C#でPluralizeとSingularizeしてみた

http://d.hatena.ne.jp/nagakura_eil/20080713/p3 の続き
またRailsの実装をそのままC#で書き直したような感じです。
微妙に終わってないけど。。。

Inflections.csを作成

static class Inflections
{
	static List<string> _uncountable = new List<string>(new string[] { "equipment","information","rice","money","species","series","fish","sheep" });
	public static List<string> Uncountable {
		get { return Inflections._uncountable; }
		set { Inflections._uncountable = value; }
	}
	
	static List<string[]> _pluralPatterns = new List<string[]>();
	public static List<string[]> PluralPatterns {
		get { return Inflections._pluralPatterns; }
		set { Inflections._pluralPatterns = value; }
	}

	static List<string[]> _singlePatterns = new List<string[]>();
	public static List<string[]> SinglePatterns {
		get { return Inflections._singlePatterns; }
		set { Inflections._singlePatterns = value; }
	}

	static List<string[]> _irregularPattern = new List<string[]>();
	public static List<string[]> IrregularPattern {
		get { return Inflections._irregularPattern; }
		set { Inflections._irregularPattern = value; }
	}
	static Inflections() {
		if(_pluralPatterns.Count != 0) { return; }
		//イレギュラーパターン
		_irregularPattern.AddRange(new[] { 
			new []{"person","people"},
			new []{"man","men"},
			new []{"child","children"},
			new []{"sex","sexes"},
			new []{"move","moves"}
		
		});

		//複数型変化
		_pluralPatterns.AddRange(new[]{
			new[]{"$", "s"},
			new[]{"s$", "s"},
			new[]{"(ax|test)is$", "$1es"},
			new[]{"(octop|vir)us$", "$1i"},
			new[]{"(alias|status)$", "$1es"},
			new[]{"(bu)s$", "$1ses"},
			new[]{"(buffal|tomat)o$", "$1oes"},
			new[]{"([ti])um$", "$1a"},
			new[]{"sis$", "ses"},
			new[]{"(?:([^f])fe|([lr])f)$", "$1$2ves"},
			new[]{"(hive)$", "$1s"},
			new[]{"([^aeiouy]|qu)y$", "$1ies"},
			new[]{"(x|ch|ss|sh)$", "$1es"},
			new[]{"(matr|vert|ind)(?:ix|ex)$", "$1ices"},
			new[]{"([m|l])ouse$", "$1ice"},
			new[]{"^(ox)$", "$1en"},
			new[]{"(quiz)$", "$1zes"}});

		//単数形変化
		_singlePatterns.AddRange(new[] { 
			new[]{"s$", ""},
			new[]{"(n)ews$", "$1ews"},
			new[]{"([ti])a$", "$1um"},
			new[]{"((a)naly|(b)a|(d)iagno|(p)arenthe|(p)rogno|(s)ynop|(t)he)ses$", "$1$2sis"},
			new[]{"(^analy)ses$", "$1sis"},
			new[]{"([^f])ves$", "$1fe"},
			new[]{"(hive)s$", "$1"},
			new[]{"(tive)s$", "$1"},
			new[]{"([lr])ves$", "$1f"},
			new[]{"([^aeiouy]|qu)ies$", "$1y"},
			new[]{"(s)eries$", "$1eries"},
			new[]{"(m)ovies$", "$1ovie"},
			new[]{"(x|ch|ss|sh)es$", "$1"},
			new[]{"([m|l])ice$", "$1ouse"},
			new[]{"(bus)es$", "$1"},
			new[]{"(o)es$", "$1"},
			new[]{"(shoe)s$", "$1"},
			new[]{"(cris|ax|test)es$", "$1is"},
			new[]{"(octop|vir)i$", "$1us"},
			new[]{"(alias|status)es$", "$1"},
			new[]{"^(ox)en", "$1"},
			new[]{"(vert|ind)ices$", "$1ex"},
			new[]{"(matr)ices$", "$1ix"},
			new[]{"(quiz)zes$", "$1"}});

		_irregularPattern.ForEach(ss => {
			string s1 = ss[0];
			string s2 = ss[1];
			if(s1[0].ToString().ToUpper() == s2[0].ToString().ToUpper()) {
				_pluralPatterns.Add(new[] { "(" + s1[0] + ")" + s1.Substring(1) + "$","$1" + s2.Substring(1) });
				_singlePatterns.Add(new[] { "(" + s2[0] + ")" + s2.Substring(1) + "$","$1" + s1.Substring(1) });
			} else {
				//先頭文字が違うのには未対応、そのうちやる
			}
		});
		_pluralPatterns.Reverse();
		_singlePatterns.Reverse();
	}
}

で、実際に変換するメソッド

private string Pluralize(string word) {
	if(word == "" | Inflections.Uncountable.Contains(word)) { return word; }
	foreach(var ss in Inflections.PluralPatterns) {
		Regex r = new Regex(ss[0],RegexOptions.IgnoreCase);
		if(r.IsMatch(word)) {
			var m = r.Replace(word,ss[1]);
			return m;
		}
	}
	return word;
}

private string Singularize(string word) {
	if(Inflections.Uncountable.Contains(word)) { return word; }
	foreach(var ss in Inflections.SinglePatterns) {
		Regex r = new Regex(ss[0],RegexOptions.IgnoreCase);
		if(r.IsMatch(word)) {
			var m = r.Replace(word,ss[1]);
			return m;
		}
	}
	return word;
}

このままだとRegexのインスタンスを何度も生成しちゃうんで、キャッシュ取るとかしたほうがいいかも。
まぁ用途的に何度もループの中で呼び出されたりするようなことは少ないと思うので放置で。
あとC#での記述法がよくわからなかったので1文字目の文字が違う場合のイレギュラーパターンの実装がまだです。