/// <summary>
/// <para>Fills <see cref="inputWindow"/> with input stream tokens, if available,
/// shifting to the right if the window was previously full.
/// </para>
/// <para>
/// Resets <see cref="gramSize"/> to its minimum value.
/// </para>
/// </summary>
/// <exception cref="IOException"> if there's a problem getting the next token </exception>
private void ShiftInputWindow()
{
InputWindowToken firstToken = null;
if (inputWindow.Count > 0)
{
firstToken = inputWindow.First.Value;
inputWindow.Remove(firstToken);
}
while (inputWindow.Count < maxShingleSize)
{
if (null != firstToken) // recycle the firstToken, if available
{
if (null != GetNextToken(firstToken))
{
inputWindow.AddLast(firstToken); // the firstToken becomes the last
firstToken = null;
}
else
{
break; // end of input stream
}
}
else
{
InputWindowToken nextToken = GetNextToken(null);
if (null != nextToken)
{
inputWindow.AddLast(nextToken);
}
else
{
break; // end of input stream
}
}
}
if (outputUnigramsIfNoShingles && noShingleOutput && gramSize.MinValue > 1 && inputWindow.Count < minShingleSize)
{
gramSize.MinValue = 1;
}
gramSize.Reset();
isOutputHere = false;
}
/// <summary>
/// <para>Get the next token from the input stream.
/// </para>
/// <para>If the next token has <c>positionIncrement > 1</c>,
/// <c>positionIncrement - 1</c> <see cref="fillerToken"/>s are
/// inserted first.
/// </para>
/// </summary>
/// <param name="target"> Where to put the new token; if null, a new instance is created. </param>
/// <returns> On success, the populated token; null otherwise </returns>
/// <exception cref="IOException"> if the input stream has a problem </exception>
private InputWindowToken GetNextToken(InputWindowToken target)
{
InputWindowToken newTarget = target;
if (numFillerTokensToInsert > 0)
{
if (null == target)
{
newTarget = new InputWindowToken(this, nextInputStreamToken.CloneAttributes());
}
else
{
nextInputStreamToken.CopyTo(target.attSource);
}
// A filler token occupies no space
newTarget.offsetAtt.SetOffset(newTarget.offsetAtt.StartOffset, newTarget.offsetAtt.StartOffset);
newTarget.termAtt.CopyBuffer(fillerToken, 0, fillerToken.Length);
newTarget.isFiller = true;
--numFillerTokensToInsert;
}
else if (isNextInputStreamToken)
{
if (null == target)
{
newTarget = new InputWindowToken(this, nextInputStreamToken.CloneAttributes());
}
else
{
nextInputStreamToken.CopyTo(target.attSource);
}
isNextInputStreamToken = false;
newTarget.isFiller = false;
}
else if (!exhausted)
{
if (m_input.IncrementToken())
{
if (null == target)
{
newTarget = new InputWindowToken(this, CloneAttributes());
}
else
{
this.CopyTo(target.attSource);
}
if (posIncrAtt.PositionIncrement > 1)
{
// Each output shingle must contain at least one input token,
// so no more than (maxShingleSize - 1) filler tokens will be inserted.
numFillerTokensToInsert = Math.Min(posIncrAtt.PositionIncrement - 1, maxShingleSize - 1);
// Save the current token as the next input stream token
if (null == nextInputStreamToken)
{
nextInputStreamToken = CloneAttributes();
}
else
{
this.CopyTo(nextInputStreamToken);
}
isNextInputStreamToken = true;
// A filler token occupies no space
newTarget.offsetAtt.SetOffset(offsetAtt.StartOffset, offsetAtt.StartOffset);
newTarget.termAtt.CopyBuffer(fillerToken, 0, fillerToken.Length);
newTarget.isFiller = true;
--numFillerTokensToInsert;
}
else
{
newTarget.isFiller = false;
}
}
else
{
exhausted = true;
m_input.End();
endState = CaptureState();
numFillerTokensToInsert = Math.Min(posIncrAtt.PositionIncrement, maxShingleSize - 1);
if (numFillerTokensToInsert > 0)
{
nextInputStreamToken = new AttributeSource(this.GetAttributeFactory());
nextInputStreamToken.AddAttribute <ICharTermAttribute>();
IOffsetAttribute newOffsetAtt = nextInputStreamToken.AddAttribute <IOffsetAttribute>();
newOffsetAtt.SetOffset(offsetAtt.EndOffset, offsetAtt.EndOffset);
// Recurse/loop just once:
return(GetNextToken(target));
}
else
{
newTarget = null;
}
}
}
else
{
newTarget = null;
}
return(newTarget);
}
public override bool IncrementToken()
{
bool tokenAvailable = false;
int builtGramSize = 0;
if (gramSize.AtMinValue() || inputWindow.Count < gramSize.Value)
{
ShiftInputWindow();
gramBuilder.Length = 0;
}
else
{
builtGramSize = gramSize.PreviousValue;
}
if (inputWindow.Count >= gramSize.Value)
{
bool isAllFiller = true;
InputWindowToken nextToken = null;
IEnumerator <InputWindowToken> iter = inputWindow.GetEnumerator();
for (int gramNum = 1; iter.MoveNext() && builtGramSize < gramSize.Value; ++gramNum)
{
nextToken = iter.Current;
if (builtGramSize < gramNum)
{
if (builtGramSize > 0)
{
gramBuilder.Append(tokenSeparator);
}
gramBuilder.Append(nextToken.termAtt.Buffer, 0, nextToken.termAtt.Length);
++builtGramSize;
}
if (isAllFiller && nextToken.isFiller)
{
if (gramNum == gramSize.Value)
{
gramSize.Advance();
}
}
else
{
isAllFiller = false;
}
}
if (!isAllFiller && builtGramSize == gramSize.Value)
{
inputWindow.First.Value.attSource.CopyTo(this);
posIncrAtt.PositionIncrement = isOutputHere ? 0 : 1;
termAtt.SetEmpty().Append(gramBuilder);
if (gramSize.Value > 1)
{
typeAtt.Type = tokenType;
noShingleOutput = false;
}
offsetAtt.SetOffset(offsetAtt.StartOffset, nextToken.offsetAtt.EndOffset);
posLenAtt.PositionLength = builtGramSize;
isOutputHere = true;
gramSize.Advance();
tokenAvailable = true;
}
}
return(tokenAvailable);
}
public override bool IncrementToken()
{
bool tokenAvailable = false;
int builtGramSize = 0;
if (gramSize.atMinValue() || inputWindow.Count < gramSize.Value)
{
shiftInputWindow();
gramBuilder.Length = 0;
}
else
{
builtGramSize = gramSize.PreviousValue;
}
if (inputWindow.Count >= gramSize.Value)
{
bool isAllFiller = true;
InputWindowToken nextToken = null;
IEnumerator <InputWindowToken> iter = inputWindow.GetEnumerator();
//JAVA TO C# CONVERTER TODO TASK: Java iterators are only converted within the context of 'while' and 'for' loops:
for (int gramNum = 1; iter.hasNext() && builtGramSize < gramSize.Value; ++gramNum)
{
nextToken = iter.Current;
if (builtGramSize < gramNum)
{
if (builtGramSize > 0)
{
gramBuilder.Append(tokenSeparator);
}
gramBuilder.Append(nextToken.termAtt.buffer(), 0, nextToken.termAtt.length());
++builtGramSize;
}
if (isAllFiller && nextToken.isFiller)
{
if (gramNum == gramSize.Value)
{
gramSize.advance();
}
}
else
{
isAllFiller = false;
}
}
if (!isAllFiller && builtGramSize == gramSize.Value)
{
inputWindow.First.Value.attSource.copyTo(this);
posIncrAtt.PositionIncrement = isOutputHere ? 0 : 1;
termAtt.setEmpty().append(gramBuilder);
if (gramSize.Value > 1)
{
typeAtt.Type = tokenType;
noShingleOutput = false;
}
offsetAtt.setOffset(offsetAtt.startOffset(), nextToken.offsetAtt.endOffset());
posLenAtt.PositionLength = builtGramSize;
isOutputHere = true;
gramSize.advance();
tokenAvailable = true;
}
}
return(tokenAvailable);
}
/// <summary>
/// <para>Get the next token from the input stream.
/// </para>
/// <para>If the next token has <code>positionIncrement > 1</code>,
/// <code>positionIncrement - 1</code> <seealso cref="#fillerToken"/>s are
/// inserted first.
/// </para>
/// </summary>
/// <param name="target"> Where to put the new token; if null, a new instance is created. </param>
/// <returns> On success, the populated token; null otherwise </returns>
/// <exception cref="IOException"> if the input stream has a problem </exception>
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: private InputWindowToken getNextToken(InputWindowToken target) throws java.io.IOException
private InputWindowToken getNextToken(InputWindowToken target)
{
InputWindowToken newTarget = target;
if (numFillerTokensToInsert > 0)
{
if (null == target)
{
newTarget = new InputWindowToken(this, nextInputStreamToken.cloneAttributes());
}
else
{
nextInputStreamToken.copyTo(target.attSource);
}
// A filler token occupies no space
newTarget.offsetAtt.setOffset(newTarget.offsetAtt.startOffset(), newTarget.offsetAtt.startOffset());
newTarget.termAtt.copyBuffer(fillerToken, 0, fillerToken.Length);
newTarget.isFiller = true;
--numFillerTokensToInsert;
}
else if (isNextInputStreamToken)
{
if (null == target)
{
newTarget = new InputWindowToken(this, nextInputStreamToken.cloneAttributes());
}
else
{
nextInputStreamToken.copyTo(target.attSource);
}
isNextInputStreamToken = false;
newTarget.isFiller = false;
}
else if (!exhausted)
{
if (input.incrementToken())
{
if (null == target)
{
newTarget = new InputWindowToken(this, cloneAttributes());
}
else
{
this.copyTo(target.attSource);
}
if (posIncrAtt.PositionIncrement > 1)
{
// Each output shingle must contain at least one input token,
// so no more than (maxShingleSize - 1) filler tokens will be inserted.
numFillerTokensToInsert = Math.Min(posIncrAtt.PositionIncrement - 1, maxShingleSize - 1);
// Save the current token as the next input stream token
if (null == nextInputStreamToken)
{
nextInputStreamToken = cloneAttributes();
}
else
{
this.copyTo(nextInputStreamToken);
}
isNextInputStreamToken = true;
// A filler token occupies no space
newTarget.offsetAtt.setOffset(offsetAtt.startOffset(), offsetAtt.startOffset());
newTarget.termAtt.copyBuffer(fillerToken, 0, fillerToken.Length);
newTarget.isFiller = true;
--numFillerTokensToInsert;
}
else
{
newTarget.isFiller = false;
}
}
else
{
exhausted = true;
input.end();
endState = captureState();
numFillerTokensToInsert = Math.Min(posIncrAtt.PositionIncrement, maxShingleSize - 1);
if (numFillerTokensToInsert > 0)
{
nextInputStreamToken = new AttributeSource(AttributeFactory);
nextInputStreamToken.addAttribute(typeof(CharTermAttribute));
OffsetAttribute newOffsetAtt = nextInputStreamToken.addAttribute(typeof(OffsetAttribute));
newOffsetAtt.setOffset(offsetAtt.endOffset(), offsetAtt.endOffset());
// Recurse/loop just once:
return getNextToken(target);
}
else
{
newTarget = null;
}
}
}
else
{
newTarget = null;
}
return newTarget;
}
/// <summary>
/// <para>Get the next token from the input stream.
/// </para>
/// <para>If the next token has <code>positionIncrement > 1</code>,
/// <code>positionIncrement - 1</code> <seealso cref="#fillerToken"/>s are
/// inserted first.
/// </para>
/// </summary>
/// <param name="target"> Where to put the new token; if null, a new instance is created. </param>
/// <returns> On success, the populated token; null otherwise </returns>
/// <exception cref="IOException"> if the input stream has a problem </exception>
private InputWindowToken GetNextToken(InputWindowToken target)
{
InputWindowToken newTarget = target;
if (numFillerTokensToInsert > 0)
{
if (null == target)
{
newTarget = new InputWindowToken(this, nextInputStreamToken.CloneAttributes());
}
else
{
nextInputStreamToken.CopyTo(target.attSource);
}
// A filler token occupies no space
newTarget.offsetAtt.SetOffset(newTarget.offsetAtt.StartOffset(), newTarget.offsetAtt.StartOffset());
newTarget.termAtt.CopyBuffer(fillerToken, 0, fillerToken.Length);
newTarget.isFiller = true;
--numFillerTokensToInsert;
}
else if (isNextInputStreamToken)
{
if (null == target)
{
newTarget = new InputWindowToken(this, nextInputStreamToken.CloneAttributes());
}
else
{
nextInputStreamToken.CopyTo(target.attSource);
}
isNextInputStreamToken = false;
newTarget.isFiller = false;
}
else if (!exhausted)
{
if (input.IncrementToken())
{
if (null == target)
{
newTarget = new InputWindowToken(this, CloneAttributes());
}
else
{
this.CopyTo(target.attSource);
}
if (posIncrAtt.PositionIncrement > 1)
{
// Each output shingle must contain at least one input token,
// so no more than (maxShingleSize - 1) filler tokens will be inserted.
numFillerTokensToInsert = Math.Min(posIncrAtt.PositionIncrement - 1, maxShingleSize - 1);
// Save the current token as the next input stream token
if (null == nextInputStreamToken)
{
nextInputStreamToken = CloneAttributes();
}
else
{
this.CopyTo(nextInputStreamToken);
}
isNextInputStreamToken = true;
// A filler token occupies no space
newTarget.offsetAtt.SetOffset(offsetAtt.StartOffset(), offsetAtt.StartOffset());
newTarget.termAtt.CopyBuffer(fillerToken, 0, fillerToken.Length);
newTarget.isFiller = true;
--numFillerTokensToInsert;
}
else
{
newTarget.isFiller = false;
}
}
else
{
exhausted = true;
input.End();
endState = CaptureState();
numFillerTokensToInsert = Math.Min(posIncrAtt.PositionIncrement, maxShingleSize - 1);
if (numFillerTokensToInsert > 0)
{
// LUCENENET TODO: Property attributeFactory should begin with uppre case character
nextInputStreamToken = new AttributeSource(this.attributeFactory);
nextInputStreamToken.AddAttribute<ICharTermAttribute>();
IOffsetAttribute newOffsetAtt = nextInputStreamToken.AddAttribute<IOffsetAttribute>();
newOffsetAtt.SetOffset(offsetAtt.EndOffset(), offsetAtt.EndOffset());
// Recurse/loop just once:
return GetNextToken(target);
}
else
{
newTarget = null;
}
}
}
else
{
newTarget = null;
}
return newTarget;
}
