public Normalizer(UnicodeDataRecord[] unicodeData, NormalizationProps normalizationProps)
{
// CCCテーブル (共通)
_canonicalCombiningClassTable = unicodeData.Where(x => x.CanonicalCombiningClass != 0)
.ToDictionary(x => x.CodePoint, x => x.CanonicalCombiningClass);
// 分解テーブル (Decomposer)
_decompositionTable = CreateDecompositionTable(unicodeData);
// 合成テーブル (Composer, OptimizedComposer)
_compositionTable = CreateCompositionTable(unicodeData, normalizationProps);
// クイックチェックテーブル (OptimizedComposer)
_compositionQuickCheckTable = CreateCompositionQuickCheckTable(unicodeData, normalizationProps);
}
static void Main(string[] args)
{
var norm = new Normalizer(
UnicodeData.Load("UnicodeData.txt"),
NormalizationProps.Load("DerivedNormalizationProps.txt")
);
var tests = NormalizationTest.Load("NormalizationTest.txt");
TestNFD(norm, tests);
TestNFKD(norm, tests);
TestNFC(norm, tests);
TestNFKC(norm, tests);
TestOptimizedNFC(norm, tests);
TestOptimizedNFKC(norm, tests);
BenchNFC(norm, tests);
BenchOptimizedNFC(norm, tests);
}
/// <summary>合成テーブルを作成します。</summary>
/// <param name="unicodeData">UnicodeData.txt の内容</param>
/// <param name="normalizationProps">DerivedNormalizationProps.txt の内容</param>
private static IReadOnlyDictionary <CodePointPair, uint> CreateCompositionTable(UnicodeDataRecord[] unicodeData, NormalizationProps normalizationProps)
{
var fullCompositionExclusion = new HashSet <uint>(normalizationProps.FullCompositionExclusion);
return(unicodeData.Where(x => x.DecompositionMapping != null &&
x.DecompositionMapping.Type == null && // 互換分解でない
x.DecompositionMapping.Mapping.Length == 2 && // 1文字の分解でない -> 2文字
!fullCompositionExclusion.Contains(x.CodePoint) // Full_Composition_Exclusion
).ToDictionary(
x => new CodePointPair(x.DecompositionMapping.Mapping[0], x.DecompositionMapping.Mapping[1]),
x => x.CodePoint
));
}
/// <summary>合成のクイックチェック用CCCテーブルを作成します。</summary>
/// <param name="unicodeData">UnicodeData.txt の内容</param>
/// <param name="normalizationProps">DerivedNormalizationProps.txt の内容</param>
/// <returns>Key: コードポイント, Value: CCC または、 NFC_QCがNかMなら255、NFKC_QCがNかMなら256</returns>
private static IReadOnlyDictionary <uint, int> CreateCompositionQuickCheckTable(UnicodeDataRecord[] unicodeData, NormalizationProps normalizationProps)
{
return(unicodeData.Where(x => x.CanonicalCombiningClass != 0).Where(x => !normalizationProps.NfcQC.ContainsKey(x.CodePoint)).Select(x => new { x.CodePoint, Ccc = x.CanonicalCombiningClass }) // CCC
.Concat(normalizationProps.NfcQC.Select(x => new { CodePoint = x.Key, Ccc = 255 })) // NFC_QC
.Concat(normalizationProps.NfkcQC.Where(x => !normalizationProps.NfcQC.ContainsKey(x.Key)).Select(x => new { CodePoint = x.Key, Ccc = 256 })) // NFKC_QC
.ToDictionary(x => x.CodePoint, x => x.Ccc));
}
