public virtual void TestRead1waySynonymRules()
{
SlowSynonymMap synMap;
// (a)->[a]
// (b)->[a]
IList <string> rules = new JCG.List <string>();
rules.Add("a,b");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", false, null);
assertEquals(2, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "a");
AssertTokIncludes(synMap, "b", "a");
// (a)->[a]
// (b)->[a]
// (c)->[a]
rules.Clear();
rules.Add("a,b,c");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", false, null);
assertEquals(3, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "a");
AssertTokIncludes(synMap, "b", "a");
AssertTokIncludes(synMap, "c", "a");
// (a)->[a]
// (b1)->(b2)->[a]
rules.Clear();
rules.Add("a,b1 b2");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", false, null);
assertEquals(2, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "a");
assertEquals(1, GetSubSynonymMap(synMap, "b1").Submap.size());
AssertTokIncludes(GetSubSynonymMap(synMap, "b1"), "b2", "a");
// (a1)->(a2)->[a1][a2]
// (b)->[a1][a2]
rules.Clear();
rules.Add("a1 a2,b");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", false, null);
assertEquals(2, synMap.Submap.size());
assertEquals(1, GetSubSynonymMap(synMap, "a1").Submap.size());
AssertTokIncludes(GetSubSynonymMap(synMap, "a1"), "a2", "a1");
AssertTokIncludes(GetSubSynonymMap(synMap, "a1"), "a2", "a2");
AssertTokIncludes(synMap, "b", "a1");
AssertTokIncludes(synMap, "b", "a2");
}
public void TestSearchPhraseSlop()
{
// "a b c"~0
Query query = pqF("a", "b", "c");
// phraseHighlight = true, fieldMatch = true
FieldQuery fq = new FieldQuery(query, true, true);
// "a b c" w/ position-gap = 2
IList <TermInfo> phraseCandidate = new JCG.List <TermInfo>();
phraseCandidate.Add(new TermInfo("a", 0, 1, 0, 1));
phraseCandidate.Add(new TermInfo("b", 2, 3, 2, 1));
phraseCandidate.Add(new TermInfo("c", 4, 5, 4, 1));
assertNull(fq.SearchPhrase(F, phraseCandidate));
// "a b c"~1
query = pqF(1F, 1, "a", "b", "c");
// phraseHighlight = true, fieldMatch = true
fq = new FieldQuery(query, true, true);
// "a b c" w/ position-gap = 2
assertNotNull(fq.SearchPhrase(F, phraseCandidate));
// "a b c" w/ position-gap = 3
phraseCandidate.Clear();
phraseCandidate.Add(new TermInfo("a", 0, 1, 0, 1));
phraseCandidate.Add(new TermInfo("b", 2, 3, 3, 1));
phraseCandidate.Add(new TermInfo("c", 4, 5, 6, 1));
assertNull(fq.SearchPhrase(F, phraseCandidate));
}
/// <summary>Resets stems accumulator and hands over to superclass.</summary>
public override void Reset()
{
lemmaListIndex = 0;
lemmaList = new List <WordData>();
tagsList.Clear();
base.Reset();
}
public virtual void Clear()
{
lock (this)
{
updates.Clear();
nextGen = 1;
numTerms.Value = 0;
bytesUsed.Value = 0;
}
}
/// <summary>
/// Move to the next match, returning true iff any such exists. </summary>
public override bool MoveNext()
{
if (firstTime)
{
firstTime = false;
for (int i = 0; i < subSpans.Length; i++)
{
if (!subSpans[i].MoveNext())
{
more = false;
return(false);
}
}
more = true;
}
if (collectPayloads)
{
matchPayload.Clear();
}
return(AdvanceAfterOrdered());
}
public override bool IncrementToken()
{
if (buffer != null && buffer.Count > 0)
{
CharsRef nextStem = buffer[0];
buffer.RemoveAt(0);
RestoreState(savedState);
posIncAtt.PositionIncrement = 0;
termAtt.SetEmpty().Append(nextStem);
return(true);
}
if (!m_input.IncrementToken())
{
return(false);
}
if (keywordAtt.IsKeyword)
{
return(true);
}
buffer = new JCG.List <CharsRef>(dedup ? stemmer.UniqueStems(termAtt.Buffer, termAtt.Length) : stemmer.Stem(termAtt.Buffer, termAtt.Length));
if (buffer.Count == 0) // we do not know this word, return it unchanged
{
return(true);
}
if (longestOnly && buffer.Count > 1)
{
buffer.Sort(lengthComparer);
}
CharsRef stem = buffer[0];
buffer.RemoveAt(0);
termAtt.SetEmpty().Append(stem);
if (longestOnly)
{
buffer.Clear();
}
else
{
if (buffer.Count > 0)
{
savedState = CaptureState();
}
}
return(true);
}
public virtual void Clear()
{
UninterruptableMonitor.Enter(this);
try
{
updates.Clear();
nextGen = 1;
numTerms.Value = 0;
bytesUsed.Value = 0;
}
finally
{
UninterruptableMonitor.Exit(this);
}
}
public virtual void TestAppendIterator()
{
Random random = Random;
BytesRefArray list = new BytesRefArray(Util.Counter.NewCounter());
IList <string> stringList = new JCG.List <string>();
for (int j = 0; j < 2; j++)
{
if (j > 0 && random.NextBoolean())
{
list.Clear();
stringList.Clear();
}
int entries = AtLeast(500);
BytesRef spare = new BytesRef();
int initSize = list.Length;
for (int i = 0; i < entries; i++)
{
string randomRealisticUnicodeString = TestUtil.RandomRealisticUnicodeString(random);
spare.CopyChars(randomRealisticUnicodeString);
Assert.AreEqual(i + initSize, list.Append(spare));
stringList.Add(randomRealisticUnicodeString);
}
for (int i = 0; i < entries; i++)
{
Assert.IsNotNull(list.Get(spare, i));
Assert.AreEqual(stringList[i], spare.Utf8ToString(), "entry " + i + " doesn't match");
}
// check random
for (int i = 0; i < entries; i++)
{
int e = random.Next(entries);
Assert.IsNotNull(list.Get(spare, e));
Assert.AreEqual(stringList[e], spare.Utf8ToString(), "entry " + i + " doesn't match");
}
for (int i = 0; i < 2; i++)
{
IBytesRefEnumerator iterator = list.GetEnumerator();
foreach (string @string in stringList)
{
Assert.IsTrue(iterator.MoveNext());
Assert.AreEqual(@string, iterator.Current.Utf8ToString());
}
}
}
}
public void TestSearchPhrase()
{
Query query = pqF("a", "b", "c");
// phraseHighlight = true, fieldMatch = true
FieldQuery fq = new FieldQuery(query, true, true);
// "a"
IList <TermInfo> phraseCandidate = new JCG.List <TermInfo>();
phraseCandidate.Add(new TermInfo("a", 0, 1, 0, 1));
assertNull(fq.SearchPhrase(F, phraseCandidate));
// "a b"
phraseCandidate.Add(new TermInfo("b", 2, 3, 1, 1));
assertNull(fq.SearchPhrase(F, phraseCandidate));
// "a b c"
phraseCandidate.Add(new TermInfo("c", 4, 5, 2, 1));
assertNotNull(fq.SearchPhrase(F, phraseCandidate));
assertNull(fq.SearchPhrase("x", phraseCandidate));
// phraseHighlight = true, fieldMatch = false
fq = new FieldQuery(query, true, false);
// "a b c"
assertNotNull(fq.SearchPhrase(F, phraseCandidate));
assertNotNull(fq.SearchPhrase("x", phraseCandidate));
// phraseHighlight = false, fieldMatch = true
fq = new FieldQuery(query, false, true);
// "a"
phraseCandidate.Clear();
phraseCandidate.Add(new TermInfo("a", 0, 1, 0, 1));
assertNotNull(fq.SearchPhrase(F, phraseCandidate));
// "a b"
phraseCandidate.Add(new TermInfo("b", 2, 3, 1, 1));
assertNull(fq.SearchPhrase(F, phraseCandidate));
// "a b c"
phraseCandidate.Add(new TermInfo("c", 4, 5, 2, 1));
assertNotNull(fq.SearchPhrase(F, phraseCandidate));
assertNull(fq.SearchPhrase("x", phraseCandidate));
}
public virtual void TestSortIterator()
{
Random random = Random;
BytesRefArray list = new BytesRefArray(Util.Counter.NewCounter());
IList <string> stringList = new JCG.List <string>();
for (int j = 0; j < 2; j++)
{
if (j > 0 && random.NextBoolean())
{
list.Clear();
stringList.Clear();
}
int entries = AtLeast(500);
BytesRef spare = new BytesRef();
int initSize = list.Length;
for (int i = 0; i < entries; i++)
{
string randomRealisticUnicodeString = TestUtil.RandomRealisticUnicodeString(random);
spare.CopyChars(randomRealisticUnicodeString);
Assert.AreEqual(initSize + i, list.Append(spare));
stringList.Add(randomRealisticUnicodeString);
}
// LUCENENET NOTE: Must sort using ArrayUtil.GetNaturalComparator<T>()
// to ensure culture isn't taken into consideration during the sort,
// which will match the sort order of BytesRef.UTF8SortedAsUTF16Comparer.
CollectionUtil.TimSort(stringList);
#pragma warning disable 612, 618
IBytesRefIterator iter = list.GetIterator(BytesRef.UTF8SortedAsUTF16Comparer);
#pragma warning restore 612, 618
int a = 0;
while ((spare = iter.Next()) != null)
{
Assert.AreEqual(stringList[a], spare.Utf8ToString(), "entry " + a + " doesn't match");
a++;
}
Assert.IsNull(iter.Next());
Assert.AreEqual(a, stringList.Count);
}
}
/// <summary>
/// Perform the actual DM Soundex algorithm on the input string.
/// </summary>
/// <param name="source">A string to encode.</param>
/// <param name="branching">If branching shall be performed.</param>
/// <returns>A string array containing all DM Soundex codes corresponding to the string supplied depending on the selected branching mode.</returns>
/// <exception cref="ArgumentException">If a character is not mapped.</exception>
private string[] GetSoundex(string source, bool branching)
{
if (source == null)
{
return(null);
}
string input = Cleanup(source);
// LinkedHashSet preserves input order. In .NET we can use List for that purpose.
IList <Branch> currentBranches = new JCG.List <Branch>
{
new Branch()
};
char lastChar = '\0';
for (int index = 0; index < input.Length; index++)
{
char ch = input[index];
// ignore whitespace inside a name
if (char.IsWhiteSpace(ch))
{
continue;
}
string inputContext = input.Substring(index);
if (!RULES.TryGetValue(ch, out IList <Rule> rules) || rules == null)
{
continue;
}
// use an EMPTY_LIST to avoid false positive warnings wrt potential null pointer access
IList <Branch> nextBranches = branching ? new JCG.List <Branch>() : Collections.EmptyList <Branch>() as IList <Branch>;
foreach (Rule rule in rules)
{
if (rule.Matches(inputContext))
{
if (branching)
{
nextBranches.Clear();
}
string[] replacements = rule.GetReplacements(inputContext, lastChar == '\0');
bool branchingRequired = replacements.Length > 1 && branching;
foreach (Branch branch in currentBranches)
{
foreach (string nextReplacement in replacements)
{
// if we have multiple replacements, always create a new branch
Branch nextBranch = branchingRequired ? branch.CreateBranch() : branch;
// special rule: occurrences of mn or nm are treated differently
bool force = (lastChar == 'm' && ch == 'n') || (lastChar == 'n' && ch == 'm');
nextBranch.ProcessNextReplacement(nextReplacement, force);
if (branching)
{
if (!nextBranches.Contains(nextBranch))
{
nextBranches.Add(nextBranch);
}
}
else
{
break;
}
}
}
if (branching)
{
currentBranches.Clear();
currentBranches.AddRange(nextBranches);
}
index += rule.PatternLength - 1;
break;
}
}
lastChar = ch;
}
string[] result = new string[currentBranches.Count];
int idx = 0;
foreach (Branch branch in currentBranches)
{
branch.Finish();
result[idx++] = branch.ToString();
}
return(result);
}
/// <summary>
/// Returns <c>true</c> if the given string is accepted by the automaton.
/// <para/>
/// Complexity: linear in the length of the string.
/// <para/>
/// <b>Note:</b> for full performance, use the <see cref="RunAutomaton"/> class.
/// </summary>
public static bool Run(Automaton a, string s)
{
if (a.IsSingleton)
{
return(s.Equals(a.singleton, StringComparison.Ordinal));
}
if (a.deterministic)
{
State p = a.initial;
int cp; // LUCENENET: Removed unnecessary assignment
for (int i = 0; i < s.Length; i += Character.CharCount(cp))
{
State q = p.Step(cp = Character.CodePointAt(s, i));
if (q == null)
{
return(false);
}
p = q;
}
return(p.accept);
}
else
{
State[] states = a.GetNumberedStates();
LinkedList <State> pp = new LinkedList <State>();
LinkedList <State> pp_other = new LinkedList <State>();
OpenBitSet bb = new OpenBitSet(states.Length);
OpenBitSet bb_other = new OpenBitSet(states.Length);
pp.AddLast(a.initial);
JCG.List <State> dest = new JCG.List <State>();
bool accept = a.initial.accept;
int c; // LUCENENET: Removed unnecessary assignment
for (int i = 0; i < s.Length; i += Character.CharCount(c))
{
c = Character.CodePointAt(s, i);
accept = false;
pp_other.Clear();
bb_other.Clear(0, bb_other.Length - 1);
foreach (State p in pp)
{
dest.Clear();
p.Step(c, dest);
foreach (State q in dest)
{
if (q.accept)
{
accept = true;
}
if (!bb_other.Get(q.number))
{
bb_other.Set(q.number);
pp_other.AddLast(q);
}
}
}
LinkedList <State> tp = pp;
pp = pp_other;
pp_other = tp;
OpenBitSet tb = bb;
bb = bb_other;
bb_other = tb;
}
return(accept);
}
}
protected override IQueryNode PostProcessNode(IQueryNode node)
{
if (node is ITextableQueryNode &&
!(node is WildcardQueryNode) &&
!(node is FuzzyQueryNode) &&
!(node is RegexpQueryNode) &&
!(node.Parent is IRangeQueryNode))
{
FieldQueryNode fieldNode = ((FieldQueryNode)node);
string text = fieldNode.GetTextAsString();
string field = fieldNode.GetFieldAsString();
CachingTokenFilter buffer = null;
IPositionIncrementAttribute posIncrAtt = null;
int numTokens = 0;
int positionCount = 0;
bool severalTokensAtSamePosition = false;
TokenStream source = null;
try
{
source = this.analyzer.GetTokenStream(field, text);
source.Reset();
buffer = new CachingTokenFilter(source);
if (buffer.HasAttribute <IPositionIncrementAttribute>())
{
posIncrAtt = buffer.GetAttribute <IPositionIncrementAttribute>();
}
try
{
while (buffer.IncrementToken())
{
numTokens++;
int positionIncrement = (posIncrAtt != null) ? posIncrAtt
.PositionIncrement : 1;
if (positionIncrement != 0)
{
positionCount += positionIncrement;
}
else
{
severalTokensAtSamePosition = true;
}
}
}
catch (Exception e) when(e.IsIOException())
{
// ignore
}
}
catch (Exception e) when(e.IsIOException())
{
throw RuntimeException.Create(e);
}
finally
{
IOUtils.DisposeWhileHandlingException(source);
}
// rewind the buffer stream
buffer.Reset();
if (!buffer.HasAttribute <ICharTermAttribute>())
{
return(new NoTokenFoundQueryNode());
}
ICharTermAttribute termAtt = buffer.GetAttribute <ICharTermAttribute>();
if (numTokens == 0)
{
return(new NoTokenFoundQueryNode());
}
else if (numTokens == 1)
{
string term = null;
try
{
bool hasNext;
hasNext = buffer.IncrementToken();
if (Debugging.AssertsEnabled)
{
Debugging.Assert(hasNext == true);
}
term = termAtt.ToString();
}
catch (Exception e) when(e.IsIOException())
{
// safe to ignore, because we know the number of tokens
}
fieldNode.Text = term.AsCharSequence();
return(fieldNode);
}
else if (severalTokensAtSamePosition || !(node is QuotedFieldQueryNode))
{
if (positionCount == 1 || !(node is QuotedFieldQueryNode))
{
// no phrase query:
if (positionCount == 1)
{
// simple case: only one position, with synonyms
IList <IQueryNode> children = new JCG.List <IQueryNode>();
for (int i = 0; i < numTokens; i++)
{
string term = null;
try
{
bool hasNext = buffer.IncrementToken();
if (Debugging.AssertsEnabled)
{
Debugging.Assert(hasNext == true);
}
term = termAtt.ToString();
}
catch (Exception e) when(e.IsIOException())
{
// safe to ignore, because we know the number of tokens
}
children.Add(new FieldQueryNode(field, term, -1, -1));
}
return(new GroupQueryNode(
new StandardBooleanQueryNode(children, positionCount == 1)));
}
else
{
// multiple positions
IQueryNode q = new StandardBooleanQueryNode(Collections.EmptyList <IQueryNode>(), false);
IQueryNode currentQuery = null;
for (int i = 0; i < numTokens; i++)
{
string term = null;
try
{
bool hasNext = buffer.IncrementToken();
if (Debugging.AssertsEnabled)
{
Debugging.Assert(hasNext == true);
}
term = termAtt.ToString();
}
catch (Exception e) when(e.IsIOException())
{
// safe to ignore, because we know the number of tokens
}
if (posIncrAtt != null && posIncrAtt.PositionIncrement == 0)
{
if (!(currentQuery is BooleanQueryNode))
{
IQueryNode t = currentQuery;
currentQuery = new StandardBooleanQueryNode(Collections.EmptyList <IQueryNode>(), true);
((BooleanQueryNode)currentQuery).Add(t);
}
((BooleanQueryNode)currentQuery).Add(new FieldQueryNode(field, term, -1, -1));
}
else
{
if (currentQuery != null)
{
if (this.defaultOperator == Operator.OR)
{
q.Add(currentQuery);
}
else
{
q.Add(new ModifierQueryNode(currentQuery, Modifier.MOD_REQ));
}
}
currentQuery = new FieldQueryNode(field, term, -1, -1);
}
}
if (this.defaultOperator == Operator.OR)
{
q.Add(currentQuery);
}
else
{
q.Add(new ModifierQueryNode(currentQuery, Modifier.MOD_REQ));
}
if (q is BooleanQueryNode)
{
q = new GroupQueryNode(q);
}
return(q);
}
}
else
{
// phrase query:
MultiPhraseQueryNode mpq = new MultiPhraseQueryNode();
IList <FieldQueryNode> multiTerms = new JCG.List <FieldQueryNode>();
int position = -1;
int i = 0;
int termGroupCount = 0;
for (; i < numTokens; i++)
{
string term = null;
int positionIncrement = 1;
try
{
bool hasNext = buffer.IncrementToken();
if (Debugging.AssertsEnabled)
{
Debugging.Assert(hasNext == true);
}
term = termAtt.ToString();
if (posIncrAtt != null)
{
positionIncrement = posIncrAtt.PositionIncrement;
}
}
catch (Exception e) when(e.IsIOException())
{
// safe to ignore, because we know the number of tokens
}
if (positionIncrement > 0 && multiTerms.Count > 0)
{
foreach (FieldQueryNode termNode in multiTerms)
{
if (this.positionIncrementsEnabled)
{
termNode.PositionIncrement = position;
}
else
{
termNode.PositionIncrement = termGroupCount;
}
mpq.Add(termNode);
}
// Only increment once for each "group" of
// terms that were in the same position:
termGroupCount++;
multiTerms.Clear();
}
position += positionIncrement;
multiTerms.Add(new FieldQueryNode(field, term, -1, -1));
}
foreach (FieldQueryNode termNode in multiTerms)
{
if (this.positionIncrementsEnabled)
{
termNode.PositionIncrement = position;
}
else
{
termNode.PositionIncrement = termGroupCount;
}
mpq.Add(termNode);
}
return(mpq);
}
}
else
{
TokenizedPhraseQueryNode pq = new TokenizedPhraseQueryNode();
int position = -1;
for (int i = 0; i < numTokens; i++)
{
string term = null;
int positionIncrement = 1;
try
{
bool hasNext = buffer.IncrementToken();
if (Debugging.AssertsEnabled)
{
Debugging.Assert(hasNext == true);
}
term = termAtt.ToString();
if (posIncrAtt != null)
{
positionIncrement = posIncrAtt.PositionIncrement;
}
}
catch (Exception e) when(e.IsIOException())
{
// safe to ignore, because we know the number of tokens
}
FieldQueryNode newFieldNode = new FieldQueryNode(field, term, -1, -1);
if (this.positionIncrementsEnabled)
{
position += positionIncrement;
newFieldNode.PositionIncrement = position;
}
else
{
newFieldNode.PositionIncrement = i;
}
pq.Add(newFieldNode);
}
return(pq);
}
}
return(node);
}
/// <summary>
/// Sort input to output, explicit hint for the buffer size. The amount of allocated
/// memory may deviate from the hint (may be smaller or larger).
/// </summary>
public SortInfo Sort(FileInfo input, FileInfo output)
{
sortInfo = new SortInfo(this)
{
TotalTime = J2N.Time.NanoTime() / J2N.Time.MillisecondsPerNanosecond
}; // LUCENENET: Use NanoTime() rather than CurrentTimeMilliseconds() for more accurate/reliable results
output.Delete();
var merges = new JCG.List <FileInfo>();
bool success2 = false;
try
{
var inputStream = new ByteSequencesReader(input);
bool success = false;
try
{
int lines = 0;
while ((lines = ReadPartition(inputStream)) > 0)
{
merges.Add(SortPartition(/*lines*/)); // LUCENENET specific - removed unused parameter
sortInfo.TempMergeFiles++;
sortInfo.Lines += lines;
// Handle intermediate merges.
if (merges.Count == maxTempFiles)
{
var intermediate = new FileInfo(Path.GetTempFileName());
try
{
MergePartitions(merges, intermediate);
}
finally
{
foreach (var file in merges)
{
file.Delete();
}
merges.Clear();
merges.Add(intermediate);
}
sortInfo.TempMergeFiles++;
}
}
success = true;
}
finally
{
if (success)
{
IOUtils.Dispose(inputStream);
}
else
{
IOUtils.DisposeWhileHandlingException(inputStream);
}
}
// One partition, try to rename or copy if unsuccessful.
if (merges.Count == 1)
{
FileInfo single = merges[0];
Copy(single, output);
try
{
File.Delete(single.FullName);
}
#pragma warning disable CA1031 // Do not catch general exception types
catch
{
// ignored
}
#pragma warning restore CA1031 // Do not catch general exception types
}
else
{
// otherwise merge the partitions with a priority queue.
MergePartitions(merges, output);
}
success2 = true;
}
finally
{
foreach (FileInfo file in merges)
{
file.Delete();
}
if (!success2)
{
output.Delete();
}
}
sortInfo.TotalTime = ((J2N.Time.NanoTime() / J2N.Time.MillisecondsPerNanosecond) - sortInfo.TotalTime); // LUCENENET: Use NanoTime() rather than CurrentTimeMilliseconds() for more accurate/reliable results
return(sortInfo);
}
/// <summary>
/// a constructor.
/// </summary>
/// <param name="fieldTermStack"><see cref="FieldTermStack"/> object</param>
/// <param name="fieldQuery"><see cref="FieldQuery"/> object</param>
/// <param name="phraseLimit">maximum size of phraseList</param>
public FieldPhraseList(FieldTermStack fieldTermStack, FieldQuery fieldQuery, int phraseLimit)
{
string field = fieldTermStack.FieldName;
IList <TermInfo> phraseCandidate = new JCG.List <TermInfo>();
QueryPhraseMap currMap; // LUCENENET: IDE0059: Remove unnecessary value assignment
QueryPhraseMap nextMap; // LUCENENET: IDE0059: Remove unnecessary value assignment
while (!fieldTermStack.IsEmpty && (phraseList.Count < phraseLimit))
{
phraseCandidate.Clear();
TermInfo ti; // LUCENENET: IDE0059: Remove unnecessary value assignment
TermInfo first; // LUCENENET: IDE0059: Remove unnecessary value assignment
first = ti = fieldTermStack.Pop();
currMap = fieldQuery.GetFieldTermMap(field, ti.Text);
while (currMap is null && ti.Next != first)
{
ti = ti.Next;
currMap = fieldQuery.GetFieldTermMap(field, ti.Text);
}
// if not found, discard top TermInfo from stack, then try next element
if (currMap is null)
{
continue;
}
// if found, search the longest phrase
phraseCandidate.Add(ti);
while (true)
{
first = ti = fieldTermStack.Pop();
nextMap = null;
if (ti != null)
{
nextMap = currMap.GetTermMap(ti.Text);
while (nextMap is null && ti.Next != first)
{
ti = ti.Next;
nextMap = currMap.GetTermMap(ti.Text);
}
}
if (ti is null || nextMap is null)
{
if (ti != null)
{
fieldTermStack.Push(ti);
}
if (currMap.IsValidTermOrPhrase(phraseCandidate))
{
AddIfNoOverlap(new WeightedPhraseInfo(phraseCandidate, currMap.Boost, currMap.TermOrPhraseNumber));
}
else
{
while (phraseCandidate.Count > 1)
{
//fieldTermStack.Push(phraseCandidate.Last.Value);
//phraseCandidate.RemoveLast();
TermInfo last = phraseCandidate[phraseCandidate.Count - 1];
phraseCandidate.Remove(last);
fieldTermStack.Push(last);
currMap = fieldQuery.SearchPhrase(field, phraseCandidate);
if (currMap != null)
{
AddIfNoOverlap(new WeightedPhraseInfo(phraseCandidate, currMap.Boost, currMap.TermOrPhraseNumber));
break;
}
}
}
break;
}
protected virtual FieldFragList CreateFieldFragList(FieldPhraseList fieldPhraseList, FieldFragList fieldFragList, int fragCharSize)
{
// LUCENENET specific - added guard clauses to check for nulls
if (fieldPhraseList is null)
{
throw new ArgumentNullException(nameof(fieldPhraseList));
}
if (fieldFragList is null)
{
throw new ArgumentNullException(nameof(fieldFragList));
}
if (fragCharSize < minFragCharSize)
{
throw new ArgumentOutOfRangeException(nameof(fragCharSize), "fragCharSize(" + fragCharSize + ") is too small. It must be " + minFragCharSize + " or higher."); // LUCENENET specific - changed from IllegalArgumentException to ArgumentOutOfRangeException (.NET convention)
}
JCG.List <WeightedPhraseInfo> wpil = new JCG.List <WeightedPhraseInfo>();
using (IteratorQueue <WeightedPhraseInfo> queue = new IteratorQueue <WeightedPhraseInfo>(fieldPhraseList.PhraseList.GetEnumerator()))
{
WeightedPhraseInfo phraseInfo = null;
int startOffset = 0;
while ((phraseInfo = queue.Top()) != null)
{
// if the phrase violates the border of previous fragment, discard it and try next phrase
if (phraseInfo.StartOffset < startOffset)
{
queue.RemoveTop();
continue;
}
wpil.Clear();
int currentPhraseStartOffset = phraseInfo.StartOffset;
int currentPhraseEndOffset = phraseInfo.EndOffset;
int spanStart = Math.Max(currentPhraseStartOffset - margin, startOffset);
int spanEnd = Math.Max(currentPhraseEndOffset, spanStart + fragCharSize);
if (AcceptPhrase(queue.RemoveTop(), currentPhraseEndOffset - currentPhraseStartOffset, fragCharSize))
{
wpil.Add(phraseInfo);
}
while ((phraseInfo = queue.Top()) != null)
{ // pull until we crossed the current spanEnd
if (phraseInfo.EndOffset <= spanEnd)
{
currentPhraseEndOffset = phraseInfo.EndOffset;
if (AcceptPhrase(queue.RemoveTop(), currentPhraseEndOffset - currentPhraseStartOffset, fragCharSize))
{
wpil.Add(phraseInfo);
}
}
else
{
break;
}
}
if (wpil.Count == 0)
{
continue;
}
int matchLen = currentPhraseEndOffset - currentPhraseStartOffset;
// now recalculate the start and end position to "center" the result
int newMargin = Math.Max(0, (fragCharSize - matchLen) / 2); // matchLen can be > fragCharSize prevent IAOOB here
spanStart = currentPhraseStartOffset - newMargin;
if (spanStart < startOffset)
{
spanStart = startOffset;
}
// whatever is bigger here we grow this out
spanEnd = spanStart + Math.Max(matchLen, fragCharSize);
startOffset = spanEnd;
fieldFragList.Add(spanStart, spanEnd, wpil);
}
}
return(fieldFragList);
}
/// <summary>
/// Merging constructor.
/// </summary>
/// <param name="toMerge"><see cref="FieldPhraseList"/>s to merge to build this one</param>
public FieldPhraseList(FieldPhraseList[] toMerge)
{
// Merge all overlapping WeightedPhraseInfos
// Step 1. Sort by startOffset, endOffset, and boost, in that order.
IEnumerator <WeightedPhraseInfo>[] allInfos = new IEnumerator <WeightedPhraseInfo> [toMerge.Length];
try
{
int index = 0;
foreach (FieldPhraseList fplToMerge in toMerge)
{
allInfos[index++] = fplToMerge.phraseList.GetEnumerator();
}
using MergedEnumerator <WeightedPhraseInfo> itr = new MergedEnumerator <WeightedPhraseInfo>(false, allInfos);
// Step 2. Walk the sorted list merging infos that overlap
phraseList = new JCG.List <WeightedPhraseInfo>();
if (!itr.MoveNext())
{
return;
}
IList <WeightedPhraseInfo> work = new JCG.List <WeightedPhraseInfo>();
WeightedPhraseInfo first = itr.Current;
work.Add(first);
int workEndOffset = first.EndOffset;
while (itr.MoveNext())
{
WeightedPhraseInfo current = itr.Current;
if (current.StartOffset <= workEndOffset)
{
workEndOffset = Math.Max(workEndOffset, current.EndOffset);
work.Add(current);
}
else
{
if (work.Count == 1)
{
phraseList.Add(work[0]);
work[0] = current;
}
else
{
phraseList.Add(new WeightedPhraseInfo(work));
work.Clear();
work.Add(current);
}
workEndOffset = current.EndOffset;
}
}
if (work.Count == 1)
{
phraseList.Add(work[0]);
}
else
{
phraseList.Add(new WeightedPhraseInfo(work));
work.Clear();
}
}
finally
{
IOUtils.Dispose(allInfos);
}
}
public virtual void TestReadMappingRules()
{
SlowSynonymMap synMap;
// (a)->[b]
IList <string> rules = new JCG.List <string>();
rules.Add("a=>b");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", true, null);
assertEquals(1, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "b");
// (a)->[c]
// (b)->[c]
rules.Clear();
rules.Add("a,b=>c");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", true, null);
assertEquals(2, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "c");
AssertTokIncludes(synMap, "b", "c");
// (a)->[b][c]
rules.Clear();
rules.Add("a=>b,c");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", true, null);
assertEquals(1, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "b");
AssertTokIncludes(synMap, "a", "c");
// (a)->(b)->[a2]
// [a1]
rules.Clear();
rules.Add("a=>a1");
rules.Add("a b=>a2");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", true, null);
assertEquals(1, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "a1");
assertEquals(1, GetSubSynonymMap(synMap, "a").Submap.size());
AssertTokIncludes(GetSubSynonymMap(synMap, "a"), "b", "a2");
// (a)->(b)->[a2]
// (c)->[a3]
// [a1]
rules.Clear();
rules.Add("a=>a1");
rules.Add("a b=>a2");
rules.Add("a c=>a3");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", true, null);
assertEquals(1, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "a1");
assertEquals(2, GetSubSynonymMap(synMap, "a").Submap.size());
AssertTokIncludes(GetSubSynonymMap(synMap, "a"), "b", "a2");
AssertTokIncludes(GetSubSynonymMap(synMap, "a"), "c", "a3");
// (a)->(b)->[a2]
// [a1]
// (b)->(c)->[b2]
// [b1]
rules.Clear();
rules.Add("a=>a1");
rules.Add("a b=>a2");
rules.Add("b=>b1");
rules.Add("b c=>b2");
synMap = new SlowSynonymMap(true);
SlowSynonymFilterFactory.ParseRules(rules, synMap, "=>", ",", true, null);
assertEquals(2, synMap.Submap.size());
AssertTokIncludes(synMap, "a", "a1");
assertEquals(1, GetSubSynonymMap(synMap, "a").Submap.size());
AssertTokIncludes(GetSubSynonymMap(synMap, "a"), "b", "a2");
AssertTokIncludes(synMap, "b", "b1");
assertEquals(1, GetSubSynonymMap(synMap, "b").Submap.size());
AssertTokIncludes(GetSubSynonymMap(synMap, "b"), "c", "b2");
}
public virtual void TestBasic()
{
string groupField = "author";
FieldType customType = new FieldType();
customType.IsStored = (true);
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(
Random,
dir,
NewIndexWriterConfig(TEST_VERSION_CURRENT,
new MockAnalyzer(Random)).SetMergePolicy(NewLogMergePolicy()));
bool canUseIDV = !"Lucene3x".Equals(w.IndexWriter.Config.Codec.Name, StringComparison.Ordinal);
JCG.List <Document> documents = new JCG.List <Document>();
// 0
Document doc = new Document();
AddGroupField(doc, groupField, "author1", canUseIDV);
doc.Add(new TextField("content", "random text", Field.Store.YES));
doc.Add(new Field("id", "1", customType));
documents.Add(doc);
// 1
doc = new Document();
AddGroupField(doc, groupField, "author1", canUseIDV);
doc.Add(new TextField("content", "some more random text", Field.Store.YES));
doc.Add(new Field("id", "2", customType));
documents.Add(doc);
// 2
doc = new Document();
AddGroupField(doc, groupField, "author1", canUseIDV);
doc.Add(new TextField("content", "some more random textual data", Field.Store.YES));
doc.Add(new Field("id", "3", customType));
doc.Add(new StringField("groupend", "x", Field.Store.NO));
documents.Add(doc);
w.AddDocuments(documents);
documents.Clear();
// 3
doc = new Document();
AddGroupField(doc, groupField, "author2", canUseIDV);
doc.Add(new TextField("content", "some random text", Field.Store.YES));
doc.Add(new Field("id", "4", customType));
doc.Add(new StringField("groupend", "x", Field.Store.NO));
w.AddDocument(doc);
// 4
doc = new Document();
AddGroupField(doc, groupField, "author3", canUseIDV);
doc.Add(new TextField("content", "some more random text", Field.Store.YES));
doc.Add(new Field("id", "5", customType));
documents.Add(doc);
// 5
doc = new Document();
AddGroupField(doc, groupField, "author3", canUseIDV);
doc.Add(new TextField("content", "random", Field.Store.YES));
doc.Add(new Field("id", "6", customType));
doc.Add(new StringField("groupend", "x", Field.Store.NO));
documents.Add(doc);
w.AddDocuments(documents);
documents.Clear();
// 6 -- no author field
doc = new Document();
doc.Add(new TextField("content", "random word stuck in alot of other text", Field.Store.YES));
doc.Add(new Field("id", "6", customType));
doc.Add(new StringField("groupend", "x", Field.Store.NO));
w.AddDocument(doc);
IndexSearcher indexSearcher = NewSearcher(w.GetReader());
w.Dispose();
Sort groupSort = Sort.RELEVANCE;
GroupingSearch groupingSearch = CreateRandomGroupingSearch(groupField, groupSort, 5, canUseIDV);
ITopGroups <object> groups = groupingSearch.Search(indexSearcher, (Filter)null, new TermQuery(new Index.Term("content", "random")), 0, 10);
assertEquals(7, groups.TotalHitCount);
assertEquals(7, groups.TotalGroupedHitCount);
assertEquals(4, groups.Groups.Length);
// relevance order: 5, 0, 3, 4, 1, 2, 6
// the later a document is added the higher this docId
// value
IGroupDocs <object> group = groups.Groups[0];
CompareGroupValue("author3", group);
assertEquals(2, group.ScoreDocs.Length);
assertEquals(5, group.ScoreDocs[0].Doc);
assertEquals(4, group.ScoreDocs[1].Doc);
assertTrue(group.ScoreDocs[0].Score > group.ScoreDocs[1].Score);
group = groups.Groups[1];
CompareGroupValue("author1", group);
assertEquals(3, group.ScoreDocs.Length);
assertEquals(0, group.ScoreDocs[0].Doc);
assertEquals(1, group.ScoreDocs[1].Doc);
assertEquals(2, group.ScoreDocs[2].Doc);
assertTrue(group.ScoreDocs[0].Score > group.ScoreDocs[1].Score);
assertTrue(group.ScoreDocs[1].Score > group.ScoreDocs[2].Score);
group = groups.Groups[2];
CompareGroupValue("author2", group);
assertEquals(1, group.ScoreDocs.Length);
assertEquals(3, group.ScoreDocs[0].Doc);
group = groups.Groups[3];
CompareGroupValue(null, group);
assertEquals(1, group.ScoreDocs.Length);
assertEquals(6, group.ScoreDocs[0].Doc);
Filter lastDocInBlock = new CachingWrapperFilter(new QueryWrapperFilter(new TermQuery(new Index.Term("groupend", "x"))));
groupingSearch = new GroupingSearch(lastDocInBlock);
groups = groupingSearch.Search(indexSearcher, null, new TermQuery(new Index.Term("content", "random")), 0, 10);
assertEquals(7, groups.TotalHitCount);
assertEquals(7, groups.TotalGroupedHitCount);
assertEquals(4, groups.TotalGroupCount.GetValueOrDefault());
assertEquals(4, groups.Groups.Length);
indexSearcher.IndexReader.Dispose();
dir.Dispose();
}
/// <summary>
/// Dumps an <see cref="FST{T}"/> to a GraphViz's <c>dot</c> language description
/// for visualization. Example of use:
///
/// <code>
/// using (TextWriter sw = new StreamWriter("out.dot"))
/// {
/// Util.ToDot(fst, sw, true, true);
/// }
/// </code>
///
/// and then, from command line:
///
/// <code>
/// dot -Tpng -o out.png out.dot
/// </code>
///
/// <para/>
/// Note: larger FSTs (a few thousand nodes) won't even
/// render, don't bother. If the FST is > 2.1 GB in size
/// then this method will throw strange exceptions.
/// <para/>
/// See also <a href="http://www.graphviz.org/">http://www.graphviz.org/</a>.
/// </summary>
/// <param name="sameRank">
/// If <c>true</c>, the resulting <c>dot</c> file will try
/// to order states in layers of breadth-first traversal. This may
/// mess up arcs, but makes the output FST's structure a bit clearer.
/// </param>
/// <param name="labelStates">
/// If <c>true</c> states will have labels equal to their offsets in their
/// binary format. Expands the graph considerably.
/// </param>
public static void ToDot <T>(FST <T> fst, TextWriter @out, bool sameRank, bool labelStates)
{
const string expandedNodeColor = "blue";
// this is the start arc in the automaton (from the epsilon state to the first state
// with outgoing transitions.
FST.Arc <T> startArc = fst.GetFirstArc(new FST.Arc <T>());
// A queue of transitions to consider for the next level.
IList <FST.Arc <T> > thisLevelQueue = new JCG.List <FST.Arc <T> >();
// A queue of transitions to consider when processing the next level.
IList <FST.Arc <T> > nextLevelQueue = new JCG.List <FST.Arc <T> >();
nextLevelQueue.Add(startArc);
//System.out.println("toDot: startArc: " + startArc);
// A list of states on the same level (for ranking).
IList <int?> sameLevelStates = new JCG.List <int?>();
// A bitset of already seen states (target offset).
BitArray seen = new BitArray(32);
seen.SafeSet((int)startArc.Target, true);
// Shape for states.
const string stateShape = "circle";
const string finalStateShape = "doublecircle";
// Emit DOT prologue.
@out.Write("digraph FST {\n");
@out.Write(" rankdir = LR; splines=true; concentrate=true; ordering=out; ranksep=2.5; \n");
if (!labelStates)
{
@out.Write(" node [shape=circle, width=.2, height=.2, style=filled]\n");
}
EmitDotState(@out, "initial", "point", "white", "");
T NO_OUTPUT = fst.Outputs.NoOutput;
var r = fst.GetBytesReader();
// final FST.Arc<T> scratchArc = new FST.Arc<>();
{
string stateColor;
if (fst.IsExpandedTarget(startArc, r))
{
stateColor = expandedNodeColor;
}
else
{
stateColor = null;
}
bool isFinal;
T finalOutput;
if (startArc.IsFinal)
{
isFinal = true;
finalOutput = startArc.NextFinalOutput.Equals(NO_OUTPUT) ? default(T) : startArc.NextFinalOutput;
}
else
{
isFinal = false;
finalOutput = default(T);
}
EmitDotState(@out, Convert.ToString(startArc.Target), isFinal ? finalStateShape : stateShape, stateColor, finalOutput == null ? "" : fst.Outputs.OutputToString(finalOutput));
}
@out.Write(" initial -> " + startArc.Target + "\n");
int level = 0;
while (nextLevelQueue.Count > 0)
{
// we could double buffer here, but it doesn't matter probably.
//System.out.println("next level=" + level);
thisLevelQueue.AddRange(nextLevelQueue);
nextLevelQueue.Clear();
level++;
@out.Write("\n // Transitions and states at level: " + level + "\n");
while (thisLevelQueue.Count > 0)
{
FST.Arc <T> arc = thisLevelQueue[thisLevelQueue.Count - 1];
thisLevelQueue.RemoveAt(thisLevelQueue.Count - 1);
//System.out.println(" pop: " + arc);
if (FST <T> .TargetHasArcs(arc))
{
// scan all target arcs
//System.out.println(" readFirstTarget...");
long node = arc.Target;
fst.ReadFirstRealTargetArc(arc.Target, arc, r);
//System.out.println(" firstTarget: " + arc);
while (true)
{
//System.out.println(" cycle arc=" + arc);
// Emit the unseen state and add it to the queue for the next level.
if (arc.Target >= 0 && !seen.SafeGet((int)arc.Target))
{
/*
* boolean isFinal = false;
* T finalOutput = null;
* fst.readFirstTargetArc(arc, scratchArc);
* if (scratchArc.isFinal() && fst.targetHasArcs(scratchArc)) {
* // target is final
* isFinal = true;
* finalOutput = scratchArc.output == NO_OUTPUT ? null : scratchArc.output;
* System.out.println("dot hit final label=" + (char) scratchArc.label);
* }
*/
string stateColor;
if (fst.IsExpandedTarget(arc, r))
{
stateColor = expandedNodeColor;
}
else
{
stateColor = null;
}
string finalOutput;
if (arc.NextFinalOutput != null && !arc.NextFinalOutput.Equals(NO_OUTPUT))
{
finalOutput = fst.Outputs.OutputToString(arc.NextFinalOutput);
}
else
{
finalOutput = "";
}
EmitDotState(@out, Convert.ToString(arc.Target), stateShape, stateColor, finalOutput);
// To see the node address, use this instead:
//emitDotState(out, Integer.toString(arc.target), stateShape, stateColor, String.valueOf(arc.target));
seen.SafeSet((int)arc.Target, true);
nextLevelQueue.Add((new FST.Arc <T>()).CopyFrom(arc));
sameLevelStates.Add((int)arc.Target);
}
string outs;
if (!arc.Output.Equals(NO_OUTPUT))
{
outs = "/" + fst.Outputs.OutputToString(arc.Output);
}
else
{
outs = "";
}
if (!FST <T> .TargetHasArcs(arc) && arc.IsFinal && !arc.NextFinalOutput.Equals(NO_OUTPUT))
{
// Tricky special case: sometimes, due to
// pruning, the builder can [sillily] produce
// an FST with an arc into the final end state
// (-1) but also with a next final output; in
// this case we pull that output up onto this
// arc
outs = outs + "/[" + fst.Outputs.OutputToString(arc.NextFinalOutput) + "]";
}
string arcColor;
if (arc.Flag(FST.BIT_TARGET_NEXT))
{
arcColor = "red";
}
else
{
arcColor = "black";
}
Debug.Assert(arc.Label != FST.END_LABEL);
@out.Write(" " + node + " -> " + arc.Target + " [label=\"" + PrintableLabel(arc.Label) + outs + "\"" + (arc.IsFinal ? " style=\"bold\"" : "") + " color=\"" + arcColor + "\"]\n");
// Break the loop if we're on the last arc of this state.
if (arc.IsLast)
{
//System.out.println(" break");
break;
}
fst.ReadNextRealArc(arc, r);
}
}
}
// Emit state ranking information.
if (sameRank && sameLevelStates.Count > 1)
{
@out.Write(" {rank=same; ");
foreach (int state in sameLevelStates)
{
@out.Write(state + "; ");
}
@out.Write(" }\n");
}
sameLevelStates.Clear();
}
// Emit terminating state (always there anyway).
@out.Write(" -1 [style=filled, color=black, shape=doublecircle, label=\"\"]\n\n");
@out.Write(" {rank=sink; -1 }\n");
@out.Write("}\n");
@out.Flush();
}
public override void Run()
{
// TODO: would be better if this were cross thread, so that we make sure one thread deleting anothers added docs works:
IList <string> toDeleteIDs = new JCG.List <string>();
IList <SubDocs> toDeleteSubDocs = new JCG.List <SubDocs>();
while (J2N.Time.NanoTime() / J2N.Time.MillisecondsPerNanosecond < stopTime && !outerInstance.m_failed) // LUCENENET: Use NanoTime() rather than CurrentTimeMilliseconds() for more accurate/reliable results
{
try
{
// Occasional longish pause if running
// nightly
if (LuceneTestCase.TestNightly && Random.Next(6) == 3)
{
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": now long sleep");
}
//Thread.Sleep(TestUtil.NextInt32(Random, 50, 500));
// LUCENENET specific - Reduced amount of pause to keep the total
// Nightly test time under 1 hour
Thread.Sleep(TestUtil.NextInt32(Random, 50, 250));
}
// Rate limit ingest rate:
if (Random.Next(7) == 5)
{
Thread.Sleep(TestUtil.NextInt32(Random, 1, 10));
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": done sleep");
}
}
Document doc = docs.NextDoc();
if (doc is null)
{
break;
}
// Maybe add randomly named field
string addedField;
if (Random.NextBoolean())
{
addedField = "extra" + Random.Next(40);
doc.Add(NewTextField(addedField, "a random field", Field.Store.YES));
}
else
{
addedField = null;
}
if (Random.NextBoolean())
{
if (Random.NextBoolean())
{
// Add/update doc block:
string packID;
SubDocs delSubDocs;
if (toDeleteSubDocs.Count > 0 && Random.NextBoolean())
{
delSubDocs = toDeleteSubDocs[Random.Next(toDeleteSubDocs.Count)];
if (Debugging.AssertsEnabled)
{
Debugging.Assert(!delSubDocs.Deleted);
}
toDeleteSubDocs.Remove(delSubDocs);
// Update doc block, replacing prior packID
packID = delSubDocs.PackID;
}
else
{
delSubDocs = null;
// Add doc block, using new packID
packID = outerInstance.m_packCount.GetAndIncrement().ToString(CultureInfo.InvariantCulture);
}
Field packIDField = NewStringField("packID", packID, Field.Store.YES);
IList <string> docIDs = new JCG.List <string>();
SubDocs subDocs = new SubDocs(packID, docIDs);
IList <Document> docsList = new JCG.List <Document>();
allSubDocs.Enqueue(subDocs);
doc.Add(packIDField);
docsList.Add(TestUtil.CloneDocument(doc));
docIDs.Add(doc.Get("docid"));
int maxDocCount = TestUtil.NextInt32(Random, 1, 10);
while (docsList.Count < maxDocCount)
{
doc = docs.NextDoc();
if (doc is null)
{
break;
}
docsList.Add(TestUtil.CloneDocument(doc));
docIDs.Add(doc.Get("docid"));
}
outerInstance.m_addCount.AddAndGet(docsList.Count);
Term packIDTerm = new Term("packID", packID);
if (delSubDocs != null)
{
delSubDocs.Deleted = true;
delIDs.UnionWith(delSubDocs.SubIDs);
outerInstance.m_delCount.AddAndGet(delSubDocs.SubIDs.Count);
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": update pack packID=" + delSubDocs.PackID +
" count=" + docsList.Count + " docs=" + string.Format(J2N.Text.StringFormatter.InvariantCulture, "{0}", docIDs));
}
outerInstance.UpdateDocuments(packIDTerm, docsList);
}
else
{
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": add pack packID=" + packID +
" count=" + docsList.Count + " docs=" + string.Format(J2N.Text.StringFormatter.InvariantCulture, "{0}", docIDs));
}
outerInstance.AddDocuments(packIDTerm, docsList);
}
doc.RemoveField("packID");
if (Random.Next(5) == 2)
{
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": buffer del id:" + packID);
}
toDeleteSubDocs.Add(subDocs);
}
}
else
{
// Add single doc
string docid = doc.Get("docid");
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": add doc docid:" + docid);
}
outerInstance.AddDocument(new Term("docid", docid), doc);
outerInstance.m_addCount.GetAndIncrement();
if (Random.Next(5) == 3)
{
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": buffer del id:" + doc.Get("docid"));
}
toDeleteIDs.Add(docid);
}
}
}
else
{
// Update single doc, but we never re-use
// and ID so the delete will never
// actually happen:
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": update doc id:" + doc.Get("docid"));
}
string docid = doc.Get("docid");
outerInstance.UpdateDocument(new Term("docid", docid), doc);
outerInstance.m_addCount.GetAndIncrement();
if (Random.Next(5) == 3)
{
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": buffer del id:" + doc.Get("docid"));
}
toDeleteIDs.Add(docid);
}
}
if (Random.Next(30) == 17)
{
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": apply " + toDeleteIDs.Count + " deletes");
}
foreach (string id in toDeleteIDs)
{
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": del term=id:" + id);
}
outerInstance.DeleteDocuments(new Term("docid", id));
}
int count = outerInstance.m_delCount.AddAndGet(toDeleteIDs.Count);
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": tot " + count + " deletes");
}
delIDs.UnionWith(toDeleteIDs);
toDeleteIDs.Clear();
foreach (SubDocs subDocs in toDeleteSubDocs)
{
if (Debugging.AssertsEnabled)
{
Debugging.Assert(!subDocs.Deleted);
}
delPackIDs.Add(subDocs.PackID);
outerInstance.DeleteDocuments(new Term("packID", subDocs.PackID));
subDocs.Deleted = true;
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": del subs: " + subDocs.SubIDs + " packID=" + subDocs.PackID);
}
delIDs.UnionWith(subDocs.SubIDs);
outerInstance.m_delCount.AddAndGet(subDocs.SubIDs.Count);
}
toDeleteSubDocs.Clear();
}
if (addedField != null)
{
doc.RemoveField(addedField);
}
}
catch (Exception t) when(t.IsThrowable())
{
Console.WriteLine(Thread.CurrentThread.Name + ": hit exc");
Console.WriteLine(t.ToString());
Console.Write(t.StackTrace);
outerInstance.m_failed.Value = (true);
throw RuntimeException.Create(t);
}
}
if (Verbose)
{
Console.WriteLine(Thread.CurrentThread.Name + ": indexing done");
}
outerInstance.DoAfterIndexingThreadDone();
}
/// <summary>
/// Minimizes the given automaton using Hopcroft's algorithm.
/// </summary>
public static void MinimizeHopcroft(Automaton a)
{
a.Determinize();
if (a.initial.numTransitions == 1)
{
Transition t = a.initial.TransitionsArray[0];
if (t.to == a.initial && t.min == Character.MinCodePoint && t.max == Character.MaxCodePoint)
{
return;
}
}
a.Totalize();
// initialize data structures
int[] sigma = a.GetStartPoints();
State[] states = a.GetNumberedStates();
int sigmaLen = sigma.Length, statesLen = states.Length;
JCG.List <State>[,] reverse = new JCG.List <State> [statesLen, sigmaLen];
ISet <State>[] partition = new JCG.HashSet <State> [statesLen];
JCG.List <State>[] splitblock = new JCG.List <State> [statesLen];
int[] block = new int[statesLen];
StateList[,] active = new StateList[statesLen, sigmaLen];
StateListNode[,] active2 = new StateListNode[statesLen, sigmaLen];
Queue <Int32Pair> pending = new Queue <Int32Pair>(); // LUCENENET specific - Queue is much more performant than LinkedList
OpenBitSet pending2 = new OpenBitSet(sigmaLen * statesLen);
OpenBitSet split = new OpenBitSet(statesLen),
refine = new OpenBitSet(statesLen), refine2 = new OpenBitSet(statesLen);
for (int q = 0; q < statesLen; q++)
{
splitblock[q] = new JCG.List <State>();
partition[q] = new JCG.HashSet <State>();
for (int x = 0; x < sigmaLen; x++)
{
active[q, x] = new StateList();
}
}
// find initial partition and reverse edges
for (int q = 0; q < statesLen; q++)
{
State qq = states[q];
int j = qq.accept ? 0 : 1;
partition[j].Add(qq);
block[q] = j;
for (int x = 0; x < sigmaLen; x++)
{
//JCG.List<State>[] r = reverse[qq.Step(sigma[x]).number];
var r = qq.Step(sigma[x]).number;
if (reverse[r, x] == null)
{
reverse[r, x] = new JCG.List <State>();
}
reverse[r, x].Add(qq);
}
}
// initialize active sets
for (int j = 0; j <= 1; j++)
{
for (int x = 0; x < sigmaLen; x++)
{
foreach (State qq in partition[j])
{
if (reverse[qq.number, x] != null)
{
active2[qq.number, x] = active[j, x].Add(qq);
}
}
}
}
// initialize pending
for (int x = 0; x < sigmaLen; x++)
{
int j = (active[0, x].Count <= active[1, x].Count) ? 0 : 1;
pending.Enqueue(new Int32Pair(j, x));
pending2.Set(x * statesLen + j);
}
// process pending until fixed point
int k = 2;
while (pending.Count > 0)
{
Int32Pair ip = pending.Dequeue();
int p = ip.n1;
int x = ip.n2;
pending2.Clear(x * statesLen + p);
// find states that need to be split off their blocks
for (StateListNode m = active[p, x].First; m != null; m = m.Next)
{
JCG.List <State> r = reverse[m.Q.number, x];
if (r != null)
{
foreach (State s in r)
{
int i = s.number;
if (!split.Get(i))
{
split.Set(i);
int j = block[i];
splitblock[j].Add(s);
if (!refine2.Get(j))
{
refine2.Set(j);
refine.Set(j);
}
}
}
}
}
// refine blocks
for (int j = refine.NextSetBit(0); j >= 0; j = refine.NextSetBit(j + 1))
{
JCG.List <State> sb = splitblock[j];
if (sb.Count < partition[j].Count)
{
ISet <State> b1 = partition[j];
ISet <State> b2 = partition[k];
foreach (State s in sb)
{
b1.Remove(s);
b2.Add(s);
block[s.number] = k;
for (int c = 0; c < sigmaLen; c++)
{
StateListNode sn = active2[s.number, c];
if (sn != null && sn.Sl == active[j, c])
{
sn.Remove();
active2[s.number, c] = active[k, c].Add(s);
}
}
}
// update pending
for (int c = 0; c < sigmaLen; c++)
{
int aj = active[j, c].Count, ak = active[k, c].Count, ofs = c * statesLen;
if (!pending2.Get(ofs + j) && 0 < aj && aj <= ak)
{
pending2.Set(ofs + j);
pending.Enqueue(new Int32Pair(j, c));
}
else
{
pending2.Set(ofs + k);
pending.Enqueue(new Int32Pair(k, c));
}
}
k++;
}
refine2.Clear(j);
foreach (State s in sb)
{
split.Clear(s.number);
}
sb.Clear();
}
refine.Clear(0, refine.Length);
}
// make a new state for each equivalence class, set initial state
State[] newstates = new State[k];
for (int n = 0; n < newstates.Length; n++)
{
State s = new State();
newstates[n] = s;
foreach (State q in partition[n])
{
if (q == a.initial)
{
a.initial = s;
}
s.accept = q.accept;
s.number = q.number; // select representative
q.number = n;
}
}
// build transitions and set acceptance
for (int n = 0; n < newstates.Length; n++)
{
State s = newstates[n];
s.accept = states[s.number].accept;
foreach (Transition t in states[s.number].GetTransitions())
{
s.AddTransition(new Transition(t.min, t.max, newstates[t.to.number]));
}
}
a.ClearNumberedStates();
a.RemoveDeadTransitions();
}
public void TestNestedSorting()
{
Directory dir = NewDirectory();
RandomIndexWriter w = new RandomIndexWriter(Random, dir, NewIndexWriterConfig(TEST_VERSION_CURRENT,
new MockAnalyzer(Random)).SetMergePolicy(NoMergePolicy.COMPOUND_FILES));
IList <Document> docs = new JCG.List <Document>();
Document document = new Document();
document.Add(new StringField("field2", "a", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "b", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "c", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("__type", "parent", Field.Store.NO));
document.Add(new StringField("field1", "a", Field.Store.NO));
docs.Add(document);
w.AddDocuments(docs);
w.Commit();
docs.Clear();
document = new Document();
document.Add(new StringField("field2", "c", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "d", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "e", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("__type", "parent", Field.Store.NO));
document.Add(new StringField("field1", "b", Field.Store.NO));
docs.Add(document);
w.AddDocuments(docs);
docs.Clear();
document = new Document();
document.Add(new StringField("field2", "e", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "f", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "g", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("__type", "parent", Field.Store.NO));
document.Add(new StringField("field1", "c", Field.Store.NO));
docs.Add(document);
w.AddDocuments(docs);
docs.Clear();
document = new Document();
document.Add(new StringField("field2", "g", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "h", Field.Store.NO));
document.Add(new StringField("filter_1", "F", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "i", Field.Store.NO));
document.Add(new StringField("filter_1", "F", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("__type", "parent", Field.Store.NO));
document.Add(new StringField("field1", "d", Field.Store.NO));
docs.Add(document);
w.AddDocuments(docs);
w.Commit();
docs.Clear();
document = new Document();
document.Add(new StringField("field2", "i", Field.Store.NO));
document.Add(new StringField("filter_1", "F", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "j", Field.Store.NO));
document.Add(new StringField("filter_1", "F", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "k", Field.Store.NO));
document.Add(new StringField("filter_1", "F", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("__type", "parent", Field.Store.NO));
document.Add(new StringField("field1", "f", Field.Store.NO));
docs.Add(document);
w.AddDocuments(docs);
docs.Clear();
document = new Document();
document.Add(new StringField("field2", "k", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "l", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "m", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("__type", "parent", Field.Store.NO));
document.Add(new StringField("field1", "g", Field.Store.NO));
docs.Add(document);
w.AddDocuments(docs);
// This doc will not be included, because it doesn't have nested docs
document = new Document();
document.Add(new StringField("__type", "parent", Field.Store.NO));
document.Add(new StringField("field1", "h", Field.Store.NO));
w.AddDocument(document);
docs.Clear();
document = new Document();
document.Add(new StringField("field2", "m", Field.Store.NO));
document.Add(new StringField("filter_1", "T", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "n", Field.Store.NO));
document.Add(new StringField("filter_1", "F", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("field2", "o", Field.Store.NO));
document.Add(new StringField("filter_1", "F", Field.Store.NO));
docs.Add(document);
document = new Document();
document.Add(new StringField("__type", "parent", Field.Store.NO));
document.Add(new StringField("field1", "i", Field.Store.NO));
docs.Add(document);
w.AddDocuments(docs);
w.Commit();
// Some garbage docs, just to check if the NestedFieldComparer can deal with this.
document = new Document();
document.Add(new StringField("fieldXXX", "x", Field.Store.NO));
w.AddDocument(document);
document = new Document();
document.Add(new StringField("fieldXXX", "x", Field.Store.NO));
w.AddDocument(document);
document = new Document();
document.Add(new StringField("fieldXXX", "x", Field.Store.NO));
w.AddDocument(document);
IndexSearcher searcher = new IndexSearcher(DirectoryReader.Open(w.IndexWriter, false));
w.Dispose();
Filter parentFilter = new QueryWrapperFilter(new TermQuery(new Term("__type", "parent")));
Filter childFilter = new QueryWrapperFilter(new PrefixQuery(new Term("field2")));
ToParentBlockJoinQuery query = new ToParentBlockJoinQuery(new FilteredQuery(new MatchAllDocsQuery(), childFilter), new FixedBitSetCachingWrapperFilter(parentFilter), ScoreMode.None);
// Sort by field ascending, order first
ToParentBlockJoinSortField sortField = new ToParentBlockJoinSortField("field2", SortFieldType.STRING, false, Wrap(parentFilter), Wrap(childFilter));
Sort sort = new Sort(sortField);
TopFieldDocs topDocs = searcher.Search(query, 5, sort);
assertEquals(7, topDocs.TotalHits);
assertEquals(5, topDocs.ScoreDocs.Length);
assertEquals(3, topDocs.ScoreDocs[0].Doc);
assertEquals("a", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[0]).Fields[0]).Utf8ToString());
assertEquals(7, topDocs.ScoreDocs[1].Doc);
assertEquals("c", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[1]).Fields[0]).Utf8ToString());
assertEquals(11, topDocs.ScoreDocs[2].Doc);
assertEquals("e", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[2]).Fields[0]).Utf8ToString());
assertEquals(15, topDocs.ScoreDocs[3].Doc);
assertEquals("g", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[3]).Fields[0]).Utf8ToString());
assertEquals(19, topDocs.ScoreDocs[4].Doc);
assertEquals("i", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[4]).Fields[0]).Utf8ToString());
// Sort by field ascending, order last
sortField = new ToParentBlockJoinSortField("field2", SortFieldType.STRING, false, true, Wrap(parentFilter), Wrap(childFilter));
sort = new Sort(sortField);
topDocs = searcher.Search(query, 5, sort);
assertEquals(7, topDocs.TotalHits);
assertEquals(5, topDocs.ScoreDocs.Length);
assertEquals(3, topDocs.ScoreDocs[0].Doc);
assertEquals("c", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[0]).Fields[0]).Utf8ToString());
assertEquals(7, topDocs.ScoreDocs[1].Doc);
assertEquals("e", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[1]).Fields[0]).Utf8ToString());
assertEquals(11, topDocs.ScoreDocs[2].Doc);
assertEquals("g", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[2]).Fields[0]).Utf8ToString());
assertEquals(15, topDocs.ScoreDocs[3].Doc);
assertEquals("i", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[3]).Fields[0]).Utf8ToString());
assertEquals(19, topDocs.ScoreDocs[4].Doc);
assertEquals("k", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[4]).Fields[0]).Utf8ToString());
// Sort by field descending, order last
sortField = new ToParentBlockJoinSortField("field2", SortFieldType.STRING, true, Wrap(parentFilter), Wrap(childFilter));
sort = new Sort(sortField);
topDocs = searcher.Search(query, 5, sort);
assertEquals(topDocs.TotalHits, 7);
assertEquals(5, topDocs.ScoreDocs.Length);
assertEquals(28, topDocs.ScoreDocs[0].Doc);
assertEquals("o", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[0]).Fields[0]).Utf8ToString());
assertEquals(23, topDocs.ScoreDocs[1].Doc);
assertEquals("m", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[1]).Fields[0]).Utf8ToString());
assertEquals(19, topDocs.ScoreDocs[2].Doc);
assertEquals("k", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[2]).Fields[0]).Utf8ToString());
assertEquals(15, topDocs.ScoreDocs[3].Doc);
assertEquals("i", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[3]).Fields[0]).Utf8ToString());
assertEquals(11, topDocs.ScoreDocs[4].Doc);
assertEquals("g", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[4]).Fields[0]).Utf8ToString());
// Sort by field descending, order last, sort filter (filter_1:T)
childFilter = new QueryWrapperFilter(new TermQuery((new Term("filter_1", "T"))));
query = new ToParentBlockJoinQuery(
new FilteredQuery(new MatchAllDocsQuery(), childFilter),
new FixedBitSetCachingWrapperFilter(parentFilter),
ScoreMode.None);
sortField = new ToParentBlockJoinSortField("field2", SortFieldType.STRING, true, Wrap(parentFilter), Wrap(childFilter));
sort = new Sort(sortField);
topDocs = searcher.Search(query, 5, sort);
assertEquals(6, topDocs.TotalHits);
assertEquals(5, topDocs.ScoreDocs.Length);
assertEquals(23, topDocs.ScoreDocs[0].Doc);
assertEquals("m", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[0]).Fields[0]).Utf8ToString());
assertEquals(28, topDocs.ScoreDocs[1].Doc);
assertEquals("m", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[1]).Fields[0]).Utf8ToString());
assertEquals(11, topDocs.ScoreDocs[2].Doc);
assertEquals("g", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[2]).Fields[0]).Utf8ToString());
assertEquals(15, topDocs.ScoreDocs[3].Doc);
assertEquals("g", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[3]).Fields[0]).Utf8ToString());
assertEquals(7, topDocs.ScoreDocs[4].Doc);
assertEquals("e", ((BytesRef)((FieldDoc)topDocs.ScoreDocs[4]).Fields[0]).Utf8ToString());
searcher.IndexReader.Dispose();
dir.Dispose();
}
