public SnowballFilter(TokenStream input, SnowballProgram stemmer)
: base(input)
{
this.stemmer = stemmer;
this.termAtt = AddAttribute<ICharTermAttribute>();
this.keywordAttr = AddAttribute<IKeywordAttribute>();
}
public SnowballFilter(TokenStream input, SnowballProgram stemmer)
: base(input)
{
this.stemmer = stemmer;
this.termAtt = AddAttribute <ICharTermAttribute>();
this.keywordAttr = AddAttribute <IKeywordAttribute>();
}
/// <summary>
/// Construct the named stemming filter.
///
/// Available stemmers are listed in <seealso cref="org.tartarus.snowball.ext"/>.
/// The name of a stemmer is the part of the class name before "Stemmer",
/// e.g., the stemmer in <seealso cref="org.tartarus.snowball.ext.EnglishStemmer"/> is named "English".
/// </summary>
/// <param name="in"> the input tokens to stem </param>
/// <param name="name"> the name of a stemmer </param>
public SnowballFilter(TokenStream @in, string name) : base(@in)
{
//Class.forName is frowned upon in place of the ResourceLoader but in this case,
// the factory will use the other constructor so that the program is already loaded.
try
{
Type stemClass = Type.GetType("org.tartarus.snowball.ext." + name + "Stemmer").asSubclass(typeof(SnowballProgram));
stemmer = stemClass.newInstance();
}
catch (Exception e)
{
throw new System.ArgumentException("Invalid stemmer class specified: " + name, e);
}
}
/// <summary>
/// Construct the named stemming filter.
///
/// Available stemmers are listed in Lucene.Net.Tartarus.Snowball.Ext.
/// The name of a stemmer is the part of the class name before "Stemmer",
/// e.g., the stemmer in <see cref="Tartarus.Snowball.Ext.EnglishStemmer"/> is named "English".
/// </summary>
/// <param name="in"> the input tokens to stem </param>
/// <param name="name"> the name of a stemmer </param>
public SnowballFilter(TokenStream @in, string name)
: base(@in)
{
try
{
string className = typeof(SnowballProgram).Namespace + ".Ext." +
name + "Stemmer, " + this.GetType().GetTypeInfo().Assembly.GetName().Name;
Type stemClass = Type.GetType(className);
stemmer = (SnowballProgram)Activator.CreateInstance(stemClass);
}
catch (Exception e)
{
throw new System.ArgumentException("Invalid stemmer class specified: " + name, e);
}
this.termAtt = AddAttribute <ICharTermAttribute>();
this.keywordAttr = AddAttribute <IKeywordAttribute>();
}
/// <summary>
/// Construct the named stemming filter.
///
/// Available stemmers are listed in <seealso cref="org.tartarus.snowball.ext"/>.
/// The name of a stemmer is the part of the class name before "Stemmer",
/// e.g., the stemmer in <seealso cref="org.tartarus.snowball.ext.EnglishStemmer"/> is named "English".
/// </summary>
/// <param name="in"> the input tokens to stem </param>
/// <param name="name"> the name of a stemmer </param>
public SnowballFilter(TokenStream @in, string name)
: base(@in)
{
try
{
// LUCENENET TODO: There should probably be a way to make this an extesibility point so
// custom extensions can be loaded.
string className = typeof(SnowballProgram).Namespace + ".Ext." +
name + "Stemmer, " + this.GetType().Assembly.GetName().Name;
Type stemClass = Type.GetType(className);
stemmer = (SnowballProgram)Activator.CreateInstance(stemClass);
}
catch (Exception e)
{
throw new System.ArgumentException("Invalid stemmer class specified: " + name, e);
}
this.termAtt = AddAttribute <ICharTermAttribute>();
this.keywordAttr = AddAttribute <IKeywordAttribute>();
}
public override TokenStream TokenStream(string fieldName, TextReader reader)
{
if (!this._Language.Equals("JA", StringComparison.CurrentCultureIgnoreCase))
{//不是日语的需要重建Reader。wangyunpeng
reader = this.InitReader(reader);
}
TokenStream result = this._Analyzer.TokenStream(fieldName, reader);
result = new StopFilter(this._EnableStopPositionIncrements, result, this._StopCharArraySet, true);
SnowballProgram snowballProgram = SnowballDict.GetSnowball(this._Language);//词干。wangyunpeng，2015-8-17改成线程安全的调用方式。
if (snowballProgram != null)
{
result = new SnowballFilter(result, snowballProgram);
}
if (_UseIndexSynonyms)
{//在创建索引的时候，将同义词，近义词，相关词存入索引。
result = new SynonymsFilter(this._Language, result);
}
return(result);
}
/// <summary>
/// Construct the named stemming filter.
///
/// Available stemmers are listed in <seealso cref="org.tartarus.snowball.ext"/>.
/// The name of a stemmer is the part of the class name before "Stemmer",
/// e.g., the stemmer in <seealso cref="org.tartarus.snowball.ext.EnglishStemmer"/> is named "English".
/// </summary>
/// <param name="in"> the input tokens to stem </param>
/// <param name="name"> the name of a stemmer </param>
public SnowballFilter(TokenStream @in, string name)
: base(@in)
{
try
{
// LUCENENET TODO: There should probably be a way to make this an extesibility point so
// custom extensions can be loaded.
string className = typeof(SnowballProgram).Namespace + ".Ext." +
name + "Stemmer, " + this.GetType().Assembly.GetName().Name;
Type stemClass = Type.GetType(className);
stemmer = (SnowballProgram)Activator.CreateInstance(stemClass);
}
catch (Exception e)
{
throw new System.ArgumentException("Invalid stemmer class specified: " + name, e);
}
this.termAtt = AddAttribute<ICharTermAttribute>();
this.keywordAttr = AddAttribute<IKeywordAttribute>();
}
public Among(string s, int substring_i, int result,
string methodname, SnowballProgram methodobject)
{
this.s_size = s.Length;
this.s = s.ToCharArray();
this.substring_i = substring_i;
this.result = result;
this.methodobject = methodobject;
if (methodname.Length == 0)
{
this.method = null;
}
else
{
try
{
this.method = methodobject.GetType().GetMethod(methodname, EMPTY_PARAMS);
}
catch (MissingMethodException e)
{
throw new Exception(e.ToString(), e);
}
}
}
public SnowballFilter(TokenStream input, SnowballProgram stemmer) : base(input)
{
this.stemmer = stemmer;
}
/// <summary>
/// 获取词干对象(线程不安全)
/// </summary>
/// <param name="language"></param>
/// <returns></returns>
//public static SnowballProgram GetSnowball(string language)
//{
// if (_dictSnowball.ContainsKey(language))
// return _dictSnowball[language];
// return null;
//}
/// <summary>
/// 获取词干对象(线程安全)
/// </summary>
/// <param name="language"></param>
/// <returns></returns>
public static SnowballProgram GetSnowball(string language)
{
SnowballProgram result = null;
switch (language)
{
case "DA":
result = new DanishStemmer();
break; //丹麦语
case "NL":
result = new DutchStemmer();
break; //荷兰语
case "EN":
result = new EnglishStemmer();
break; //英语
case "FI":
result = new FinnishStemmer();
break; //芬兰语
case "FR":
result = new FrenchStemmer();
break; //法语
case "DE2":
result = new German2Stemmer();
break; //德语2
case "DE":
result = new GermanStemmer();
break; //德语
case "HU":
result = new HungarianStemmer();
break;
case "IT":
result = new ItalianStemmer();
break;
case "文斯语":
result = new LovinsStemmer();
break;
case "NO":
result = new NorwegianStemmer();
break;
case "波特语":
result = new PorterStemmer();
break; //英语的
case "PT":
result = new PortugueseStemmer();
break; //葡萄牙语
case "RO":
result = new RomanianStemmer();
break;
case "RU":
result = new RussianStemmer();
break; //俄语
case "ES":
result = new SpanishStemmer();
break; //西班牙语
case "SV":
result = new SwedishStemmer();
break;
case "TR":
result = new TurkishStemmer();
break; //土耳其语
}
return(result);
}
