public AddSuffixFilter(TokenStream input, Dictionary<string, char[]> _suffixByTokenType)
: base(input)
{
termAtt = (TermAttribute)AddAttribute(typeof(TermAttribute));
typeAtt = (TypeAttribute)AddAttribute(typeof(TypeAttribute));
this.suffixByTokenType = _suffixByTokenType;
}
override public object Clone()
{
TypeAttribute impl = new TypeAttribute();
impl.type = type;
return(impl);
}
public HyphenationTokenFilter(TokenStream input, Hyphenator hyphenator)
: base(input)
{
_hyphenator = hyphenator;
_termAtt = (TermAttribute)AddAttribute(typeof(TermAttribute));
_typeAtt = (TypeAttribute)AddAttribute(typeof(TypeAttribute));
_ofsAtt = (OffsetAttribute)AddAttribute(typeof(OffsetAttribute));
}
public override bool Equals(object other)
{
if (other == this)
{
return(true);
}
if (other is TypeAttribute)
{
TypeAttribute o = (TypeAttribute)other;
return(this.type == null ? o.type == null : this.type.Equals(o.type));
}
return(false);
}
public virtual void TestTypeAttribute()
{
TypeAttribute att = new TypeAttribute();
Assert.AreEqual(TypeAttribute.DEFAULT_TYPE, att.Type);
att.Type = "hallo";
Assert.AreEqual("type=hallo", att.ToString());
TypeAttribute att2 = (TypeAttribute) AssertCloneIsEqual(att);
Assert.AreEqual("hallo", att2.Type);
att2 = (TypeAttribute) AssertCopyIsEqual(att);
Assert.AreEqual("hallo", att2.Type);
att.Clear();
Assert.AreEqual(TypeAttribute.DEFAULT_TYPE, att.Type);
}
public virtual void TestTypeAttribute()
{
TypeAttribute att = new TypeAttribute();
Assert.AreEqual(TypeAttribute.DEFAULT_TYPE, att.Type);
att.Type = "hallo";
Assert.AreEqual("type=hallo", att.ToString());
TypeAttribute att2 = (TypeAttribute)AssertCloneIsEqual(att);
Assert.AreEqual("hallo", att2.Type);
att2 = (TypeAttribute)AssertCopyIsEqual(att);
Assert.AreEqual("hallo", att2.Type);
att.Clear();
Assert.AreEqual(TypeAttribute.DEFAULT_TYPE, att.Type);
}
public PrefixAwareTokenFilter(TokenStream prefix, TokenStream suffix) : base(suffix)
{
Suffix = suffix;
Prefix = prefix;
_prefixExhausted = false;
// ReSharper disable DoNotCallOverridableMethodsInConstructor
_termAtt = (TermAttribute) AddAttribute(typeof (TermAttribute));
_posIncrAtt = (PositionIncrementAttribute) AddAttribute(typeof (PositionIncrementAttribute));
_payloadAtt = (PayloadAttribute) AddAttribute(typeof (PayloadAttribute));
_offsetAtt = (OffsetAttribute) AddAttribute(typeof (OffsetAttribute));
_typeAtt = (TypeAttribute) AddAttribute(typeof (TypeAttribute));
_flagsAtt = (FlagsAttribute) AddAttribute(typeof (FlagsAttribute));
// ReSharper restore DoNotCallOverridableMethodsInConstructor
_pTermAtt = (TermAttribute) prefix.AddAttribute(typeof (TermAttribute));
_pPosIncrAtt = (PositionIncrementAttribute) prefix.AddAttribute(typeof (PositionIncrementAttribute));
_pPayloadAtt = (PayloadAttribute) prefix.AddAttribute(typeof (PayloadAttribute));
_pOffsetAtt = (OffsetAttribute) prefix.AddAttribute(typeof (OffsetAttribute));
_pTypeAtt = (TypeAttribute) prefix.AddAttribute(typeof (TypeAttribute));
_pFlagsAtt = (FlagsAttribute) prefix.AddAttribute(typeof (FlagsAttribute));
}
/// <summary>
/// Creates a shingle filter with ad hoc parameter settings.
/// </summary>
/// <param name="input">stream from which to construct the matrix</param>
/// <param name="minimumShingleSize">minimum number of tokens in any shingle.</param>
/// <param name="maximumShingleSize">maximum number of tokens in any shingle.</param>
/// <param name="spacerCharacter">character to use between texts of the token parts in a shingle. null for none.</param>
/// <param name="ignoringSinglePrefixOrSuffixShingle">if true, shingles that only contains permutation of the first of the last column will not be produced as shingles. Useful when adding boundary marker tokens such as '^' and '$'.</param>
/// <param name="settingsCodec">codec used to read input token weight and matrix positioning.</param>
public ShingleMatrixFilter(TokenStream input, int minimumShingleSize, int maximumShingleSize, Char? spacerCharacter, bool ignoringSinglePrefixOrSuffixShingle, TokenSettingsCodec settingsCodec)
{
_input = input;
MinimumShingleSize = minimumShingleSize;
MaximumShingleSize = maximumShingleSize;
SpacerCharacter = spacerCharacter;
IsIgnoringSinglePrefixOrSuffixShingle = ignoringSinglePrefixOrSuffixShingle;
_settingsCodec = settingsCodec;
// ReSharper disable DoNotCallOverridableMethodsInConstructor
_termAtt = (TermAttribute) AddAttribute(typeof (TermAttribute));
_posIncrAtt = (PositionIncrementAttribute) AddAttribute(typeof (PositionIncrementAttribute));
_payloadAtt = (PayloadAttribute) AddAttribute(typeof (PayloadAttribute));
_offsetAtt = (OffsetAttribute) AddAttribute(typeof (OffsetAttribute));
_typeAtt = (TypeAttribute) AddAttribute(typeof (TypeAttribute));
_flagsAtt = (FlagsAttribute) AddAttribute(typeof (FlagsAttribute));
// ReSharper restore DoNotCallOverridableMethodsInConstructor
_inTermAtt = (TermAttribute) input.AddAttribute(typeof (TermAttribute));
_inPosIncrAtt = (PositionIncrementAttribute) input.AddAttribute(typeof (PositionIncrementAttribute));
_inPayloadAtt = (PayloadAttribute) input.AddAttribute(typeof (PayloadAttribute));
_inOffsetAtt = (OffsetAttribute) input.AddAttribute(typeof (OffsetAttribute));
_inTypeAtt = (TypeAttribute) input.AddAttribute(typeof (TypeAttribute));
_inFlagsAtt = (FlagsAttribute) input.AddAttribute(typeof (FlagsAttribute));
}
/// <summary>
/// Creates a shingle filter based on a user defined matrix.
///
/// The filter /will/ delete columns from the input matrix! You will not be able to reset the filter if you used this constructor.
/// todo: don't touch the matrix! use a bool, set the input stream to null or something, and keep track of where in the matrix we are at.
///
/// </summary>
/// <param name="matrix">the input based for creating shingles. Does not need to contain any information until ShingleMatrixFilter.Next(Token) is called the first time.</param>
/// <param name="minimumShingleSize">minimum number of tokens in any shingle.</param>
/// <param name="maximumShingleSize">maximum number of tokens in any shingle.</param>
/// <param name="spacerCharacter">character to use between texts of the token parts in a shingle. null for none.</param>
/// <param name="ignoringSinglePrefixOrSuffixShingle">if true, shingles that only contains permutation of the first of the last column will not be produced as shingles. Useful when adding boundary marker tokens such as '^' and '$'.</param>
/// <param name="settingsCodec">codec used to read input token weight and matrix positioning.</param>
public ShingleMatrixFilter(Matrix.Matrix matrix, int minimumShingleSize, int maximumShingleSize, Char spacerCharacter, bool ignoringSinglePrefixOrSuffixShingle, TokenSettingsCodec settingsCodec)
{
Matrix = matrix;
MinimumShingleSize = minimumShingleSize;
MaximumShingleSize = maximumShingleSize;
SpacerCharacter = spacerCharacter;
IsIgnoringSinglePrefixOrSuffixShingle = ignoringSinglePrefixOrSuffixShingle;
_settingsCodec = settingsCodec;
// ReSharper disable DoNotCallOverridableMethodsInConstructor
_termAtt = (TermAttribute) AddAttribute(typeof (TermAttribute));
_posIncrAtt = (PositionIncrementAttribute) AddAttribute(typeof (PositionIncrementAttribute));
_payloadAtt = (PayloadAttribute) AddAttribute(typeof (PayloadAttribute));
_offsetAtt = (OffsetAttribute) AddAttribute(typeof (OffsetAttribute));
_typeAtt = (TypeAttribute) AddAttribute(typeof (TypeAttribute));
_flagsAtt = (FlagsAttribute) AddAttribute(typeof (FlagsAttribute));
// ReSharper restore DoNotCallOverridableMethodsInConstructor
// set the input to be an empty token stream, we already have the data.
_input = new EmptyTokenStream();
_inTermAtt = (TermAttribute) _input.AddAttribute(typeof (TermAttribute));
_inPosIncrAtt = (PositionIncrementAttribute) _input.AddAttribute(typeof (PositionIncrementAttribute));
_inPayloadAtt = (PayloadAttribute) _input.AddAttribute(typeof (PayloadAttribute));
_inOffsetAtt = (OffsetAttribute) _input.AddAttribute(typeof (OffsetAttribute));
_inTypeAtt = (TypeAttribute) _input.AddAttribute(typeof (TypeAttribute));
_inFlagsAtt = (FlagsAttribute) _input.AddAttribute(typeof (FlagsAttribute));
}
public override void CopyTo(Attribute target)
{
TypeAttribute t = (TypeAttribute)target;
t.type = type;
}
private void Init(System.IO.TextReader input, HebMorph.StreamLemmatizer _lemmatizer,
HebMorph.LemmaFilters.LemmaFilterBase _lemmaFilter, bool AlwaysSaveMarkedOriginal)
{
termAtt = (TermAttribute)AddAttribute(typeof(TermAttribute));
offsetAtt = (OffsetAttribute)AddAttribute(typeof(OffsetAttribute));
posIncrAtt = (PositionIncrementAttribute)AddAttribute(typeof(PositionIncrementAttribute));
typeAtt = (TypeAttribute)AddAttribute(typeof(TypeAttribute));
//payAtt = (PayloadAttribute)AddAttribute(typeof(PayloadAttribute));
this._streamLemmatizer = _lemmatizer;
this._streamLemmatizer.SetStream(input);
this.alwaysSaveMarkedOriginal = AlwaysSaveMarkedOriginal;
this.lemmaFilter = _lemmaFilter;
}
override public System.Object Clone()
{
TypeAttribute impl = new TypeAttribute();
impl.type = type;
return impl;
}
public TokenListStream(ICollection<Token> tokens)
{
_tokens = tokens;
_termAtt = (TermAttribute) AddAttribute(typeof (TermAttribute));
_posIncrAtt = (PositionIncrementAttribute) AddAttribute(typeof (PositionIncrementAttribute));
_payloadAtt = (PayloadAttribute) AddAttribute(typeof (PayloadAttribute));
_offsetAtt = (OffsetAttribute) AddAttribute(typeof (OffsetAttribute));
_typeAtt = (TypeAttribute) AddAttribute(typeof (TypeAttribute));
_flagsAtt = (FlagsAttribute) AddAttribute(typeof (FlagsAttribute));
}
/// <summary>
/// Constructs a ShingleFilter with the specified single size from the TokenStream
/// </summary>
/// <param name="input">input token stream</param>
/// <param name="maxShingleSize">maximum shingle size produced by the filter.</param>
public ShingleFilter(TokenStream input, int maxShingleSize) : base(input)
{
SetMaxShingleSize(maxShingleSize);
// ReSharper disable DoNotCallOverridableMethodsInConstructor
_termAtt = (TermAttribute) AddAttribute(typeof (TermAttribute));
_offsetAtt = (OffsetAttribute) AddAttribute(typeof (OffsetAttribute));
_posIncrAtt = (PositionIncrementAttribute) AddAttribute(typeof (PositionIncrementAttribute));
_typeAtt = (TypeAttribute) AddAttribute(typeof (TypeAttribute));
// ReSharper restore DoNotCallOverridableMethodsInConstructor
}
public MetaphoneReplacementFilter(TokenStream input)
: base(input)
{
termAttr = AddAttribute(typeof (TermAttribute)) as TermAttribute;
typeAttr = AddAttribute(typeof (TypeAttribute)) as TypeAttribute;
}
public override void CopyTo(AttributeImpl target)
{
TypeAttribute t = (TypeAttribute)target;
t.SetType(type);
}
