protected override TokenStreamComponents CreateComponents(string fieldName, TextReader reader)
{
var tokenizer = new AdrivaTokenizer(this.LuceneVersion, reader);
var components = new TokenStreamComponents(tokenizer);
return(components);
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
if (!stopWordsPerField.TryGetValue(fieldName, out ISet <string> stopWords) || stopWords is null)
{
return(components);
}
var stopFilter = new StopFilter(matchVersion, components.TokenStream, new CharArraySet(matchVersion, stopWords, false));
return(new TokenStreamComponents(components.Tokenizer, stopFilter));
}
public void Main()
{
const string s = "関西国際空港";
Console.WriteLine($"対象の文字列:{s}");
using var reader = new StringReader(s);
Tokenizer tokenizer = new JapaneseTokenizer(reader, ReadDict(), false, JapaneseTokenizerMode.NORMAL);
var tokenStreamComponents = new TokenStreamComponents(tokenizer, tokenizer);
using var tokenStream = tokenStreamComponents.TokenStream;
// note:処理の実行前にResetを実行する必要がある
tokenStream.Reset();
while (tokenStream.IncrementToken())
{
Console.WriteLine("---");
Console.WriteLine(
$"ICharTermAttribute=>{tokenStream.GetAttribute<ICharTermAttribute>().ToString()}");
Console.WriteLine(
$"ITermToBytesRefAttribute#BytesRef=>{tokenStream.GetAttribute<ITermToBytesRefAttribute>().BytesRef}");
Console.WriteLine(
$"IOffsetAttribute#StartOffset=>{tokenStream.GetAttribute<IOffsetAttribute>().StartOffset}");
Console.WriteLine(
$"IOffsetAttribute#EndOffset=>{tokenStream.GetAttribute<IOffsetAttribute>().EndOffset}");
Console.WriteLine(
$"IPositionIncrementAttribute=>{tokenStream.GetAttribute<IPositionIncrementAttribute>().PositionIncrement}");
Console.WriteLine(
$"IPositionLengthAttribute=>{tokenStream.GetAttribute<IPositionLengthAttribute>().PositionLength}");
Console.WriteLine(
$"IBaseFormAttribute#GetBaseForm=>{tokenStream.GetAttribute<IBaseFormAttribute>().GetBaseForm()}");
Console.WriteLine(
$"IPartOfSpeechAttribute#GetPartOfSpeech=>{tokenStream.GetAttribute<IPartOfSpeechAttribute>().GetPartOfSpeech()}");
Console.WriteLine(
$"IReadingAttribute#GetReading=>{tokenStream.GetAttribute<IReadingAttribute>().GetReading()}");
Console.WriteLine(
$"IReadingAttribute#GetPronunciation=>{tokenStream.GetAttribute<IReadingAttribute>().GetPronunciation()}");
Console.WriteLine(
$"IInflectionAttribute#GetInflectionForm=>{tokenStream.GetAttribute<IInflectionAttribute>().GetInflectionForm()}");
Console.WriteLine(
$"IInflectionAttribute#GetInflectionType=>{tokenStream.GetAttribute<IInflectionAttribute>().GetInflectionType()}");
Console.WriteLine("---");
}
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
var stopWords = stopWordsPerField.ContainsKey(fieldName) ? stopWordsPerField[fieldName] : null;
if (stopWords == null)
{
return(components);
}
var stopFilter = new StopFilter(matchVersion, components.TokenStream, new CharArraySet(matchVersion, stopWords, false));
return(new TokenStreamComponents(components.Tokenizer, stopFilter));
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
HashSet <string> stopWords = stopWordsPerField[fieldName];
if (stopWords == null)
{
return(components);
}
StopFilter stopFilter = new StopFilter(matchVersion, components.TokenStream, new CharArraySet(matchVersion, stopWords, false));
return(new TokenStreamComponents(components.Tokenizer, stopFilter));
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
ShingleFilter filter = new ShingleFilter(components.TokenStream, minShingleSize, maxShingleSize);
filter.SetMinShingleSize(minShingleSize);
filter.SetMaxShingleSize(maxShingleSize);
filter.SetTokenSeparator(tokenSeparator);
filter.SetOutputUnigrams(outputUnigrams);
filter.SetOutputUnigramsIfNoShingles(outputUnigramsIfNoShingles);
filter.SetFillerToken(fillerToken);
return(new TokenStreamComponents(components.Tokenizer, filter));
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
ShingleFilter filter = new ShingleFilter(components.TokenStream, minShingleSize, maxShingleSize);
filter.MinShingleSize = minShingleSize;
filter.MaxShingleSize = maxShingleSize;
filter.TokenSeparator = tokenSeparator;
filter.OutputUnigrams = outputUnigrams;
filter.OutputUnigramsIfNoShingles = outputUnigramsIfNoShingles;
filter.FillerToken = fillerToken;
return(new TokenStreamComponents(components.Tokenizer, filter));
}
protected override TokenStreamComponents CreateComponents(string fieldName, TextReader reader)
{
TokenStreamComponents tokenStreamComponents = null;
Tokenizer tokenizer = new StandardTokenizer(matchVersion, reader);
TokenStream stream = new LowerCaseFilter(matchVersion, tokenizer);
stream = new CyrllicToLatinFilter(stream);
stream = new StopFilter(matchVersion, stream, StopFilter.MakeStopSet(matchVersion, STOP_WORDS));
stream = new SnowballFilter(stream, new SimpleSerbianStemmer());
stream = new ASCIIFoldingFilter(stream);
tokenStreamComponents = new TokenStreamComponents(tokenizer, stream);
return(tokenStreamComponents);
}
protected override TokenStreamComponents CreateComponents(string fieldName, TextReader reader)
{
TokenStreamComponents tokenStreamComponents = null;
Tokenizer tokenizer = new StandardTokenizer(matchVersion, reader);
TokenStream stream = new StandardFilter(matchVersion, tokenizer);
stream = new LowerCaseFilter(matchVersion, stream);
stream = new StopFilter(matchVersion, stream, StopAnalyzer.ENGLISH_STOP_WORDS_SET);
stream = new PorterStemFilter(stream);
stream = new SnowballFilter(stream, new EnglishStemmer());
tokenStreamComponents = new TokenStreamComponents(tokenizer, stream);
return(tokenStreamComponents);
}
public override void SetReusableComponents(Analyzer analyzer, string fieldName, TokenStreamComponents components)
{
SetStoredValue(analyzer, components);
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
if (fieldName.Equals("textgrams") && outerInstance.minPrefixChars > 0)
{
return new TokenStreamComponents(components.Tokenizer,
new EdgeNGramTokenFilter(
outerInstance.matchVersion,
components.TokenStream,
1,
outerInstance.minPrefixChars));
}
else
{
return components;
}
}
/// <summary>
/// Wraps / alters the given TokenStreamComponents, taken from the wrapped
/// Analyzer, to form new components. It is through this method that new
/// TokenFilters can be added by AnalyzerWrappers. By default, the given
/// components are returned.
/// </summary>
/// <param name="fieldName">
/// Name of the field which is to be analyzed </param>
/// <param name="components">
/// TokenStreamComponents taken from the wrapped Analyzer </param>
/// <returns> Wrapped / altered TokenStreamComponents. </returns>
protected virtual TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
return components;
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
return new TokenStreamComponents(components.Tokenizer, new LimitTokenCountFilter(components.TokenStream, maxTokenCount, consumeAllTokens));
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
return(new TokenStreamComponents(components.Tokenizer, new LimitTokenCountFilter(components.TokenStream, maxTokenCount, consumeAllTokens)));
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
ShingleFilter shingles = new ShingleFilter(components.TokenStream, 2, outerInstance.grams);
shingles.SetTokenSeparator(char.ToString((char)outerInstance.separator));
return(new TokenStreamComponents(components.Tokenizer, shingles));
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
if (fieldName.Equals("textgrams") && outerInstance.minPrefixChars > 0)
{
return(new TokenStreamComponents(components.Tokenizer,
new EdgeNGramTokenFilter(
outerInstance.matchVersion,
components.TokenStream,
1,
outerInstance.minPrefixChars)));
}
else
{
return(components);
}
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
var stopWords = stopWordsPerField[fieldName];
if (stopWords == null)
{
return components;
}
var stopFilter = new StopFilter(matchVersion, components.TokenStream, new CharArraySet(matchVersion, stopWords, false));
return new TokenStreamComponents(components.Tokenizer, stopFilter);
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
ShingleFilter filter = new ShingleFilter(components.TokenStream, minShingleSize, maxShingleSize);
filter.MinShingleSize = minShingleSize;
filter.MaxShingleSize = maxShingleSize;
filter.TokenSeparator = tokenSeparator;
filter.OutputUnigrams = outputUnigrams;
filter.OutputUnigramsIfNoShingles = outputUnigramsIfNoShingles;
filter.FillerToken = fillerToken;
return new TokenStreamComponents(components.Tokenizer, filter);
}
public override void SetReusableComponents(Analyzer analyzer, string fieldName, TokenStreamComponents components)
{
var componentsPerField = (IDictionary<string, TokenStreamComponents>)GetStoredValue(analyzer);
if (componentsPerField == null)
{
componentsPerField = new Dictionary<string, TokenStreamComponents>();
SetStoredValue(analyzer, componentsPerField);
}
componentsPerField[fieldName] = components;
}
protected override TokenStreamComponents WrapComponents(string fieldName, TokenStreamComponents components)
{
return components;
}
/// <summary> /// Stores the given TokenStreamComponents as the reusable components for the /// field with the give name. /// </summary> /// <param name="fieldName"> Name of the field whose TokenStreamComponents are being set </param> /// <param name="components"> TokenStreamComponents which are to be reused for the field </param> public abstract void SetReusableComponents(Analyzer analyzer, string fieldName, TokenStreamComponents components);