public void TestElision2()
{
String test = "Plop, juste pour voir l'embrouille avec O'brian. M'enfin.";
Tokenizer tokenizer = new StandardTokenizer(Version.LUCENE_CURRENT, new StringReader(test));
HashSet<String> articles = new HashSet<String>();
articles.Add("l");
articles.Add("M");
TokenFilter filter = new ElisionFilter(tokenizer, articles);
List<string> tas = Filtre(filter);
Assert.AreEqual("embrouille", tas[4]);
Assert.AreEqual("O'brian", tas[6]);
Assert.AreEqual("enfin", tas[7]);
}
/// <summary>
/// Creates
/// <see cref="TokenStreamComponents"/>
/// used to tokenize all the text in the provided <see cref="TextReader"/>.
/// </summary>
/// <returns> <see cref="TokenStreamComponents"/>
/// built from a <see cref="StandardTokenizer"/> filtered with
/// <see cref="StandardFilter"/>, <see cref="ElisionFilter"/>,
/// <see cref="LowerCaseFilter"/>, <see cref="StopFilter"/>,
/// <see cref="SetKeywordMarkerFilter"/> if a stem exclusion set is
/// provided, and <see cref="FrenchLightStemFilter"/> </returns>
///
protected override TokenStreamComponents CreateComponents(string fieldName, TextReader reader)
{
#pragma warning disable 612, 618
if (m_matchVersion.OnOrAfter(LuceneVersion.LUCENE_31))
#pragma warning restore 612, 618
{
Tokenizer source = new StandardTokenizer(m_matchVersion, reader);
TokenStream result = new StandardFilter(m_matchVersion, source);
result = new ElisionFilter(result, DEFAULT_ARTICLES);
result = new LowerCaseFilter(m_matchVersion, result);
result = new StopFilter(m_matchVersion, result, m_stopwords);
if (excltable.Count > 0)
{
result = new SetKeywordMarkerFilter(result, excltable);
}
#pragma warning disable 612, 618
if (m_matchVersion.OnOrAfter(LuceneVersion.LUCENE_36))
#pragma warning restore 612, 618
{
result = new FrenchLightStemFilter(result);
}
else
{
result = new SnowballFilter(result, new Tartarus.Snowball.Ext.FrenchStemmer());
}
return(new TokenStreamComponents(source, result));
}
else
{
Tokenizer source = new StandardTokenizer(m_matchVersion, reader);
TokenStream result = new StandardFilter(m_matchVersion, source);
result = new StopFilter(m_matchVersion, result, m_stopwords);
if (excltable.Count > 0)
{
result = new SetKeywordMarkerFilter(result, excltable);
}
#pragma warning disable 612, 618
result = new FrenchStemFilter(result);
#pragma warning restore 612, 618
// Convert to lowercase after stemming!
return(new TokenStreamComponents(source, new LowerCaseFilter(m_matchVersion, result)));
}
}
