("Indiana", "Jurassib"), "title:[indiana TO jurassib]", 0);
}
@Test
public void In() throws Exception {
testQuery(title.in(Arrays.asList("jurassic", "park")), "title:jurassic title:park", 1);
testQuery(title.in("jurassic", "park"), "title:jurassic title:park", 1);
testQuery(title.eq("jurassic").or(title.eq("park")), "title:jurassic title:park", 1);
}
@Test
public void Lt() throws Exception {
testQuery(rating.lt("Superb"), "rating:{* TO superb}", 1);
}
@Test
public void Lt_Numeric_Integer() throws Exception {
testQuery(year.lt(1991), "year:{* TO 1991}", 1);
}
@Test
public void Lt_Numeric_Double() throws Exception {
testQuery(gross.lt(10000.0), "gross:{* TO 10000.0}", 1);
}
@Test
public void Lt_Not_In_Range_Because_Equal() throws Exception {
testQuery(rating.lt("Good"), "rating:{* TO good}", 0);
}
@Test
public void Lt_Numeric_Integer_Not_In_Range_Because_Equal() throws Exception {
testQuery(year.lt(1990), "year:{* TO 1990}", 0);
}
@Test
public void Lt_Numeric_Double_Not_In_Range_Because_Equal() throws Exception {
testQuery(gross.lt(900.0), "gross:{* TO 900.0}", 0);
}
@Test
public void Loe() throws Exception {
testQuery(rating.loe("Superb"), "rating:[* TO superb]", 1);
}
@Test
public void Loe_Numeric_Integer() throws Exception {
testQuery(year.loe(1991), "year:[* TO 1991]", 1);
}
@Test
public void Loe_Numeric_Double() throws Exception {
testQuery(gross.loe(903.0), "gross:[* TO 903.0]", 1);
}
@Test
public void Loe_Equal() throws Exception {
testQuery(rating.loe("Good"), "rating:[* TO good]", 1);
}
@Test
public void Loe_Numeric_Integer_Equal() throws Exception {
testQuery(year.loe(1990), "year:[* TO 1990]", 1);
}
@Test
public void Loe_Numeric_Double_Equal() throws Exception {
testQuery(gross.loe(900.0), "gross:[* TO 900.0]", 1);
}
@Test
public void Loe_Not_Found() throws Exception {
testQuery(rating.loe("Bad"), "rating:[* TO bad]", 0);
}
@Test
public void Loe_Numeric_Integer_Not_Found() throws Exception {
testQuery(year.loe(1989), "year:[* TO 1989]", 0);
}
@Test
public void Loe_Numeric_Double_Not_Found() throws Exception {
testQuery(gross.loe(899.9), "gross:[* TO 899.9]", 0);
}
@Test
public void Gt() throws Exception {
testQuery(rating.gt("Bad"), "rating:{bad TO *}", 1);
}
@Test
public void Gt_Numeric_Integer() throws Exception {
testQuery(year.gt(1989), "year:{1989 TO *}", 1);
}
@Test
public void Gt_Numeric_Double() throws Exception {
testQuery(gross.gt(100.00), "gross:{100.0 TO *}", 1);
}
@Test
public void Gt_Not_In_Range_Because_Equal() throws Exception {
testQuery(rating.gt("Good"), "rating:{good TO *}", 0);
}
@Test
public void Gt_Numeric_Integer_Not_In_Range_Because_Equal() throws Exception {
testQuery(year.gt(1990), "year:{1990 TO *}", 0);
}
@Test
public void Gt_Numeric_Double_Not_In_Range_Because_Equal() throws Exception {
testQuery(gross.gt(900.00), "gross:{900.0 TO *}", 0);
}
@Test
public void Goe() throws Exception {
testQuery(rating.goe("Bad"), "rating:[bad TO *]", 1);
}
@Test
public void Goe_Numeric_Integer() throws Exception {
testQuery(year.goe(1989), "year:[1989 TO *]", 1);
}
@Test
public void Goe_Numeric_Double() throws Exception {
testQuery(gross.goe(320.50), "gross:[320.5 TO *]", 1);
}
@Test
public void Goe_Equal() throws Exception {
testQuery(rating.goe("Good"), "rating:[good TO *]", 1);
}
@Test
public void Goe_Numeric_Integer_Equal() throws Exception {
testQuery(year.goe(1990), "year:[1990 TO *]", 1);
}
@Test
public void Goe_Numeric_Double_Equal() throws Exception {
testQuery(gross.goe(900.00), "gross:[900.0 TO *]", 1);
}
@Test
public void Goe_Not_Found() throws Exception {
testQuery(rating.goe("Hood"), "rating:[hood TO *]", 0);
}
@Test
public void Goe_Numeric_Integer_Not_Found() throws Exception {
testQuery(year.goe(1991), "year:[1991 TO *]", 0);
}
@Test
public void Goe_Numeric_Double_Not_Found() throws Exception {
testQuery(gross.goe(900.10), "gross:[900.1 TO *]", 0);
}
@Test
public void Equals_Empty_String() throws Exception {
testQuery(title.eq(""), "title:", 0);
}
@Test
public void Not_Equals_Empty_String() throws Exception {
testQuery(title.ne(""), "-title: +*:*", 1);
}
@Test
public void Contains_Empty_String() throws Exception {
testQuery(title.contains(""), "title:**", 1);
}
@Test
public void Like_Empty_String() throws Exception {
testQuery(title.like(""), "title:", 0);
}
@Test
public void Starts_With_Empty_String() throws Exception {
testQuery(title.startsWith(""), "title:*", 1);
}
@Test
public void Ends_With_Empty_String() throws Exception {
testQuery(title.endsWith(""), "title:*", 1);
}
@Test
public void Between_Empty_Strings() throws Exception {
testQuery(title.between("", ""), "title:[ TO ]", 0);
}
@Test
public void BooleanBuilder() throws Exception {
testQuery(new BooleanBuilder(gross.goe(900.10)), "gross:[900.1 TO *]", 0);
}
@Test
@Ignore
public void Fuzzy() throws Exception {
fail("Not yet implemented!");
}
@Test
@Ignore
public void Proximity() throws Exception {
fail("Not yet implemented!");
}
@Test
@Ignore
public void Boost() throws Exception {
fail("Not yet implemented!");
}
private boolean unsupportedOperation(Predicate filter) {
if (filter instanceof Operation<?>) {
Operator<?> op = ((Operation<?>) filter).getOperator();
if (op == Ops.STARTS_WITH_IC
|| op == Ops.EQ_IGNORE_CASE
|| op == Ops.STARTS_WITH_IC
|| op == Ops.ENDS_WITH_IC
|| op == Ops.STRING_CONTAINS_IC) {
return true;
}
}
return false;
}
@Test
public void various() throws Exception {
MatchingFiltersFactory filters = new MatchingFiltersFactory(Module.LUCENE, Target.LUCENE);
for (Predicate filter : filters.string(title, StringConstant.create("jurassic park"))) {
if (unsupportedOperation(filter)) {
continue;
}
testQuery(filter, 1);
}
for (Predicate filter : filters.string(author, StringConstant.create("michael crichton"))) {
if (unsupportedOperation(filter)) {
continue;
}
testQuery(filter, 1);
}
for (Predicate filter : filters.string(title, StringConstant.create("1990"))) {
if (unsupportedOperation(filter)) {
continue;
}
testQuery(filter, 0);
}
}
}
private void testRandomTrieAndClassicRangeQuery(int precisionStep) throws Exception {
final Random rnd = newRandom();
String field = "field" + precisionStep;
int termCountT = 0, termCountC = 0;
for (int i = 0; i < 50; i++) {
long lower = (long) (rnd.nextDouble() * noDocs * distance) + startOffset;
long upper = (long) (rnd.nextDouble() * noDocs * distance) + startOffset;
if (lower > upper) {
long a = lower;
lower = upper;
upper = a;
}
// test inclusive range
NumericRangeQuery<Long> tq =
NumericRangeQuery.newLongRange(field, precisionStep, lower, upper, true, true);
TermRangeQuery cq =
new TermRangeQuery(
field,
NumericUtils.longToPrefixCoded(lower),
NumericUtils.longToPrefixCoded(upper),
true,
true);
TopDocs tTopDocs = searcher.search(tq, 1);
TopDocs cTopDocs = searcher.search(cq, 1);
assertEquals(
"Returned count for NumericRangeQuery and TermRangeQuery must be equal",
cTopDocs.totalHits,
tTopDocs.totalHits);
termCountT += tq.getTotalNumberOfTerms();
termCountC += cq.getTotalNumberOfTerms();
// test exclusive range
tq = NumericRangeQuery.newLongRange(field, precisionStep, lower, upper, false, false);
cq =
new TermRangeQuery(
field,
NumericUtils.longToPrefixCoded(lower),
NumericUtils.longToPrefixCoded(upper),
false,
false);
tTopDocs = searcher.search(tq, 1);
cTopDocs = searcher.search(cq, 1);
assertEquals(
"Returned count for NumericRangeQuery and TermRangeQuery must be equal",
cTopDocs.totalHits,
tTopDocs.totalHits);
termCountT += tq.getTotalNumberOfTerms();
termCountC += cq.getTotalNumberOfTerms();
// test left exclusive range
tq = NumericRangeQuery.newLongRange(field, precisionStep, lower, upper, false, true);
cq =
new TermRangeQuery(
field,
NumericUtils.longToPrefixCoded(lower),
NumericUtils.longToPrefixCoded(upper),
false,
true);
tTopDocs = searcher.search(tq, 1);
cTopDocs = searcher.search(cq, 1);
assertEquals(
"Returned count for NumericRangeQuery and TermRangeQuery must be equal",
cTopDocs.totalHits,
tTopDocs.totalHits);
termCountT += tq.getTotalNumberOfTerms();
termCountC += cq.getTotalNumberOfTerms();
// test right exclusive range
tq = NumericRangeQuery.newLongRange(field, precisionStep, lower, upper, true, false);
cq =
new TermRangeQuery(
field,
NumericUtils.longToPrefixCoded(lower),
NumericUtils.longToPrefixCoded(upper),
true,
false);
tTopDocs = searcher.search(tq, 1);
cTopDocs = searcher.search(cq, 1);
assertEquals(
"Returned count for NumericRangeQuery and TermRangeQuery must be equal",
cTopDocs.totalHits,
tTopDocs.totalHits);
termCountT += tq.getTotalNumberOfTerms();
termCountC += cq.getTotalNumberOfTerms();
}
if (precisionStep == Integer.MAX_VALUE) {
assertEquals(
"Total number of terms should be equal for unlimited precStep", termCountT, termCountC);
} else {
System.out.println("Average number of terms during random search on '" + field + "':");
System.out.println(" Trie query: " + (((double) termCountT) / (50 * 4)));
System.out.println(" Classical query: " + (((double) termCountC) / (50 * 4)));
}
}
/**
* This method produces the common part of the query handle the filter search clauses.
*
* @param andQuery The existing and-based query
* @param filter The filter search paramaters.
* @param submissions Whether this is for submissions or action logs
*/
public void buildQuery(BooleanQuery andQuery, SearchFilter filter, boolean submissions) {
QueryParser parser = new QueryParser(indexer.version, "searchText", indexer.standardAnalyzer);
// Include Submission filter
if (filter.getIncludedSubmissions().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (Submission sub : filter.getIncludedSubmissions()) {
orQuery.add(
new TermQuery(new Term("subId", NumericUtils.longToPrefixCoded(sub.getId()))),
Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Include Log filter
if (filter.getIncludedActionLogs().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (ActionLog log : filter.getIncludedActionLogs()) {
orQuery.add(
new TermQuery(new Term("logId", NumericUtils.longToPrefixCoded(log.getId()))),
Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Exclude Submission filter
for (Submission sub : filter.getExcludedSubmissions()) {
andQuery.add(
new TermQuery(new Term("subId", NumericUtils.longToPrefixCoded(sub.getId()))),
Occur.MUST_NOT);
}
// Exclude Log filter
for (ActionLog log : filter.getExcludedActionLogs()) {
andQuery.add(
new TermQuery(new Term("logId", NumericUtils.longToPrefixCoded(log.getId()))),
Occur.MUST_NOT);
}
// Search Text Filter
if (filter.getSearchText().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (String searchText : filter.getSearchText()) {
try {
// First try to interpret it as a complex lucene search string.
orQuery.add(parser.parse(searchText), Occur.SHOULD);
} catch (ParseException e) {
// If that fails just fall back to a term query.
orQuery.add(new TermQuery(new Term("searchText", searchText)), Occur.SHOULD);
}
}
andQuery.add(orQuery, Occur.MUST);
}
// State Filter
if (filter.getStates().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (String stateName : filter.getStates()) {
State state = stateManager.getState(stateName);
orQuery.add(new TermQuery(new Term("state", state.getDisplayName())), Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Assignee Filter
if (filter.getAssignees().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (Person assignee : filter.getAssignees()) {
long assigneeId = 0;
if (assignee != null) assigneeId = assignee.getId();
orQuery.add(
new TermQuery(new Term("searchAssigned", NumericUtils.longToPrefixCoded(assigneeId))),
Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Embargo Filter
if (filter.getEmbargoTypes().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (EmbargoType embargo : filter.getEmbargoTypes()) {
orQuery.add(
new TermQuery(
new Term("embargo", embargo.getName() + " " + embargo.getGuarantor().name())),
Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Graduation Semester Filter
if (filter.getGraduationSemesters().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (Semester semester : filter.getGraduationSemesters()) {
// We can't index it if it dosn't have a date.
if (semester.year == null) continue;
Calendar cal = Calendar.getInstance();
cal.clear();
cal.set(Calendar.YEAR, semester.year);
if (semester.month != null) {
cal.set(Calendar.MONTH, semester.month);
}
orQuery.add(
new TermQuery(
new Term(
"graduationSemester", NumericUtils.longToPrefixCoded(cal.getTimeInMillis()))),
Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Degree Filter
if (filter.getDegrees().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (String degree : filter.getDegrees()) {
orQuery.add(new TermQuery(new Term("degree", degree)), Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Department Filter
if (filter.getDepartments().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (String dept : filter.getDepartments()) {
orQuery.add(new TermQuery(new Term("department", dept)), Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Program Filter
if (filter.getPrograms().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (String program : filter.getPrograms()) {
orQuery.add(new TermQuery(new Term("program", program)), Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// College Filter
if (filter.getColleges().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (String college : filter.getColleges()) {
orQuery.add(new TermQuery(new Term("college", college)), Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Major Filter
if (filter.getMajors().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (String major : filter.getMajors()) {
orQuery.add(new TermQuery(new Term("major", major)), Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// Document Type Filter
if (filter.getDocumentTypes().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (String docType : filter.getDocumentTypes()) {
orQuery.add(new TermQuery(new Term("documentType", docType)), Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
// UMI Release Filter
if (filter.getUMIRelease() != null) {
if (filter.getUMIRelease()) {
andQuery.add(new TermQuery(new Term("umiRelease", "yes")), Occur.MUST);
} else {
andQuery.add(new TermQuery(new Term("umiRelease", "no")), Occur.MUST);
}
}
// Date Range Filter
if (filter.getDateRangeStart() != null || filter.getDateRangeEnd() != null) {
long startTime = 0;
if (filter.getDateRangeStart() != null) startTime = filter.getDateRangeStart().getTime();
long endTime = Long.MAX_VALUE;
if (filter.getDateRangeEnd() != null) endTime = filter.getDateRangeEnd().getTime();
if (submissions)
andQuery.add(
NumericRangeQuery.newLongRange("submissionDate", startTime, endTime, true, true),
Occur.MUST);
else
andQuery.add(
NumericRangeQuery.newLongRange("lastEventTime", startTime, endTime, true, true),
Occur.MUST);
}
// Custom Action Filter
if (filter.getCustomActions().size() > 0) {
BooleanQuery orQuery = new BooleanQuery();
for (CustomActionDefinition customAction : filter.getCustomActions()) {
orQuery.add(new TermQuery(new Term("customAction", customAction.getLabel())), Occur.SHOULD);
}
andQuery.add(orQuery, Occur.MUST);
}
}
@Override
public String storedToIndexed(Fieldable f) {
if (f instanceof NumericField) {
final Number val = ((NumericField) f).getNumericValue();
if (val == null)
throw new SolrException(
SolrException.ErrorCode.SERVER_ERROR, "Invalid field contents: " + f.name());
switch (type) {
case INTEGER:
return NumericUtils.intToPrefixCoded(val.intValue());
case FLOAT:
return NumericUtils.intToPrefixCoded(NumericUtils.floatToSortableInt(val.floatValue()));
case LONG: // fallthrough!
case DATE:
return NumericUtils.longToPrefixCoded(val.longValue());
case DOUBLE:
return NumericUtils.longToPrefixCoded(
NumericUtils.doubleToSortableLong(val.doubleValue()));
default:
throw new SolrException(
SolrException.ErrorCode.SERVER_ERROR, "Unknown type for trie field: " + f.name());
}
} else {
// the following code is "deprecated" and only to support pre-3.2 indexes using the old
// BinaryField encoding:
final byte[] arr = f.getBinaryValue();
if (arr == null)
throw new SolrException(
SolrException.ErrorCode.SERVER_ERROR, "Invalid field contents: " + f.name());
switch (type) {
case INTEGER:
return NumericUtils.intToPrefixCoded(toInt(arr));
case FLOAT:
{
// WARNING: Code Duplication! Keep in sync with o.a.l.util.NumericUtils!
// copied from NumericUtils to not convert to/from float two times
// code in next 2 lines is identical to: int v =
// NumericUtils.floatToSortableInt(Float.intBitsToFloat(toInt(arr)));
int v = toInt(arr);
if (v < 0) v ^= 0x7fffffff;
return NumericUtils.intToPrefixCoded(v);
}
case LONG: // fallthrough!
case DATE:
return NumericUtils.longToPrefixCoded(toLong(arr));
case DOUBLE:
{
// WARNING: Code Duplication! Keep in sync with o.a.l.util.NumericUtils!
// copied from NumericUtils to not convert to/from double two times
// code in next 2 lines is identical to: long v =
// NumericUtils.doubleToSortableLong(Double.longBitsToDouble(toLong(arr)));
long v = toLong(arr);
if (v < 0) v ^= 0x7fffffffffffffffL;
return NumericUtils.longToPrefixCoded(v);
}
default:
throw new SolrException(
SolrException.ErrorCode.SERVER_ERROR, "Unknown type for trie field: " + f.name());
}
}
}