protected internal virtual void ToDotFile(Analyzer a, string inputText, string localFileName)
{
StreamWriter w = new StreamWriter(new FileStream(localFileName, FileMode.Open), IOUtils.CHARSET_UTF_8);
TokenStream ts = a.TokenStream("field", new StreamReader(inputText));
ts.Reset();
(new TokenStreamToDot(inputText, ts, /* new PrintWriter(*/ w /*)*/)).ToDot();
w.Close();
}
protected internal virtual string ToDot(Analyzer a, string inputText)
{
StringWriter sw = new StringWriter();
TokenStream ts = a.TokenStream("field", new StringReader(inputText));
ts.Reset();
(new TokenStreamToDot(inputText, ts, /*new StreamWriter(*/ (TextWriter)sw /*)*/)).ToDot();
return(sw.ToString());
}
public virtual void AssertThreadSafe(Analyzer analyzer)
{
int numTestPoints = 100;
int numThreads = TestUtil.NextInt(Random(), 3, 5);
Dictionary <string, BytesRef> map = new Dictionary <string, BytesRef>();
// create a map<String,SortKey> up front.
// then with multiple threads, generate sort keys for all the keys in the map
// and ensure they are the same as the ones we produced in serial fashion.
for (int i = 0; i < numTestPoints; i++)
{
string term = TestUtil.RandomSimpleString(Random());
IOException priorException = null;
TokenStream ts = analyzer.TokenStream("fake", new StreamReader(term));
try
{
ITermToBytesRefAttribute termAtt = ts.AddAttribute <ITermToBytesRefAttribute>();
BytesRef bytes = termAtt.BytesRef;
ts.Reset();
Assert.IsTrue(ts.IncrementToken());
termAtt.FillBytesRef();
// ensure we make a copy of the actual bytes too
map[term] = BytesRef.DeepCopyOf(bytes);
Assert.IsFalse(ts.IncrementToken());
ts.End();
}
catch (IOException e)
{
priorException = e;
}
finally
{
IOUtils.CloseWhileHandlingException(priorException, ts);
}
}
ThreadClass[] threads = new ThreadClass[numThreads];
for (int i = 0; i < numThreads; i++)
{
threads[i] = new ThreadAnonymousInnerClassHelper(this, analyzer, map);
}
for (int i = 0; i < numThreads; i++)
{
threads[i].Start();
}
for (int i = 0; i < numThreads; i++)
{
threads[i].Join();
}
}
public void TestLUCENE_3042()
{
String testString = "t";
Analyzer analyzer = new MockAnalyzer(Random);
using (TokenStream stream = analyzer.GetTokenStream("dummy", testString))
{
stream.Reset();
while (stream.IncrementToken())
{
// consume
}
stream.End();
}
AssertAnalyzesTo(analyzer, testString, new String[] { "t" });
}
public void TestForwardOffsets()
{
int num = AtLeast(1000);
for (int i = 0; i < num; i++)
{
String s = TestUtil.RandomHtmlishString(Random, 20);
StringReader reader = new StringReader(s);
MockCharFilter charfilter = new MockCharFilter(reader, 2);
MockAnalyzer analyzer = new MockAnalyzer(Random);
using TokenStream ts = analyzer.GetTokenStream("bogus", charfilter);
ts.Reset();
while (ts.IncrementToken())
{
;
}
ts.End();
}
}
private static void ConsumeStreamVeryOldAPI(TokenStream stream)
{
stream.Reset();
Token token;
int i = 0;
while ((token = stream.Next()) != null)
{
System.String term = token.Term();
Payload p = token.GetPayload();
if (p != null && p.GetData().Length == 1 && p.GetData()[0] == PartOfSpeechAnnotatingFilter.PROPER_NOUN_ANNOTATION)
{
Assert.IsTrue("tokenstream".Equals(term), "only TokenStream is a proper noun");
}
else
{
Assert.IsFalse("tokenstream".Equals(term), "all other tokens (if this test fails, the special POSToken subclass is not correctly passed through the chain)");
}
Assert.AreEqual(results[i], term);
i++;
}
}
public override void Run()
{
try
{
foreach (var mapping in this.map)
{
string term = mapping.Key;
BytesRef expected = mapping.Value;
Exception priorException = null; // LUCENENET: No need to cast to IOExcpetion
TokenStream ts = this.analyzer.GetTokenStream("fake", new StringReader(term));
try
{
ITermToBytesRefAttribute termAtt = ts.AddAttribute <ITermToBytesRefAttribute>();
BytesRef bytes = termAtt.BytesRef;
ts.Reset();
Assert.IsTrue(ts.IncrementToken());
termAtt.FillBytesRef();
Assert.AreEqual(expected, bytes);
Assert.IsFalse(ts.IncrementToken());
ts.End();
}
catch (Exception e) when(e.IsIOException())
{
priorException = e;
}
finally
{
IOUtils.DisposeWhileHandlingException(priorException, ts);
}
}
}
catch (Exception e) when(e.IsIOException())
{
throw RuntimeException.Create(e);
}
}
private static void ConsumeStreamNewAPI(TokenStream stream)
{
stream.Reset();
PayloadAttribute payloadAtt = (PayloadAttribute)stream.AddAttribute(typeof(PayloadAttribute));
TermAttribute termAtt = (TermAttribute)stream.AddAttribute(typeof(TermAttribute));
int i = 0;
while (stream.IncrementToken())
{
System.String term = termAtt.Term();
Payload p = payloadAtt.GetPayload();
if (p != null && p.GetData().Length == 1 && p.GetData()[0] == PartOfSpeechAnnotatingFilter.PROPER_NOUN_ANNOTATION)
{
Assert.IsTrue("tokenstream".Equals(term), "only TokenStream is a proper noun");
}
else
{
Assert.IsFalse("tokenstream".Equals(term), "all other tokens (if this test fails, the special POSToken subclass is not correctly passed through the chain)");
}
Assert.AreEqual(results[i], term);
i++;
}
}
public override void Run()
{
try
{
foreach (KeyValuePair <string, BytesRef> mapping in Map)
{
string term = mapping.Key;
BytesRef expected = mapping.Value;
IOException priorException = null;
TokenStream ts = Analyzer.TokenStream("fake", new StreamReader(term));
try
{
ITermToBytesRefAttribute termAtt = ts.AddAttribute <ITermToBytesRefAttribute>();
BytesRef bytes = termAtt.BytesRef;
ts.Reset();
Assert.IsTrue(ts.IncrementToken());
termAtt.FillBytesRef();
Assert.AreEqual(expected, bytes);
Assert.IsFalse(ts.IncrementToken());
ts.End();
}
catch (IOException e)
{
priorException = e;
}
finally
{
IOUtils.CloseWhileHandlingException(priorException, ts);
}
}
}
catch (IOException e)
{
throw (Exception)e;
}
}
public static void AssertTokenStreamContents(TokenStream ts, System.String[] output, int[] startOffsets, int[] endOffsets, System.String[] types, int[] posIncrements, int? finalOffset)
{
Assert.IsNotNull(output);
ICheckClearAttributesAttribute checkClearAtt = ts.AddAttribute<ICheckClearAttributesAttribute>();
Assert.IsTrue(ts.HasAttribute<ITermAttribute>(), "has no TermAttribute");
ITermAttribute termAtt = ts.GetAttribute<ITermAttribute>();
IOffsetAttribute offsetAtt = null;
if (startOffsets != null || endOffsets != null || finalOffset != null)
{
Assert.IsTrue(ts.HasAttribute<IOffsetAttribute>(), "has no OffsetAttribute");
offsetAtt = ts.GetAttribute<IOffsetAttribute>();
}
ITypeAttribute typeAtt = null;
if (types != null)
{
Assert.IsTrue(ts.HasAttribute<ITypeAttribute>(), "has no TypeAttribute");
typeAtt = ts.GetAttribute<ITypeAttribute>();
}
IPositionIncrementAttribute posIncrAtt = null;
if (posIncrements != null)
{
Assert.IsTrue(ts.HasAttribute<IPositionIncrementAttribute>(), "has no PositionIncrementAttribute");
posIncrAtt = ts.GetAttribute<IPositionIncrementAttribute>();
}
ts.Reset();
for (int i = 0; i < output.Length; i++)
{
// extra safety to enforce, that the state is not preserved and also assign bogus values
ts.ClearAttributes();
termAtt.SetTermBuffer("bogusTerm");
if (offsetAtt != null) offsetAtt.SetOffset(14584724, 24683243);
if (typeAtt != null) typeAtt.Type = "bogusType";
if (posIncrAtt != null) posIncrAtt.PositionIncrement = 45987657;
checkClearAtt.GetAndResetClearCalled(); // reset it, because we called clearAttribute() before
Assert.IsTrue(ts.IncrementToken(), "token " + i + " does not exist");
Assert.IsTrue(checkClearAtt.GetAndResetClearCalled(), "clearAttributes() was not called correctly in TokenStream chain");
Assert.AreEqual(output[i], termAtt.Term, "term " + i);
if (startOffsets != null)
Assert.AreEqual(startOffsets[i], offsetAtt.StartOffset, "startOffset " + i);
if (endOffsets != null)
Assert.AreEqual(endOffsets[i], offsetAtt.EndOffset, "endOffset " + i);
if (types != null)
Assert.AreEqual(types[i], typeAtt.Type, "type " + i);
if (posIncrements != null)
Assert.AreEqual(posIncrements[i], posIncrAtt.PositionIncrement, "posIncrement " + i);
}
Assert.IsFalse(ts.IncrementToken(), "end of stream");
ts.End();
if (finalOffset.HasValue)
Assert.AreEqual(finalOffset, offsetAtt.EndOffset, "finalOffset ");
ts.Close();
}
private static void ConsumeStreamVeryOldAPI(TokenStream stream)
{
stream.Reset();
Token token;
int i = 0;
while ((token = stream.Next()) != null)
{
System.String term = token.Term();
Payload p = token.GetPayload();
if (p != null && p.GetData().Length == 1 && p.GetData()[0] == PartOfSpeechAnnotatingFilter.PROPER_NOUN_ANNOTATION)
{
Assert.IsTrue("tokenstream".Equals(term), "only TokenStream is a proper noun");
}
else
{
Assert.IsFalse("tokenstream".Equals(term), "all other tokens (if this test fails, the special POSToken subclass is not correctly passed through the chain)");
}
Assert.AreEqual(results[i], term);
i++;
}
}
private static void ConsumeStreamNewAPI(TokenStream stream)
{
stream.Reset();
PayloadAttribute payloadAtt = (PayloadAttribute) stream.AddAttribute(typeof(PayloadAttribute));
TermAttribute termAtt = (TermAttribute) stream.AddAttribute(typeof(TermAttribute));
int i = 0;
while (stream.IncrementToken())
{
System.String term = termAtt.Term();
Payload p = payloadAtt.GetPayload();
if (p != null && p.GetData().Length == 1 && p.GetData()[0] == PartOfSpeechAnnotatingFilter.PROPER_NOUN_ANNOTATION)
{
Assert.IsTrue("tokenstream".Equals(term), "only TokenStream is a proper noun");
}
else
{
Assert.IsFalse("tokenstream".Equals(term), "all other tokens (if this test fails, the special POSToken subclass is not correctly passed through the chain)");
}
Assert.AreEqual(results[i], term);
i++;
}
}
public virtual void ToDot()
{
@in.Reset();
WriteHeader();
// TODO: is there some way to tell dot that it should
// make the "main path" a straight line and have the
// non-sausage arcs not affect node placement...
int pos = -1;
int lastEndPos = -1;
while (@in.IncrementToken())
{
bool isFirst = pos == -1;
int posInc = PosIncAtt.PositionIncrement;
if (isFirst && posInc == 0)
{
// TODO: hmm are TS's still allowed to do this...?
Console.Error.WriteLine("WARNING: first posInc was 0; correcting to 1");
posInc = 1;
}
if (posInc > 0)
{
// New node:
pos += posInc;
WriteNode(pos, Convert.ToString(pos));
}
if (posInc > 1)
{
// Gap!
WriteArc(lastEndPos, pos, null, "dotted");
}
if (isFirst)
{
WriteNode(-1, null);
WriteArc(-1, pos, null, null);
}
string arcLabel = TermAtt.ToString();
if (OffsetAtt != null)
{
int startOffset = OffsetAtt.StartOffset;
int endOffset = OffsetAtt.EndOffset;
//System.out.println("start=" + startOffset + " end=" + endOffset + " len=" + inputText.length());
if (InputText != null)
{
arcLabel += " / " + InputText.Substring(startOffset, endOffset - startOffset);
}
else
{
arcLabel += " / " + startOffset + "-" + endOffset;
}
}
WriteArc(pos, pos + PosLengthAtt.PositionLength, arcLabel, null);
lastEndPos = pos + PosLengthAtt.PositionLength;
}
@in.End();
if (lastEndPos != -1)
{
// TODO: should we output any final text (from end
// offsets) on this arc...?
WriteNode(-2, null);
WriteArc(lastEndPos, -2, null, null);
}
WriteTrailer();
}
/// <summary>Reset the filter as well as the input TokenStream. </summary>
public override void Reset()
{
input.Reset();
}
/// <summary>
/// Pulls the graph (including <see cref="IPositionLengthAttribute"/>
/// from the provided <see cref="TokenStream"/>, and creates the corresponding
/// automaton where arcs are bytes (or Unicode code points
/// if unicodeArcs = true) from each term.
/// </summary>
public virtual Automaton ToAutomaton(TokenStream @in)
{
var a = new Automaton();
bool deterministic = true;
var posIncAtt = @in.AddAttribute <IPositionIncrementAttribute>();
var posLengthAtt = @in.AddAttribute <IPositionLengthAttribute>();
var offsetAtt = @in.AddAttribute <IOffsetAttribute>();
var termBytesAtt = @in.AddAttribute <ITermToBytesRefAttribute>();
BytesRef term = termBytesAtt.BytesRef;
@in.Reset();
// Only temporarily holds states ahead of our current
// position:
RollingBuffer <Position> positions = new Positions();
int pos = -1;
Position posData = null;
int maxOffset = 0;
while (@in.IncrementToken())
{
int posInc = posIncAtt.PositionIncrement;
if (!preservePositionIncrements && posInc > 1)
{
posInc = 1;
}
Debug.Assert(pos > -1 || posInc > 0);
if (posInc > 0)
{
// New node:
pos += posInc;
posData = positions.Get(pos);
Debug.Assert(posData.leaving == null);
if (posData.arriving == null)
{
// No token ever arrived to this position
if (pos == 0)
{
// OK: this is the first token
posData.leaving = a.GetInitialState();
}
else
{
// this means there's a hole (eg, StopFilter
// does this):
posData.leaving = new State();
AddHoles(a.GetInitialState(), positions, pos);
}
}
else
{
posData.leaving = new State();
posData.arriving.AddTransition(new Transition(POS_SEP, posData.leaving));
if (posInc > 1)
{
// A token spanned over a hole; add holes
// "under" it:
AddHoles(a.GetInitialState(), positions, pos);
}
}
positions.FreeBefore(pos);
}
else
{
// note: this isn't necessarily true. its just that we aren't surely det.
// we could optimize this further (e.g. buffer and sort synonyms at a position)
// but thats probably overkill. this is cheap and dirty
deterministic = false;
}
int endPos = pos + posLengthAtt.PositionLength;
termBytesAtt.FillBytesRef();
BytesRef termUTF8 = ChangeToken(term);
int[] termUnicode = null;
Position endPosData = positions.Get(endPos);
if (endPosData.arriving == null)
{
endPosData.arriving = new State();
}
State state = posData.leaving;
int termLen = termUTF8.Length;
if (unicodeArcs)
{
string utf16 = termUTF8.Utf8ToString();
termUnicode = new int[utf16.CodePointCount(0, utf16.Length)];
termLen = termUnicode.Length;
for (int cp, i = 0, j = 0; i < utf16.Length; i += Character.CharCount(cp))
{
termUnicode[j++] = cp = Character.CodePointAt(utf16, i);
}
}
else
{
termLen = termUTF8.Length;
}
for (int byteIDX = 0; byteIDX < termLen; byteIDX++)
{
State nextState = byteIDX == termLen - 1 ? endPosData.arriving : new State();
int c;
if (unicodeArcs)
{
c = termUnicode[byteIDX];
}
else
{
c = termUTF8.Bytes[termUTF8.Offset + byteIDX] & 0xff;
}
state.AddTransition(new Transition(c, nextState));
state = nextState;
}
maxOffset = Math.Max(maxOffset, offsetAtt.EndOffset);
}
@in.End();
State endState = null;
if (offsetAtt.EndOffset > maxOffset)
{
endState = new State();
endState.Accept = true;
}
pos++;
while (pos <= positions.MaxPos)
{
posData = positions.Get(pos);
if (posData.arriving != null)
{
if (endState != null)
{
posData.arriving.AddTransition(new Transition(POS_SEP, endState));
}
else
{
posData.arriving.Accept = true;
}
}
pos++;
}
//toDot(a);
a.IsDeterministic = deterministic;
return(a);
}
public static void AssertTokenStreamContents(TokenStream ts, System.String[] output, int[] startOffsets, int[] endOffsets, System.String[] types, int[] posIncrements, int?finalOffset)
{
Assert.IsNotNull(output);
CheckClearAttributesAttribute checkClearAtt = (CheckClearAttributesAttribute)ts.AddAttribute(typeof(CheckClearAttributesAttribute));
Assert.IsTrue(ts.HasAttribute(typeof(TermAttribute)), "has no TermAttribute");
TermAttribute termAtt = (TermAttribute)ts.GetAttribute(typeof(TermAttribute));
OffsetAttribute offsetAtt = null;
if (startOffsets != null || endOffsets != null || finalOffset != null)
{
Assert.IsTrue(ts.HasAttribute(typeof(OffsetAttribute)), "has no OffsetAttribute");
offsetAtt = (OffsetAttribute)ts.GetAttribute(typeof(OffsetAttribute));
}
TypeAttribute typeAtt = null;
if (types != null)
{
Assert.IsTrue(ts.HasAttribute(typeof(TypeAttribute)), "has no TypeAttribute");
typeAtt = (TypeAttribute)ts.GetAttribute(typeof(TypeAttribute));
}
PositionIncrementAttribute posIncrAtt = null;
if (posIncrements != null)
{
Assert.IsTrue(ts.HasAttribute(typeof(PositionIncrementAttribute)), "has no PositionIncrementAttribute");
posIncrAtt = (PositionIncrementAttribute)ts.GetAttribute(typeof(PositionIncrementAttribute));
}
ts.Reset();
for (int i = 0; i < output.Length; i++)
{
// extra safety to enforce, that the state is not preserved and also assign bogus values
ts.ClearAttributes();
termAtt.SetTermBuffer("bogusTerm");
if (offsetAtt != null)
{
offsetAtt.SetOffset(14584724, 24683243);
}
if (typeAtt != null)
{
typeAtt.SetType("bogusType");
}
if (posIncrAtt != null)
{
posIncrAtt.SetPositionIncrement(45987657);
}
checkClearAtt.GetAndResetClearCalled(); // reset it, because we called clearAttribute() before
Assert.IsTrue(ts.IncrementToken(), "token " + i + " does not exist");
Assert.IsTrue(checkClearAtt.GetAndResetClearCalled(), "clearAttributes() was not called correctly in TokenStream chain");
Assert.AreEqual(output[i], termAtt.Term(), "term " + i);
if (startOffsets != null)
{
Assert.AreEqual(startOffsets[i], offsetAtt.StartOffset(), "startOffset " + i);
}
if (endOffsets != null)
{
Assert.AreEqual(endOffsets[i], offsetAtt.EndOffset(), "endOffset " + i);
}
if (types != null)
{
Assert.AreEqual(types[i], typeAtt.Type(), "type " + i);
}
if (posIncrements != null)
{
Assert.AreEqual(posIncrements[i], posIncrAtt.GetPositionIncrement(), "posIncrement " + i);
}
}
Assert.IsFalse(ts.IncrementToken(), "end of stream");
ts.End();
if (finalOffset.HasValue)
{
Assert.AreEqual(finalOffset, offsetAtt.EndOffset(), "finalOffset ");
}
ts.Close();
}
// offsetsAreCorrect also validates:
// - graph offsets are correct (all tokens leaving from
// pos X have the same startOffset; all tokens
// arriving to pos Y have the same endOffset)
// - offsets only move forwards (startOffset >=
// lastStartOffset)
public static void AssertTokenStreamContents(TokenStream ts, string[] output, int[] startOffsets, int[] endOffsets, string[] types, int[] posIncrements, int[] posLengths, int? finalOffset, int? finalPosInc, bool[] keywordAtts, bool offsetsAreCorrect)
{
// LUCENENET: Bug fix: NUnit throws an exception when something fails.
// This causes Dispose() to be skipped and it pollutes other tests indicating false negatives.
// Added this try-finally block to fix this.
try
{
Assert.IsNotNull(output);
var checkClearAtt = ts.AddAttribute<ICheckClearAttributesAttribute>();
ICharTermAttribute termAtt = null;
if (output.Length > 0)
{
Assert.IsTrue(ts.HasAttribute<ICharTermAttribute>(), "has no CharTermAttribute");
termAtt = ts.GetAttribute<ICharTermAttribute>();
}
IOffsetAttribute offsetAtt = null;
if (startOffsets != null || endOffsets != null || finalOffset != null)
{
Assert.IsTrue(ts.HasAttribute<IOffsetAttribute>(), "has no OffsetAttribute");
offsetAtt = ts.GetAttribute<IOffsetAttribute>();
}
ITypeAttribute typeAtt = null;
if (types != null)
{
Assert.IsTrue(ts.HasAttribute<ITypeAttribute>(), "has no TypeAttribute");
typeAtt = ts.GetAttribute<ITypeAttribute>();
}
IPositionIncrementAttribute posIncrAtt = null;
if (posIncrements != null || finalPosInc != null)
{
Assert.IsTrue(ts.HasAttribute<IPositionIncrementAttribute>(), "has no PositionIncrementAttribute");
posIncrAtt = ts.GetAttribute<IPositionIncrementAttribute>();
}
IPositionLengthAttribute posLengthAtt = null;
if (posLengths != null)
{
Assert.IsTrue(ts.HasAttribute<IPositionLengthAttribute>(), "has no PositionLengthAttribute");
posLengthAtt = ts.GetAttribute<IPositionLengthAttribute>();
}
IKeywordAttribute keywordAtt = null;
if (keywordAtts != null)
{
Assert.IsTrue(ts.HasAttribute<IKeywordAttribute>(), "has no KeywordAttribute");
keywordAtt = ts.GetAttribute<IKeywordAttribute>();
}
// Maps position to the start/end offset:
IDictionary<int?, int?> posToStartOffset = new Dictionary<int?, int?>();
IDictionary<int?, int?> posToEndOffset = new Dictionary<int?, int?>();
ts.Reset();
int pos = -1;
int lastStartOffset = 0;
for (int i = 0; i < output.Length; i++)
{
// extra safety to enforce, that the state is not preserved and also assign bogus values
ts.ClearAttributes();
termAtt.SetEmpty().Append("bogusTerm");
if (offsetAtt != null)
{
offsetAtt.SetOffset(14584724, 24683243);
}
if (typeAtt != null)
{
typeAtt.Type = "bogusType";
}
if (posIncrAtt != null)
{
posIncrAtt.PositionIncrement = 45987657;
}
if (posLengthAtt != null)
{
posLengthAtt.PositionLength = 45987653;
}
if (keywordAtt != null)
{
keywordAtt.Keyword = (i & 1) == 0;
}
bool reset = checkClearAtt.AndResetClearCalled; // reset it, because we called clearAttribute() before
Assert.IsTrue(ts.IncrementToken(), "token " + i + " does not exist");
Assert.IsTrue(reset, "ClearAttributes() was not called correctly in TokenStream chain");
Assert.AreEqual(output[i], termAtt.ToString(), "term " + i + ", output[i] = " + output[i] + ", termAtt = " + termAtt.ToString());
if (startOffsets != null)
{
Assert.AreEqual(startOffsets[i], offsetAtt.StartOffset(), "startOffset " + i);
}
if (endOffsets != null)
{
Assert.AreEqual(endOffsets[i], offsetAtt.EndOffset(), "endOffset " + i);
}
if (types != null)
{
Assert.AreEqual(types[i], typeAtt.Type, "type " + i);
}
if (posIncrements != null)
{
Assert.AreEqual(posIncrements[i], posIncrAtt.PositionIncrement, "posIncrement " + i);
}
if (posLengths != null)
{
Assert.AreEqual(posLengths[i], posLengthAtt.PositionLength, "posLength " + i);
}
if (keywordAtts != null)
{
Assert.AreEqual(keywordAtts[i], keywordAtt.Keyword, "keywordAtt " + i);
}
// we can enforce some basic things about a few attributes even if the caller doesn't check:
if (offsetAtt != null)
{
int startOffset = offsetAtt.StartOffset();
int endOffset = offsetAtt.EndOffset();
if (finalOffset != null)
{
Assert.IsTrue(startOffset <= (int)finalOffset, "startOffset must be <= finalOffset");
Assert.IsTrue(endOffset <= (int)finalOffset, "endOffset must be <= finalOffset: got endOffset=" + endOffset + " vs finalOffset=" + (int)finalOffset);
}
if (offsetsAreCorrect)
{
Assert.IsTrue(offsetAtt.StartOffset() >= lastStartOffset, "offsets must not go backwards startOffset=" + startOffset + " is < lastStartOffset=" + lastStartOffset);
lastStartOffset = offsetAtt.StartOffset();
}
if (offsetsAreCorrect && posLengthAtt != null && posIncrAtt != null)
{
// Validate offset consistency in the graph, ie
// all tokens leaving from a certain pos have the
// same startOffset, and all tokens arriving to a
// certain pos have the same endOffset:
int posInc = posIncrAtt.PositionIncrement;
pos += posInc;
int posLength = posLengthAtt.PositionLength;
if (!posToStartOffset.ContainsKey(pos))
{
// First time we've seen a token leaving from this position:
posToStartOffset[pos] = startOffset;
//System.out.println(" + s " + pos + " -> " + startOffset);
}
else
{
// We've seen a token leaving from this position
// before; verify the startOffset is the same:
//System.out.println(" + vs " + pos + " -> " + startOffset);
Assert.AreEqual((int)posToStartOffset[pos], startOffset, "pos=" + pos + " posLen=" + posLength + " token=" + termAtt);
}
int endPos = pos + posLength;
if (!posToEndOffset.ContainsKey(endPos))
{
// First time we've seen a token arriving to this position:
posToEndOffset[endPos] = endOffset;
//System.out.println(" + e " + endPos + " -> " + endOffset);
}
else
{
// We've seen a token arriving to this position
// before; verify the endOffset is the same:
//System.out.println(" + ve " + endPos + " -> " + endOffset);
Assert.AreEqual((int)posToEndOffset[endPos], endOffset, "pos=" + pos + " posLen=" + posLength + " token=" + termAtt);
}
}
}
if (posIncrAtt != null)
{
if (i == 0)
{
Assert.IsTrue(posIncrAtt.PositionIncrement >= 1, "first posIncrement must be >= 1");
}
else
{
Assert.IsTrue(posIncrAtt.PositionIncrement >= 0, "posIncrement must be >= 0");
}
}
if (posLengthAtt != null)
{
Assert.IsTrue(posLengthAtt.PositionLength >= 1, "posLength must be >= 1");
}
}
if (ts.IncrementToken())
{
Assert.Fail("TokenStream has more tokens than expected (expected count=" + output.Length + "); extra token=" + termAtt);
}
// repeat our extra safety checks for End()
ts.ClearAttributes();
if (termAtt != null)
{
termAtt.SetEmpty().Append("bogusTerm");
}
if (offsetAtt != null)
{
offsetAtt.SetOffset(14584724, 24683243);
}
if (typeAtt != null)
{
typeAtt.Type = "bogusType";
}
if (posIncrAtt != null)
{
posIncrAtt.PositionIncrement = 45987657;
}
if (posLengthAtt != null)
{
posLengthAtt.PositionLength = 45987653;
}
var reset_ = checkClearAtt.AndResetClearCalled; // reset it, because we called clearAttribute() before
ts.End();
Assert.IsTrue(checkClearAtt.AndResetClearCalled, "super.End()/ClearAttributes() was not called correctly in End()");
if (finalOffset != null)
{
Assert.AreEqual((int)finalOffset, offsetAtt.EndOffset(), "finalOffset");
}
if (offsetAtt != null)
{
Assert.IsTrue(offsetAtt.EndOffset() >= 0, "finalOffset must be >= 0");
}
if (finalPosInc != null)
{
Assert.AreEqual((int)finalPosInc, posIncrAtt.PositionIncrement, "finalPosInc");
}
ts.Dispose();
}
catch (Exception)
{
//ts.Reset();
ts.ClearAttributes();
ts.End();
ts.Dispose();
throw;
}
}
/// <summary>
/// Reset the filter as well as the input TokenStream.
/// </summary>
public override void Reset()
{
base.Reset();
input.Reset();
}
// offsetsAreCorrect also validates:
// - graph offsets are correct (all tokens leaving from
// pos X have the same startOffset; all tokens
// arriving to pos Y have the same endOffset)
// - offsets only move forwards (startOffset >=
// lastStartOffset)
public static void AssertTokenStreamContents(TokenStream ts, string[] output, int[] startOffsets, int[] endOffsets, string[] types, int[] posIncrements, int[] posLengths, int?finalOffset, int?finalPosInc, bool[] keywordAtts, bool offsetsAreCorrect)
{
Assert.IsNotNull(output);
var checkClearAtt = ts.AddAttribute <ICheckClearAttributesAttribute>();
ICharTermAttribute termAtt = null;
if (output.Length > 0)
{
Assert.IsTrue(ts.HasAttribute <ICharTermAttribute>(), "has no CharTermAttribute");
termAtt = ts.GetAttribute <ICharTermAttribute>();
}
IOffsetAttribute offsetAtt = null;
if (startOffsets != null || endOffsets != null || finalOffset != null)
{
Assert.IsTrue(ts.HasAttribute <IOffsetAttribute>(), "has no OffsetAttribute");
offsetAtt = ts.GetAttribute <IOffsetAttribute>();
}
ITypeAttribute typeAtt = null;
if (types != null)
{
Assert.IsTrue(ts.HasAttribute <ITypeAttribute>(), "has no TypeAttribute");
typeAtt = ts.GetAttribute <ITypeAttribute>();
}
IPositionIncrementAttribute posIncrAtt = null;
if (posIncrements != null || finalPosInc != null)
{
Assert.IsTrue(ts.HasAttribute <IPositionIncrementAttribute>(), "has no PositionIncrementAttribute");
posIncrAtt = ts.GetAttribute <IPositionIncrementAttribute>();
}
IPositionLengthAttribute posLengthAtt = null;
if (posLengths != null)
{
Assert.IsTrue(ts.HasAttribute <IPositionLengthAttribute>(), "has no PositionLengthAttribute");
posLengthAtt = ts.GetAttribute <IPositionLengthAttribute>();
}
IKeywordAttribute keywordAtt = null;
if (keywordAtts != null)
{
Assert.IsTrue(ts.HasAttribute <IKeywordAttribute>(), "has no KeywordAttribute");
keywordAtt = ts.GetAttribute <IKeywordAttribute>();
}
// Maps position to the start/end offset:
IDictionary <int?, int?> posToStartOffset = new Dictionary <int?, int?>();
IDictionary <int?, int?> posToEndOffset = new Dictionary <int?, int?>();
ts.Reset();
int pos = -1;
int lastStartOffset = 0;
for (int i = 0; i < output.Length; i++)
{
// extra safety to enforce, that the state is not preserved and also assign bogus values
ts.ClearAttributes();
termAtt.SetEmpty().Append("bogusTerm");
if (offsetAtt != null)
{
offsetAtt.SetOffset(14584724, 24683243);
}
if (typeAtt != null)
{
typeAtt.Type = "bogusType";
}
if (posIncrAtt != null)
{
posIncrAtt.PositionIncrement = 45987657;
}
if (posLengthAtt != null)
{
posLengthAtt.PositionLength = 45987653;
}
if (keywordAtt != null)
{
keywordAtt.Keyword = (i & 1) == 0;
}
bool reset = checkClearAtt.AndResetClearCalled; // reset it, because we called clearAttribute() before
Assert.IsTrue(ts.IncrementToken(), "token " + i + " does not exist");
Assert.IsTrue(reset, "ClearAttributes() was not called correctly in TokenStream chain");
Assert.AreEqual(output[i], termAtt.ToString(), "term " + i + ", output[i] = " + output[i] + ", termAtt = " + termAtt.ToString());
if (startOffsets != null)
{
Assert.AreEqual(startOffsets[i], offsetAtt.StartOffset(), "startOffset " + i);
}
if (endOffsets != null)
{
Assert.AreEqual(endOffsets[i], offsetAtt.EndOffset(), "endOffset " + i);
}
if (types != null)
{
Assert.AreEqual(types[i], typeAtt.Type, "type " + i);
}
if (posIncrements != null)
{
Assert.AreEqual(posIncrements[i], posIncrAtt.PositionIncrement, "posIncrement " + i);
}
if (posLengths != null)
{
Assert.AreEqual(posLengths[i], posLengthAtt.PositionLength, "posLength " + i);
}
if (keywordAtts != null)
{
Assert.AreEqual(keywordAtts[i], keywordAtt.Keyword, "keywordAtt " + i);
}
// we can enforce some basic things about a few attributes even if the caller doesn't check:
if (offsetAtt != null)
{
int startOffset = offsetAtt.StartOffset();
int endOffset = offsetAtt.EndOffset();
if (finalOffset != null)
{
Assert.IsTrue(startOffset <= (int)finalOffset, "startOffset must be <= finalOffset");
Assert.IsTrue(endOffset <= (int)finalOffset, "endOffset must be <= finalOffset: got endOffset=" + endOffset + " vs finalOffset=" + (int)finalOffset);
}
if (offsetsAreCorrect)
{
Assert.IsTrue(offsetAtt.StartOffset() >= lastStartOffset, "offsets must not go backwards startOffset=" + startOffset + " is < lastStartOffset=" + lastStartOffset);
lastStartOffset = offsetAtt.StartOffset();
}
if (offsetsAreCorrect && posLengthAtt != null && posIncrAtt != null)
{
// Validate offset consistency in the graph, ie
// all tokens leaving from a certain pos have the
// same startOffset, and all tokens arriving to a
// certain pos have the same endOffset:
int posInc = posIncrAtt.PositionIncrement;
pos += posInc;
int posLength = posLengthAtt.PositionLength;
if (!posToStartOffset.ContainsKey(pos))
{
// First time we've seen a token leaving from this position:
posToStartOffset[pos] = startOffset;
//System.out.println(" + s " + pos + " -> " + startOffset);
}
else
{
// We've seen a token leaving from this position
// before; verify the startOffset is the same:
//System.out.println(" + vs " + pos + " -> " + startOffset);
Assert.AreEqual((int)posToStartOffset[pos], startOffset, "pos=" + pos + " posLen=" + posLength + " token=" + termAtt);
}
int endPos = pos + posLength;
if (!posToEndOffset.ContainsKey(endPos))
{
// First time we've seen a token arriving to this position:
posToEndOffset[endPos] = endOffset;
//System.out.println(" + e " + endPos + " -> " + endOffset);
}
else
{
// We've seen a token arriving to this position
// before; verify the endOffset is the same:
//System.out.println(" + ve " + endPos + " -> " + endOffset);
Assert.AreEqual((int)posToEndOffset[endPos], endOffset, "pos=" + pos + " posLen=" + posLength + " token=" + termAtt);
}
}
}
if (posIncrAtt != null)
{
if (i == 0)
{
Assert.IsTrue(posIncrAtt.PositionIncrement >= 1, "first posIncrement must be >= 1");
}
else
{
Assert.IsTrue(posIncrAtt.PositionIncrement >= 0, "posIncrement must be >= 0");
}
}
if (posLengthAtt != null)
{
Assert.IsTrue(posLengthAtt.PositionLength >= 1, "posLength must be >= 1");
}
}
if (ts.IncrementToken())
{
Assert.Fail("TokenStream has more tokens than expected (expected count=" + output.Length + "); extra token=" + termAtt);
}
// repeat our extra safety checks for End()
ts.ClearAttributes();
if (termAtt != null)
{
termAtt.SetEmpty().Append("bogusTerm");
}
if (offsetAtt != null)
{
offsetAtt.SetOffset(14584724, 24683243);
}
if (typeAtt != null)
{
typeAtt.Type = "bogusType";
}
if (posIncrAtt != null)
{
posIncrAtt.PositionIncrement = 45987657;
}
if (posLengthAtt != null)
{
posLengthAtt.PositionLength = 45987653;
}
var reset_ = checkClearAtt.AndResetClearCalled; // reset it, because we called clearAttribute() before
ts.End();
Assert.IsTrue(checkClearAtt.AndResetClearCalled, "super.End()/ClearAttributes() was not called correctly in End()");
if (finalOffset != null)
{
Assert.AreEqual((int)finalOffset, offsetAtt.EndOffset(), "finalOffset");
}
if (offsetAtt != null)
{
Assert.IsTrue(offsetAtt.EndOffset() >= 0, "finalOffset must be >= 0");
}
if (finalPosInc != null)
{
Assert.AreEqual((int)finalPosInc, posIncrAtt.PositionIncrement, "finalPosInc");
}
ts.Dispose();
}
