internal FieldsReader(PerFieldDocValuesFormat outerInstance, FieldsReader other)
{
this.outerInstance = outerInstance;
IDictionary <DocValuesProducer, DocValuesProducer> oldToNew = new JCG.Dictionary <DocValuesProducer, DocValuesProducer>(IdentityEqualityComparer <DocValuesProducer> .Default);
// First clone all formats
foreach (KeyValuePair <string, DocValuesProducer> ent in other.formats)
{
DocValuesProducer values = ent.Value;
formats[ent.Key] = values;
oldToNew[ent.Value] = values;
}
// Then rebuild fields:
foreach (KeyValuePair <string, DocValuesProducer> ent in other.fields)
{
DocValuesProducer producer;
oldToNew.TryGetValue(ent.Value, out producer);
if (Debugging.AssertsEnabled)
{
Debugging.Assert(producer != null);
}
fields[ent.Key] = producer;
}
}
private IDictionary <string, Document> GenerateIndexDocuments(int ndocs)
{
IDictionary <string, Document> docs = new JCG.Dictionary <string, Document>();
for (int i = 0; i < ndocs; i++)
{
Field field = new TextField(FIELD_NAME, "field_" + i, Field.Store.YES);
Field payload = new StoredField(PAYLOAD_FIELD_NAME, new BytesRef("payload_" + i));
Field weight1 = new NumericDocValuesField(WEIGHT_FIELD_NAME_1, 10 + i);
Field weight2 = new NumericDocValuesField(WEIGHT_FIELD_NAME_2, 20 + i);
Field weight3 = new NumericDocValuesField(WEIGHT_FIELD_NAME_3, 30 + i);
Field contexts = new StoredField(CONTEXTS_FIELD_NAME, new BytesRef("ctx_" + i + "_0"));
Document doc = new Document();
doc.Add(field);
doc.Add(payload);
doc.Add(weight1);
doc.Add(weight2);
doc.Add(weight3);
doc.Add(contexts);
for (int j = 1; j < AtLeast(3); j++)
{
contexts.SetBytesValue(new BytesRef("ctx_" + i + "_" + j));
doc.Add(contexts);
}
docs.Put(field.GetStringValue(), doc);
}
return(docs);
}
public void TestUnicodeMapRandom()
{
// do random change to both, then compare
var random = new Random(12345); // reproducible results
Logln("Comparing against HashMap");
UnicodeMap <string> map1 = new UnicodeMap <string>();
IDictionary <Integer, string> map2 = new JCG.Dictionary <Integer, string>();
for (int counter = 0; counter < ITERATIONS; ++counter)
{
int start = random.Next(LIMIT);
string value = TEST_VALUES[random.Next(TEST_VALUES.Length)];
string logline = Utility.Hex(start) + "\t" + value;
if (SHOW_PROGRESS)
{
Logln(counter + "\t" + logline);
}
log.Add(logline);
if (DEBUG && counter == 144)
{
Console.Out.WriteLine(" debug");
}
map1.Put(start, value);
map2[new Integer(start)] = value;
check(map1, map2, counter);
}
checkNext(map1, map2, LIMIT);
}
/// <summary>
/// Checks some basic behaviour of an <see cref="Attribute"/>. </summary>
/// <param name="att"><see cref="Attribute"/> to reflect</param>
/// <param name="reflectedValues"> Contains a <see cref="IDictionary{String, Object}"/> with "AttributeSubclassType/key" as values.</param>
public static void AssertAttributeReflection(Attribute att, IDictionary <string, object> reflectedValues)
{
IDictionary <string, object> map = new JCG.Dictionary <string, object>();
att.ReflectWith(new AttributeReflectorAnonymousClass(map));
Assert.AreEqual(reflectedValues, map, aggressive: false, "Reflection does not produce same map");
}
public virtual void DoRandom(int iter, bool ignoreCase)
{
CharArrayMap <int?> map = new CharArrayMap <int?>(TEST_VERSION_CURRENT, 1, ignoreCase);
IDictionary <string, int?> hmap = new JCG.Dictionary <string, int?>();
char[] key;
for (int i = 0; i < iter; i++)
{
int len = Random.Next(5);
key = new char[len];
for (int j = 0; j < key.Length; j++)
{
key[j] = (char)Random.Next(127);
}
string keyStr = new string(key);
string hmapKey = ignoreCase ? keyStr.ToLowerInvariant() : keyStr;
int val = Random.Next();
object o1 = map.Put(key, val);
object o2 = hmap.Put(hmapKey, val);
assertEquals(o1, o2);
// add it again with the string method
assertEquals(val, map.Put(keyStr, val));
assertEquals(val, map.Get(key, 0, key.Length));
assertEquals(val, map.Get(key));
assertEquals(val, map.Get(keyStr));
assertEquals(hmap.Count, map.size());
}
}
public virtual void TestReuseDocsEnumNoReuse()
{
Directory dir = NewDirectory();
Codec cp = TestUtil.AlwaysPostingsFormat(new Lucene40RWPostingsFormat());
RandomIndexWriter writer = new RandomIndexWriter(Random, dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random)).SetCodec(cp));
int numdocs = AtLeast(20);
CreateRandomIndex(numdocs, writer, Random);
writer.Commit();
DirectoryReader open = DirectoryReader.Open(dir);
foreach (AtomicReaderContext ctx in open.Leaves)
{
AtomicReader indexReader = (AtomicReader)ctx.Reader;
Terms terms = indexReader.GetTerms("body");
TermsEnum iterator = terms.GetEnumerator();
IDictionary <DocsEnum, bool?> enums = new JCG.Dictionary <DocsEnum, bool?>(IdentityEqualityComparer <DocsEnum> .Default);
MatchNoBits bits = new MatchNoBits(indexReader.MaxDoc);
while (iterator.MoveNext())
{
DocsEnum docs = iterator.Docs(Random.NextBoolean() ? bits : new MatchNoBits(indexReader.MaxDoc), null, Random.NextBoolean() ? DocsFlags.FREQS : DocsFlags.NONE);
enums[docs] = true;
}
Assert.AreEqual(terms.Count, enums.Count);
}
IOUtils.Dispose(writer, open, dir);
}
public void TestBogusArguments()
{
ArgumentException expected = NUnit.Framework.Assert.Throws <ArgumentException>(() =>
{
JCG.Dictionary <String, String> @params = new JCG.Dictionary <String, String>();
@params["bogusArg"] = "bogusValue";
new MorfologikFilterFactory(@params);
});
assertTrue(expected.Message.Contains("Unknown parameters"));
}
public override void SetReusableComponents(Analyzer analyzer, string fieldName, TokenStreamComponents components)
{
var componentsPerField = (IDictionary <string, TokenStreamComponents>)GetStoredValue(analyzer);
if (componentsPerField == null)
{
// LUCENENET-615: This needs to support nullable keys
componentsPerField = new JCG.Dictionary <string, TokenStreamComponents>();
SetStoredValue(analyzer, componentsPerField);
}
componentsPerField[fieldName] = components;
}
public virtual void TestEmpty()
{
IDictionary <string, MethodInfo> functions = new JCG.Dictionary <string, MethodInfo>();
try
{
JavascriptCompiler.Compile("sqrt(20)", functions);
Assert.Fail();
}
catch (ArgumentException e)
{
Assert.IsTrue(e.Message.Contains("Unrecognized method"));
}
}
public void TestMissingDictionary()
{
IResourceLoader loader = new ClasspathResourceLoader(typeof(TestMorfologikFilterFactory));
IOException expected = NUnit.Framework.Assert.Throws <IOException>(() =>
{
IDictionary <String, String> @params = new JCG.Dictionary <String, String>();
@params[MorfologikFilterFactory.DICTIONARY_ATTRIBUTE] = "missing-dictionary-resource.dict";
MorfologikFilterFactory factory = new MorfologikFilterFactory(@params);
factory.Inform(loader);
});
assertTrue(expected.Message.Contains("Resource not found"));
}
private void Clean()
{
if (_hm.Count == 0)
{
return;
}
var newHm = new JCG.Dictionary <WeakKey <TKey>, TValue>(_hm.Count);
foreach (var entry in _hm.Where(x => x.Key != null && x.Key.IsAlive))
{
newHm.Add(entry.Key, entry.Value);
}
_hm = newHm;
}
public virtual void TestReuseDocsEnumDifferentReader()
{
Directory dir = NewDirectory();
Codec cp = TestUtil.AlwaysPostingsFormat(new Lucene40RWPostingsFormat());
MockAnalyzer analyzer = new MockAnalyzer(Random);
analyzer.MaxTokenLength = TestUtil.NextInt32(Random, 1, IndexWriter.MAX_TERM_LENGTH);
RandomIndexWriter writer = new RandomIndexWriter(Random, dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, analyzer).SetCodec(cp));
int numdocs = AtLeast(20);
CreateRandomIndex(numdocs, writer, Random);
writer.Commit();
DirectoryReader firstReader = DirectoryReader.Open(dir);
DirectoryReader secondReader = DirectoryReader.Open(dir);
IList <AtomicReaderContext> leaves = firstReader.Leaves;
IList <AtomicReaderContext> leaves2 = secondReader.Leaves;
foreach (AtomicReaderContext ctx in leaves)
{
Terms terms = ((AtomicReader)ctx.Reader).GetTerms("body");
TermsEnum iterator = terms.GetEnumerator();
IDictionary <DocsEnum, bool?> enums = new JCG.Dictionary <DocsEnum, bool?>(IdentityEqualityComparer <DocsEnum> .Default);
MatchNoBits bits = new MatchNoBits(firstReader.MaxDoc);
iterator = terms.GetEnumerator();
DocsEnum docs = null;
BytesRef term = null;
while (iterator.MoveNext())
{
term = iterator.Term;
docs = iterator.Docs(null, RandomDocsEnum("body", term, leaves2, bits), Random.NextBoolean() ? DocsFlags.FREQS : DocsFlags.NONE);
enums[docs] = true;
}
Assert.AreEqual(terms.Count, enums.Count);
iterator = terms.GetEnumerator();
enums.Clear();
docs = null;
while (iterator.MoveNext())
{
term = iterator.Term;
docs = iterator.Docs(bits, RandomDocsEnum("body", term, leaves2, bits), Random.NextBoolean() ? DocsFlags.FREQS : DocsFlags.NONE);
enums[docs] = true;
}
Assert.AreEqual(terms.Count, enums.Count);
}
IOUtils.Dispose(writer, firstReader, secondReader, dir);
}
public void TestAUnicodeMapInverse()
{
UnicodeMap <Char> foo1 = new UnicodeMap <Char>()
.PutAll('a', 'z', 'b')
.Put("ab", 'c')
.Put('x', 'b')
.Put("xy", 'c')
;
IDictionary <Char, UnicodeSet> target = new JCG.Dictionary <Char, UnicodeSet>();
foo1.AddInverseTo(target);
UnicodeMap <Char> reverse = new UnicodeMap <Char>().PutAllInverse(target);
assertEquals("", foo1, reverse);
}
private IDictionary <string, long> BytesUsedByExtension(Directory d)
{
IDictionary <string, long> bytesUsedByExtension = new JCG.Dictionary <string, long>();
foreach (string file in d.ListAll())
{
string ext = IndexFileNames.GetExtension(file);
long previousLength = bytesUsedByExtension.TryGetValue(ext, out long length) ? length : 0;
bytesUsedByExtension[ext] = previousLength + d.FileLength(file);
}
foreach (string item in ExcludedExtensionsFromByteCounts)
{
bytesUsedByExtension.Remove(item);
}
return(bytesUsedByExtension);
}
public void TestExplicitDictionary()
{
IResourceLoader loader = new ClasspathResourceLoader(typeof(TestMorfologikFilterFactory));
StringReader reader = new StringReader("inflected1 inflected2");
IDictionary <String, String> @params = new JCG.Dictionary <string, string>();
@params[MorfologikFilterFactory.DICTIONARY_ATTRIBUTE] = "custom-dictionary.dict";
MorfologikFilterFactory factory = new MorfologikFilterFactory(@params);
factory.Inform(loader);
TokenStream stream = new MockTokenizer(reader); // whitespaceMockTokenizer(reader);
stream = factory.Create(stream);
AssertTokenStreamContents(stream, new String[] { "lemma1", "lemma2" });
}
public void TestDateRange()
{
String startDate = getLocalizedDate(2002, 1, 1, false);
String endDate = getLocalizedDate(2002, 1, 4, false);
// we use the default Locale/TZ since LuceneTestCase randomizes it
//Calendar endDateExpected = new GregorianCalendar(TimeZone.getDefault(), Locale.getDefault());
//endDateExpected.set(2002, 1, 4, 23, 59, 59);
//endDateExpected.set(Calendar.MILLISECOND, 999);
DateTime endDateExpected = new GregorianCalendar().ToDateTime(2002, 1, 4, 23, 59, 59, 999);
String defaultField = "default";
String monthField = "month";
String hourField = "hour";
PrecedenceQueryParser qp = new PrecedenceQueryParser(new MockAnalyzer(Random));
IDictionary <string, DateResolution> fieldMap = new JCG.Dictionary <string, DateResolution>();
// set a field specific date resolution
fieldMap.Put(monthField, DateResolution.MONTH);
#pragma warning disable 612, 618
qp.SetDateResolution(fieldMap);
#pragma warning restore 612, 618
// set default date resolution to MILLISECOND
qp.SetDateResolution(DateResolution.MILLISECOND);
// set second field specific date resolution
fieldMap.Put(hourField, DateResolution.HOUR);
#pragma warning disable 612, 618
qp.SetDateResolution(fieldMap);
#pragma warning restore 612, 618
// for this field no field specific date resolution has been set,
// so verify if the default resolution is used
assertDateRangeQueryEquals(qp, defaultField, startDate, endDate,
endDateExpected, DateResolution.MILLISECOND);
// verify if field specific date resolutions are used for these two fields
assertDateRangeQueryEquals(qp, monthField, startDate, endDate,
endDateExpected, DateResolution.MONTH);
assertDateRangeQueryEquals(qp, hourField, startDate, endDate,
endDateExpected, DateResolution.HOUR);
}
public virtual void TestNestedPulsing()
{
// we always run this test with pulsing codec.
Codec cp = TestUtil.AlwaysPostingsFormat(new NestedPulsingPostingsFormat());
BaseDirectoryWrapper dir = NewDirectory();
RandomIndexWriter iw = new RandomIndexWriter(Random, dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random)).SetCodec(cp));
Document doc = new Document();
doc.Add(new TextField("foo", "a b b c c c d e f g g g h i i j j k l l m m m", Field.Store.NO));
// note: the reuse is imperfect, here we would have 4 enums (lost reuse when we get an enum for 'm')
// this is because we only track the 'last' enum we reused (not all).
// but this seems 'good enough' for now.
iw.AddDocument(doc);
DirectoryReader ir = iw.GetReader();
iw.Dispose();
AtomicReader segment = GetOnlySegmentReader(ir);
DocsEnum reuse = null;
IDictionary <DocsEnum, bool?> allEnums = new JCG.Dictionary <DocsEnum, bool?>(IdentityEqualityComparer <DocsEnum> .Default);
TermsEnum te = segment.GetTerms("foo").GetEnumerator();
while (te.MoveNext())
{
reuse = te.Docs(null, reuse, DocsFlags.NONE);
allEnums[reuse] = true;
}
assertEquals(4, allEnums.Count);
allEnums.Clear();
DocsAndPositionsEnum posReuse = null;
te = segment.GetTerms("foo").GetEnumerator();
while (te.MoveNext())
{
posReuse = te.DocsAndPositions(null, posReuse);
allEnums[posReuse] = true;
}
assertEquals(4, allEnums.Count);
ir.Dispose();
dir.Dispose();
}
[Timeout(120000)] // ICU4N: This test can take awhile because of the slowness of adding items to SortedSet
public void TestUnicodeMapGeneralCategory()
{
Logln("Setting General Category");
UnicodeMap <String> map1 = new UnicodeMap <string>();
IDictionary <Integer, String> map2 = new JCG.Dictionary <Integer, String>();
//Map<Integer, String> map3 = new TreeMap<Integer, String>();
map1 = new UnicodeMap <String>();
map2 = new JCG.SortedDictionary <Integer, String>();
for (int cp = 0; cp <= SET_LIMIT; ++cp)
{
int enumValue = UChar.GetIntPropertyValue(cp, propEnum);
//if (enumValue <= 0) continue; // for smaller set
String value = UChar.GetPropertyValueName(propEnum, enumValue, NameChoice.Long);
map1.Put(cp, value);
map2[new Integer(cp)] = value;
}
checkNext(map1, map2, int.MaxValue);
Logln("Comparing General Category");
check(map1, map2, -1);
Logln("Comparing Values");
ISet <String> values1 = new JCG.SortedSet <String>(StringComparer.Ordinal); map1.GetAvailableValues(values1);
ISet <String> values2 = new JCG.SortedSet <String>(map2.Values.Distinct(), StringComparer.Ordinal); // ICU4N NOTE: Added Distinct()
if (!TestBoilerplate <string> .VerifySetsIdentical(this, values1, values2))
{
throw new ArgumentException("Halting");
}
Logln("Comparing Sets");
foreach (string value in values1)
{
Logln(value == null ? "null" : value);
UnicodeSet set1 = map1.KeySet(value);
UnicodeSet set2 = TestBoilerplate <string> .GetSet(map2, value);
if (!TestBoilerplate <string> .VerifySetsIdentical(this, set1, set2))
{
throw new ArgumentException("Halting");
}
}
}
public virtual void TestSophisticatedReuse()
{
// we always run this test with pulsing codec.
Codec cp = TestUtil.AlwaysPostingsFormat(new Pulsing41PostingsFormat(1));
Directory dir = NewDirectory();
RandomIndexWriter iw = new RandomIndexWriter(Random, dir, NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(Random)).SetCodec(cp));
Document doc = new Document();
doc.Add(new TextField("foo", "a b b c c c d e f g g h i i j j k", Field.Store.NO));
iw.AddDocument(doc);
DirectoryReader ir = iw.GetReader();
iw.Dispose();
AtomicReader segment = GetOnlySegmentReader(ir);
DocsEnum reuse = null;
IDictionary <DocsEnum, bool?> allEnums = new JCG.Dictionary <DocsEnum, bool?>(IdentityEqualityComparer <DocsEnum> .Default);
TermsEnum te = segment.GetTerms("foo").GetEnumerator();
while (te.MoveNext())
{
reuse = te.Docs(null, reuse, DocsFlags.NONE);
allEnums[reuse] = true;
}
assertEquals(2, allEnums.Count);
allEnums.Clear();
DocsAndPositionsEnum posReuse = null;
te = segment.GetTerms("foo").GetEnumerator();
while (te.MoveNext())
{
posReuse = te.DocsAndPositions(null, posReuse);
allEnums[posReuse] = true;
}
assertEquals(2, allEnums.Count);
ir.Dispose();
dir.Dispose();
}
private int[] CreateExpectedGroupHeads(string searchTerm, GroupDoc[] groupDocs, Sort docSort, bool sortByScoreOnly, int[] fieldIdToDocID)
{
IDictionary <BytesRef, List <GroupDoc> > groupHeads = new JCG.Dictionary <BytesRef, List <GroupDoc> >();
foreach (GroupDoc groupDoc in groupDocs)
{
if (!groupDoc.content.StartsWith(searchTerm, StringComparison.Ordinal))
{
continue;
}
if (!groupHeads.TryGetValue(groupDoc.group, out List <GroupDoc> grouphead))
{
List <GroupDoc> list = new List <GroupDoc>();
list.Add(groupDoc);
groupHeads[groupDoc.group] = list;
continue;
}
grouphead.Add(groupDoc);
}
int[] allGroupHeads = new int[groupHeads.Count];
int i = 0;
foreach (BytesRef groupValue in groupHeads.Keys)
{
List <GroupDoc> docs = groupHeads[groupValue];
// LUCENENET TODO: The original API Collections.Sort does not currently exist.
// This call ultimately results in calling TimSort, which is why this line was replaced
// with CollectionUtil.TimSort(IList<T>, IComparer<T>).
//
// NOTE: List.Sort(comparer) won't work in this case because it calls the comparer when the
// values are the same, which results in this test failing. TimSort only calls the comparer
// when the values differ.
//Collections.Sort(docs, GetComparer(docSort, sortByScoreOnly, fieldIdToDocID));
CollectionUtil.TimSort(docs, GetComparer(docSort, sortByScoreOnly, fieldIdToDocID));
allGroupHeads[i++] = docs[0].id;
}
return(allGroupHeads);
}
private void Clean()
{
if (_hm.Count == 0)
{
return;
}
var newHm = new JCG.Dictionary <WeakKey <TKey>, TValue>(_hm.Count);
foreach (var kvp in _hm)
{
if (kvp.Key.TryGetTarget(out TKey _))
{
// LUCENENET: There is a tiny chance that a call to remove the item
// from the dictionary can happen before this line is executed. Therefore,
// just discard the reference and add it as is, even if it is no longer valid
// in this edge case. It is far more efficient to re-use the same instances, anyway.
newHm.Add(kvp.Key, kvp.Value);
}
}
_hm = newHm;
}
public void TestModify()
{
Random random = new Random(0);
UnicodeMap <string> unicodeMap = new UnicodeMap <string>();
JCG.Dictionary <int, string> hashMap = new JCG.Dictionary <int, string>();
String[] values = { null, "the", "quick", "brown", "fox" };
for (int count = 1; count <= MODIFY_TEST_ITERATIONS; ++count)
{
String value = values[random.Next(values.Length)];
int start = random.Next(MODIFY_TEST_LIMIT); // test limited range
int end = random.Next(MODIFY_TEST_LIMIT);
if (start > end)
{
int temp = start;
start = end;
end = temp;
}
int modCount = count & 0xFF;
if (modCount == 0 && IsVerbose())
{
Logln("***" + count);
Logln(unicodeMap.ToString());
}
unicodeMap.PutAll(start, end, value);
if (modCount == 1 && IsVerbose())
{
Logln(">>>\t" + Utility.Hex(start) + ".." + Utility.Hex(end) + "\t" + value);
Logln(unicodeMap.ToString());
}
for (int i = start; i <= end; ++i)
{
hashMap[i] = value;
}
if (!hasSameValues(unicodeMap, hashMap))
{
Errln("Failed at " + count);
}
}
}
// FST is pruned
private void VerifyPruned(int inputMode, FST <T> fst, int prune1, int prune2)
{
if (LuceneTestCase.VERBOSE)
{
Console.WriteLine("TEST: now verify pruned " + pairs.Count + " terms; outputs=" + outputs);
foreach (InputOutput <T> pair in pairs)
{
Console.WriteLine(" " + InputToString(inputMode, pair.Input) + ": " + outputs.OutputToString(pair.Output));
}
}
// To validate the FST, we brute-force compute all prefixes
// in the terms, matched to their "common" outputs, prune that
// set according to the prune thresholds, then assert the FST
// matches that same set.
// NOTE: Crazy RAM intensive!!
//System.out.println("TEST: tally prefixes");
// build all prefixes
IDictionary <Int32sRef, CountMinOutput <T> > prefixes = new JCG.Dictionary <Int32sRef, CountMinOutput <T> >();
Int32sRef scratch = new Int32sRef(10);
foreach (InputOutput <T> pair in pairs)
{
scratch.CopyInt32s(pair.Input);
for (int idx = 0; idx <= pair.Input.Length; idx++)
{
scratch.Length = idx;
if (!prefixes.TryGetValue(scratch, out CountMinOutput <T> cmo) || cmo == null)
{
cmo = new CountMinOutput <T>();
cmo.Count = 1;
cmo.Output = pair.Output;
prefixes[Int32sRef.DeepCopyOf(scratch)] = cmo;
}
else
{
cmo.Count++;
T output1 = cmo.Output;
if (output1.Equals(outputs.NoOutput))
{
output1 = outputs.NoOutput;
}
T output2 = pair.Output;
if (output2.Equals(outputs.NoOutput))
{
output2 = outputs.NoOutput;
}
cmo.Output = outputs.Common(output1, output2);
}
if (idx == pair.Input.Length)
{
cmo.IsFinal = true;
cmo.FinalOutput = cmo.Output;
}
}
}
if (LuceneTestCase.VERBOSE)
{
Console.WriteLine("TEST: now prune");
}
// prune 'em
// LUCENENET NOTE: Altered this a bit to go in reverse rather than use an enumerator since
// in .NET you cannot delete records while enumerating forward through a dictionary.
for (int i = prefixes.Count - 1; i >= 0; i--)
{
KeyValuePair <Int32sRef, CountMinOutput <T> > ent = prefixes.ElementAt(i);
Int32sRef prefix = ent.Key;
CountMinOutput <T> cmo = ent.Value;
if (LuceneTestCase.VERBOSE)
{
Console.WriteLine(" term prefix=" + InputToString(inputMode, prefix, false) + " count=" + cmo.Count + " isLeaf=" + cmo.IsLeaf + " output=" + outputs.OutputToString(cmo.Output) + " isFinal=" + cmo.IsFinal);
}
bool keep;
if (prune1 > 0)
{
keep = cmo.Count >= prune1;
}
else
{
Debug.Assert(prune2 > 0);
if (prune2 > 1 && cmo.Count >= prune2)
{
keep = true;
}
else if (prefix.Length > 0)
{
// consult our parent
scratch.Length = prefix.Length - 1;
Array.Copy(prefix.Int32s, prefix.Offset, scratch.Int32s, 0, scratch.Length);
prefixes.TryGetValue(scratch, out CountMinOutput <T> cmo2);
//System.out.println(" parent count = " + (cmo2 == null ? -1 : cmo2.count));
keep = cmo2 != null && ((prune2 > 1 && cmo2.Count >= prune2) || (prune2 == 1 && (cmo2.Count >= 2 || prefix.Length <= 1)));
}
else if (cmo.Count >= prune2)
{
keep = true;
}
else
{
keep = false;
}
}
if (!keep)
{
prefixes.Remove(prefix);
//System.out.println(" remove");
}
else
{
// clear isLeaf for all ancestors
//System.out.println(" keep");
scratch.CopyInt32s(prefix);
scratch.Length--;
while (scratch.Length >= 0)
{
if (prefixes.TryGetValue(scratch, out CountMinOutput <T> cmo2) && cmo2 != null)
{
//System.out.println(" clear isLeaf " + inputToString(inputMode, scratch));
cmo2.IsLeaf = false;
}
scratch.Length--;
}
}
}
if (LuceneTestCase.VERBOSE)
{
Console.WriteLine("TEST: after prune");
foreach (KeyValuePair <Int32sRef, CountMinOutput <T> > ent in prefixes)
{
Console.WriteLine(" " + InputToString(inputMode, ent.Key, false) + ": isLeaf=" + ent.Value.IsLeaf + " isFinal=" + ent.Value.IsFinal);
if (ent.Value.IsFinal)
{
Console.WriteLine(" finalOutput=" + outputs.OutputToString(ent.Value.FinalOutput));
}
}
}
if (prefixes.Count <= 1)
{
Assert.IsNull(fst);
return;
}
Assert.IsNotNull(fst);
// make sure FST only enums valid prefixes
if (LuceneTestCase.VERBOSE)
{
Console.WriteLine("TEST: check pruned enum");
}
Int32sRefFSTEnum <T> fstEnum = new Int32sRefFSTEnum <T>(fst);
Int32sRefFSTEnum.InputOutput <T> current;
while ((current = fstEnum.Next()) != null)
{
if (LuceneTestCase.VERBOSE)
{
Console.WriteLine(" fstEnum.next prefix=" + InputToString(inputMode, current.Input, false) + " output=" + outputs.OutputToString(current.Output));
}
prefixes.TryGetValue(current.Input, out CountMinOutput <T> cmo);
Assert.IsNotNull(cmo);
Assert.IsTrue(cmo.IsLeaf || cmo.IsFinal);
//if (cmo.isFinal && !cmo.isLeaf) {
if (cmo.IsFinal)
{
Assert.AreEqual(cmo.FinalOutput, current.Output);
}
else
{
Assert.AreEqual(cmo.Output, current.Output);
}
}
// make sure all non-pruned prefixes are present in the FST
if (LuceneTestCase.VERBOSE)
{
Console.WriteLine("TEST: verify all prefixes");
}
int[] stopNode = new int[1];
foreach (KeyValuePair <Int32sRef, CountMinOutput <T> > ent in prefixes)
{
if (ent.Key.Length > 0)
{
CountMinOutput <T> cmo = ent.Value;
T output = Run(fst, ent.Key, stopNode);
if (LuceneTestCase.VERBOSE)
{
Console.WriteLine("TEST: verify prefix=" + InputToString(inputMode, ent.Key, false) + " output=" + outputs.OutputToString(cmo.Output));
}
// if (cmo.isFinal && !cmo.isLeaf) {
if (cmo.IsFinal)
{
Assert.AreEqual(cmo.FinalOutput, output);
}
else
{
Assert.AreEqual(cmo.Output, output);
}
Assert.AreEqual(ent.Key.Length, stopNode[0]);
}
}
}
/// <summary>
/// Fills a <see cref="T:IDictionary{string, WeightedSpanTerm}"/> with <see cref="WeightedSpanTerm"/>s using the terms from the supplied <see cref="SpanQuery"/>.
/// </summary>
/// <param name="terms"><see cref="T:IDictionary{string, WeightedSpanTerm}"/> to place created <see cref="WeightedSpanTerm"/>s in</param>
/// <param name="spanQuery"><see cref="SpanQuery"/> to extract Terms from</param>
/// <exception cref="IOException">If there is a low-level I/O error</exception>
protected virtual void ExtractWeightedSpanTerms(IDictionary <string, WeightedSpanTerm> terms, SpanQuery spanQuery)
{
ISet <string> fieldNames;
if (fieldName == null)
{
fieldNames = new JCG.HashSet <string>();
CollectSpanQueryFields(spanQuery, fieldNames);
}
else
{
fieldNames = new JCG.HashSet <string>
{
fieldName
};
}
// To support the use of the default field name
if (defaultField != null)
{
fieldNames.Add(defaultField);
}
IDictionary <string, SpanQuery> queries = new JCG.Dictionary <string, SpanQuery>();
var nonWeightedTerms = new JCG.HashSet <Term>();
bool mustRewriteQuery = MustRewriteQuery(spanQuery);
if (mustRewriteQuery)
{
foreach (string field in fieldNames)
{
SpanQuery rewrittenQuery = (SpanQuery)spanQuery.Rewrite(GetLeafContext().Reader);
queries[field] = rewrittenQuery;
rewrittenQuery.ExtractTerms(nonWeightedTerms);
}
}
else
{
spanQuery.ExtractTerms(nonWeightedTerms);
}
List <PositionSpan> spanPositions = new List <PositionSpan>();
foreach (string field in fieldNames)
{
SpanQuery q;
q = mustRewriteQuery ? queries[field] : spanQuery;
AtomicReaderContext context = GetLeafContext();
var termContexts = new JCG.Dictionary <Term, TermContext>();
ISet <Term> extractedTerms = new JCG.SortedSet <Term>();
q.ExtractTerms(extractedTerms);
foreach (Term term in extractedTerms)
{
termContexts[term] = TermContext.Build(context, term);
}
IBits acceptDocs = context.AtomicReader.LiveDocs;
Spans.Spans spans = q.GetSpans(context, acceptDocs, termContexts);
// collect span positions
while (spans.MoveNext())
{
spanPositions.Add(new PositionSpan(spans.Start, spans.End - 1));
}
}
if (spanPositions.Count == 0)
{
// no spans found
return;
}
foreach (Term queryTerm in nonWeightedTerms)
{
if (FieldNameComparer(queryTerm.Field))
{
if (!terms.TryGetValue(queryTerm.Text, out WeightedSpanTerm weightedSpanTerm) || weightedSpanTerm == null)
{
weightedSpanTerm = new WeightedSpanTerm(spanQuery.Boost, queryTerm.Text);
weightedSpanTerm.AddPositionSpans(spanPositions);
weightedSpanTerm.IsPositionSensitive = true;
terms[queryTerm.Text] = weightedSpanTerm;
}
else
{
if (spanPositions.Count > 0)
{
weightedSpanTerm.AddPositionSpans(spanPositions);
}
}
}
}
}
private GroupedFacetResult CreateExpectedFacetResult(string searchTerm, IndexContext context, int offset, int limit, int minCount, bool orderByCount, string facetPrefix)
{
JCG.Dictionary <string, ISet <string> > facetGroups;
if (!context.searchTermToFacetGroups.TryGetValue(searchTerm, out facetGroups))
{
facetGroups = new JCG.Dictionary <string, ISet <string> >();
}
int totalCount = 0;
int totalMissCount = 0;
ISet <string> facetValues;
if (facetPrefix != null)
{
facetValues = new JCG.HashSet <string>();
foreach (string facetValue in context.facetValues)
{
if (facetValue != null && facetValue.StartsWith(facetPrefix, StringComparison.Ordinal))
{
facetValues.add(facetValue);
}
}
}
else
{
facetValues = context.facetValues;
}
List <TermGroupFacetCollector.FacetEntry> entries = new List <TermGroupFacetCollector.FacetEntry>(facetGroups.size());
// also includes facets with count 0
foreach (string facetValue in facetValues)
{
if (facetValue == null)
{
continue;
}
int count = facetGroups.TryGetValue(facetValue, out ISet <string> groups) && groups != null?groups.size() : 0;
if (count >= minCount)
{
entries.Add(new TermGroupFacetCollector.FacetEntry(new BytesRef(facetValue), count));
}
totalCount += count;
}
// Only include null count when no facet prefix is specified
if (facetPrefix == null)
{
if (facetGroups.TryGetValue(null, out ISet <string> groups) && groups != null)
{
totalMissCount = groups.size();
}
}
entries.Sort(Comparer <TermGroupFacetCollector.FacetEntry> .Create((a, b) => {
if (orderByCount)
{
int cmp = b.Count - a.Count;
if (cmp != 0)
{
return(cmp);
}
}
return(a.Value.CompareTo(b.Value));
}));
int endOffset = offset + limit;
IList <TermGroupFacetCollector.FacetEntry> entriesResult;
if (offset >= entries.size())
{
entriesResult = Collections.EmptyList <TermGroupFacetCollector.FacetEntry>();
}
else if (endOffset >= entries.size())
{
entriesResult = entries.GetRange(offset, entries.size() - offset);
}
else
{
entriesResult = entries.GetRange(offset, endOffset - offset);
}
return(new GroupedFacetResult(totalCount, totalMissCount, entriesResult));
}
/** Returns Pair(list of invalid document terms, Map of document term -> document) */
private KeyValuePair <List <string>, IDictionary <string, Document> > GenerateIndexDocuments(int ndocs, bool requiresPayload, bool requiresContexts)
{
IDictionary <string, Document> docs = new JCG.Dictionary <string, Document>();
List <string> invalidDocTerms = new List <string>();
for (int i = 0; i < ndocs; i++)
{
Document doc = new Document();
bool invalidDoc = false;
Field field = null;
// usually have valid term field in document
if (Usually())
{
field = new TextField(FIELD_NAME, "field_" + i, Field.Store.YES);
doc.Add(field);
}
else
{
invalidDoc = true;
}
// even if payload is not required usually have it
if (requiresPayload || Usually())
{
// usually have valid payload field in document
if (Usually())
{
Field payload = new StoredField(PAYLOAD_FIELD_NAME, new BytesRef("payload_" + i));
doc.Add(payload);
}
else if (requiresPayload)
{
invalidDoc = true;
}
}
if (requiresContexts || Usually())
{
if (Usually())
{
for (int j = 0; j < AtLeast(2); j++)
{
doc.Add(new StoredField(CONTEXT_FIELD_NAME, new BytesRef("context_" + i + "_" + j)));
}
}
// we should allow entries without context
}
// usually have valid weight field in document
if (Usually())
{
Field weight = (Rarely()) ?
(Field) new StoredField(WEIGHT_FIELD_NAME, 100d + i) :
(Field) new NumericDocValuesField(WEIGHT_FIELD_NAME, 100 + i);
doc.Add(weight);
}
string term = null;
if (invalidDoc)
{
term = (field != null) ? field.GetStringValue() : "invalid_" + i;
invalidDocTerms.Add(term);
}
else
{
term = field.GetStringValue();
}
docs.Put(term, doc);
}
return(new KeyValuePair <List <string>, IDictionary <string, Document> >(invalidDocTerms, docs));
}
private IndexContext CreateIndexContext()
{
Random random = Random;
DocValuesType[] dvTypes = new DocValuesType[] {
DocValuesType.BINARY,
DocValuesType.SORTED
};
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
random,
dir,
NewIndexWriterConfig(TEST_VERSION_CURRENT,
new MockAnalyzer(random)).SetMergePolicy(NewLogMergePolicy())
);
bool canUseDV = !"Lucene3x".Equals(w.IndexWriter.Config.Codec.Name, StringComparison.Ordinal);
DocValuesType dvType = canUseDV ? dvTypes[random.nextInt(dvTypes.Length)] : DocValuesType.NONE;
int numDocs = 86 + random.nextInt(1087) * RANDOM_MULTIPLIER;
string[] groupValues = new string[numDocs / 5];
string[] countValues = new string[numDocs / 10];
for (int i = 0; i < groupValues.Length; i++)
{
groupValues[i] = GenerateRandomNonEmptyString();
}
for (int i = 0; i < countValues.Length; i++)
{
countValues[i] = GenerateRandomNonEmptyString();
}
List <string> contentStrings = new List <string>();
IDictionary <string, IDictionary <string, ISet <string> > > searchTermToGroupCounts = new JCG.Dictionary <string, IDictionary <string, ISet <string> > >();
for (int i = 1; i <= numDocs; i++)
{
string groupValue = random.nextInt(23) == 14 ? null : groupValues[random.nextInt(groupValues.Length)];
string countValue = random.nextInt(21) == 13 ? null : countValues[random.nextInt(countValues.Length)];
string content = "random" + random.nextInt(numDocs / 20);
IDictionary <string, ISet <string> > groupToCounts;
if (!searchTermToGroupCounts.TryGetValue(content, out groupToCounts))
{
// Groups sort always DOCID asc...
searchTermToGroupCounts.Add(content, groupToCounts = new JCG.LinkedDictionary <string, ISet <string> >());
contentStrings.Add(content);
}
ISet <string> countsVals;
if (!groupToCounts.TryGetValue(groupValue, out countsVals))
{
groupToCounts.Add(groupValue, countsVals = new JCG.HashSet <string>());
}
countsVals.Add(countValue);
Document doc = new Document();
doc.Add(new StringField("id", string.Format(CultureInfo.InvariantCulture, "{0:D9}", i), Field.Store.YES));
if (groupValue != null)
{
AddField(doc, groupField, groupValue, dvType);
}
if (countValue != null)
{
AddField(doc, countField, countValue, dvType);
}
doc.Add(new TextField("content", content, Field.Store.YES));
w.AddDocument(doc);
}
DirectoryReader reader = w.GetReader();
if (VERBOSE)
{
for (int docID = 0; docID < reader.MaxDoc; docID++)
{
Document doc = reader.Document(docID);
Console.WriteLine("docID=" + docID + " id=" + doc.Get("id") + " content=" + doc.Get("content") + " author=" + doc.Get("author") + " publisher=" + doc.Get("publisher"));
}
}
w.Dispose();
return(new IndexContext(dir, reader, dvType, searchTermToGroupCounts, contentStrings.ToArray(/*new String[contentStrings.size()]*/)));
}
private static IDictionary <string, IList <Rule> > ParseRules(TextReader reader, string location)
{
IDictionary <string, IList <Rule> > lines = new JCG.Dictionary <string, IList <Rule> >();
int currentLine = 0;
bool inMultilineComment = false;
string rawLine;
try
{
while ((rawLine = reader.ReadLine()) != null)
{
currentLine++;
string line = rawLine;
if (inMultilineComment)
{
if (line.EndsWith(ResourceConstants.EXT_CMT_END, StringComparison.Ordinal))
{
inMultilineComment = false;
}
}
else
{
if (line.StartsWith(ResourceConstants.EXT_CMT_START, StringComparison.Ordinal))
{
inMultilineComment = true;
}
else
{
// discard comments
int cmtI = line.IndexOf(ResourceConstants.CMT, StringComparison.Ordinal);
if (cmtI >= 0)
{
line = line.Substring(0, cmtI);
}
// trim leading-trailing whitespace
line = line.Trim();
if (line.Length == 0)
{
continue; // empty lines can be safely skipped
}
if (line.StartsWith(HASH_INCLUDE, StringComparison.Ordinal))
{
// include statement
string incl = line.Substring(HASH_INCLUDE.Length).Trim();
if (incl.Contains(" "))
{
throw new ArgumentException("Malformed import statement '" + rawLine + "' in " +
location);
}
else
{
lines.PutAll(ParseRules(CreateScanner(incl), location + "->" + incl));
}
}
else
{
// rule
string[] parts = WHITESPACE.Split(line).TrimEnd();
if (parts.Length != 4)
{
throw new ArgumentException("Malformed rule statement split into " + parts.Length +
" parts: " + rawLine + " in " + location);
}
else
{
try
{
string pat = StripQuotes(parts[0]);
string lCon = StripQuotes(parts[1]);
string rCon = StripQuotes(parts[2]);
IPhonemeExpr ph = ParsePhonemeExpr(StripQuotes(parts[3]));
int cLine = currentLine;
Rule r = new RuleAnonymousClass(pat, lCon, rCon, ph, cLine, location);
string patternKey = r.pattern.Substring(0, 1 - 0);
if (!lines.TryGetValue(patternKey, out IList <Rule> rules) || rules == null)
{
rules = new List <Rule>();
lines[patternKey] = rules;
}
rules.Add(r);
}
catch (Exception e) when(e.IsIllegalArgumentException())
{
throw new InvalidOperationException("Problem parsing line '" + currentLine + "' in " +
location, e);
}
}
}
}
}
}
}
finally
{
reader.Dispose();
}
return(lines);
}
/*
* Non-recursive version of object descend. this consumes more memory than recursive in-depth
* traversal but prevents stack overflows on long chains of objects
* or complex graphs (a max. recursion depth on my machine was ~5000 objects linked in a chain
* so not too much).
*/
private static long MeasureObjectSize(object root)
{
// Objects seen so far.
IdentityHashSet <object> seen = new IdentityHashSet <object>();
// Class cache with reference Field and precalculated shallow size.
IDictionary <Type, ClassCache> classCache = new JCG.Dictionary <Type, ClassCache>(IdentityEqualityComparer <Type> .Default);
// Stack of objects pending traversal. Recursion caused stack overflows.
Stack <object> stack = new Stack <object>();
stack.Push(root);
long totalSize = 0;
while (stack.Count > 0)
{
object ob = stack.Pop();
if (ob is null || seen.Contains(ob))
{
continue;
}
seen.Add(ob);
Type obClazz = ob.GetType();
// LUCENENET specific - .NET cannot return a null type for an object, so no need to assert it
if (obClazz.Equals(typeof(string)))
{
// LUCENENET specific - we can get a closer estimate of a string
// by using simple math. Reference: http://stackoverflow.com/a/8171099.
// This fixes the TestSanity test.
totalSize += (2 * (((string)ob).Length + 1));
}
if (obClazz.IsArray)
{
/*
* Consider an array, possibly of primitive types. Push any of its references to
* the processing stack and accumulate this array's shallow size.
*/
long size = NUM_BYTES_ARRAY_HEADER;
Array array = (Array)ob;
int len = array.Length;
if (len > 0)
{
Type componentClazz = obClazz.GetElementType();
if (componentClazz.IsPrimitive)
{
size += (long)len * primitiveSizes[componentClazz];
}
else
{
size += (long)NUM_BYTES_OBJECT_REF * len;
// Push refs for traversal later.
for (int i = len; --i >= 0;)
{
object o = array.GetValue(i);
if (o != null && !seen.Contains(o))
{
stack.Push(o);
}
}
}
}
totalSize += AlignObjectSize(size);
}
else
{
/*
* Consider an object. Push any references it has to the processing stack
* and accumulate this object's shallow size.
*/
try
{
if (!classCache.TryGetValue(obClazz, out ClassCache cachedInfo) || cachedInfo is null)
{
classCache[obClazz] = cachedInfo = CreateCacheEntry(obClazz);
}
foreach (FieldInfo f in cachedInfo.ReferenceFields)
{
// Fast path to eliminate redundancies.
object o = f.GetValue(ob);
if (o != null && !seen.Contains(o))
{
stack.Push(o);
}
}
totalSize += cachedInfo.AlignedShallowInstanceSize;
}
catch (Exception e) when(e.IsIllegalAccessException())
{
// this should never happen as we enabled setAccessible().
throw RuntimeException.Create("Reflective field access failed?", e);
}
}
}
// Help the GC (?).
seen.Clear();
stack.Clear();
classCache.Clear();
return(totalSize);
}
private IndexContext CreateIndexContext(bool multipleFacetValuesPerDocument)
{
Random random = Random;
int numDocs = TestUtil.NextInt32(random, 138, 1145) * RandomMultiplier;
int numGroups = TestUtil.NextInt32(random, 1, numDocs / 4);
int numFacets = TestUtil.NextInt32(random, 1, numDocs / 6);
if (Verbose)
{
Console.WriteLine("TEST: numDocs=" + numDocs + " numGroups=" + numGroups);
}
List <string> groups = new List <string>();
for (int i = 0; i < numGroups; i++)
{
groups.Add(GenerateRandomNonEmptyString());
}
List <string> facetValues = new List <string>();
for (int i = 0; i < numFacets; i++)
{
facetValues.Add(GenerateRandomNonEmptyString());
}
string[] contentBrs = new string[TestUtil.NextInt32(random, 2, 20)];
if (Verbose)
{
Console.WriteLine("TEST: create fake content");
}
for (int contentIDX = 0; contentIDX < contentBrs.Length; contentIDX++)
{
contentBrs[contentIDX] = GenerateRandomNonEmptyString();
if (Verbose)
{
Console.WriteLine(" content=" + contentBrs[contentIDX]);
}
}
Directory dir = NewDirectory();
RandomIndexWriter writer = new RandomIndexWriter(
random,
dir,
NewIndexWriterConfig(
TEST_VERSION_CURRENT,
new MockAnalyzer(random)
)
);
bool canUseDV = !"Lucene3x".Equals(writer.IndexWriter.Config.Codec.Name, StringComparison.Ordinal);
bool useDv = canUseDV && !multipleFacetValuesPerDocument && random.nextBoolean();
Document doc = new Document();
Document docNoGroup = new Document();
Document docNoFacet = new Document();
Document docNoGroupNoFacet = new Document();
Field group = NewStringField("group", "", Field.Store.NO);
Field groupDc = new SortedDocValuesField("group_dv", new BytesRef());
if (useDv)
{
doc.Add(groupDc);
docNoFacet.Add(groupDc);
}
doc.Add(group);
docNoFacet.Add(group);
Field[] facetFields;
if (useDv)
{
Debug.Assert(!multipleFacetValuesPerDocument);
facetFields = new Field[2];
facetFields[0] = NewStringField("facet", "", Field.Store.NO);
doc.Add(facetFields[0]);
docNoGroup.Add(facetFields[0]);
facetFields[1] = new SortedDocValuesField("facet_dv", new BytesRef());
doc.Add(facetFields[1]);
docNoGroup.Add(facetFields[1]);
}
else
{
facetFields = multipleFacetValuesPerDocument ? new Field[2 + random.nextInt(6)] : new Field[1];
for (int i = 0; i < facetFields.Length; i++)
{
facetFields[i] = NewStringField("facet", "", Field.Store.NO);
doc.Add(facetFields[i]);
docNoGroup.Add(facetFields[i]);
}
}
Field content = NewStringField("content", "", Field.Store.NO);
doc.Add(content);
docNoGroup.Add(content);
docNoFacet.Add(content);
docNoGroupNoFacet.Add(content);
ISet <string> uniqueFacetValues = new JCG.SortedSet <string>(Comparer <string> .Create((a, b) => {
if (a == b)
{
return(0);
}
else if (a == null)
{
return(-1);
}
else if (b == null)
{
return(1);
}
else
{
return(a.CompareToOrdinal(b));
}
}));
// LUCENENET NOTE: Need JCG.Dictionary here because of null keys
IDictionary <string, JCG.Dictionary <string, ISet <string> > > searchTermToFacetToGroups = new Dictionary <string, JCG.Dictionary <string, ISet <string> > >();
int facetWithMostGroups = 0;
for (int i = 0; i < numDocs; i++)
{
string groupValue;
if (random.nextInt(24) == 17)
{
// So we test the "doc doesn't have the group'd
// field" case:
if (useDv)
{
groupValue = "";
}
else
{
groupValue = null;
}
}
else
{
groupValue = groups[random.nextInt(groups.size())];
}
string contentStr = contentBrs[random.nextInt(contentBrs.Length)];
if (!searchTermToFacetToGroups.TryGetValue(contentStr, out JCG.Dictionary <string, ISet <string> > facetToGroups))
{
searchTermToFacetToGroups[contentStr] = facetToGroups = new JCG.Dictionary <string, ISet <string> >();
}
List <string> facetVals = new List <string>();
if (useDv || random.nextInt(24) != 18)
{
if (useDv)
{
string facetValue = facetValues[random.nextInt(facetValues.size())];
uniqueFacetValues.Add(facetValue);
if (!facetToGroups.TryGetValue(facetValue, out ISet <string> groupsInFacet))
{
facetToGroups[facetValue] = groupsInFacet = new JCG.HashSet <string>();
}
groupsInFacet.add(groupValue);
if (groupsInFacet.size() > facetWithMostGroups)
{
facetWithMostGroups = groupsInFacet.size();
}
facetFields[0].SetStringValue(facetValue);
facetFields[1].SetBytesValue(new BytesRef(facetValue));
facetVals.Add(facetValue);
}
else
{
foreach (Field facetField in facetFields)
{
string facetValue = facetValues[random.nextInt(facetValues.size())];
uniqueFacetValues.Add(facetValue);
if (!facetToGroups.TryGetValue(facetValue, out ISet <string> groupsInFacet))
{
facetToGroups[facetValue] = groupsInFacet = new JCG.HashSet <string>();
}
groupsInFacet.add(groupValue);
if (groupsInFacet.size() > facetWithMostGroups)
{
facetWithMostGroups = groupsInFacet.size();
}
facetField.SetStringValue(facetValue);
facetVals.Add(facetValue);
}
}
}
else
{
uniqueFacetValues.Add(null);
if (!facetToGroups.TryGetValue(null, out ISet <string> groupsInFacet))
{
facetToGroups[null] = groupsInFacet = new JCG.HashSet <string>();
}
groupsInFacet.add(groupValue);
if (groupsInFacet.size() > facetWithMostGroups)
{
facetWithMostGroups = groupsInFacet.size();
}
}
if (Verbose)
{
Console.WriteLine(" doc content=" + contentStr + " group=" + (groupValue == null ? "null" : groupValue) + " facetVals=" + Collections.ToString(facetVals));
}
if (groupValue != null)
{
if (useDv)
{
groupDc.SetBytesValue(new BytesRef(groupValue));
}
group.SetStringValue(groupValue);
}
else if (useDv)
{
// DV cannot have missing values:
groupDc.SetBytesValue(new BytesRef());
}
content.SetStringValue(contentStr);
if (groupValue == null && facetVals.Count == 0)
{
writer.AddDocument(docNoGroupNoFacet);
}
else if (facetVals.Count == 0)
{
writer.AddDocument(docNoFacet);
}
else if (groupValue == null)
{
writer.AddDocument(docNoGroup);
}
else
{
writer.AddDocument(doc);
}
}
DirectoryReader reader = writer.GetReader();
writer.Dispose();
return(new IndexContext(searchTermToFacetToGroups, reader, numDocs, dir, facetWithMostGroups, numGroups, contentBrs, uniqueFacetValues, useDv));
}
