Coding4Fun - 集保罕用字轉換工具 .NET 飆速版

$sw = New-Object System.Diagnostics.Stopwatch
Get-Content .\test\B5U8.config
$sw.Start()
Start-Process java -ArgumentList "-jar FileConvertD.jar B5U8 .\test\B5U8.config `"`"" -Wait
$sw.Stop()
"B5U8 $($sw.ElapsedMilliseconds.ToString("n0"))ms"
Get-Content .\test\U8B5.config
$sw.Restart()
Start-Process java -ArgumentList "-jar FileConvertD.jar U8B5 .\test\U8B5.config" -Wait
$sw.Stop()
"U8B5 $($sw.ElapsedMilliseconds.ToString("n0"))ms"

        private static void GenClass()
        {
            Dictionary<string, string> map = new Dictionary<string, string>();
            List<string> undefChars = new List<string>();

            foreach (var line in File.ReadAllLines("Map_code_2.txt", Encoding.UTF8).Skip(4))
            {
                string convChar = line.Substring(0, line.IndexOf(" "));
                var m = Regex.Match(line.Substring(1, 16), "  0(?<u>[0-9A-F]{4})");
                if (m.Success)
                {
                    map.Add($@"\u{m.Groups["u"].Value}", convChar);
                }
                else undefChars.Add(convChar);
            }
            //注音輕聲符號
            map.Add("˙", "․");
            //額外修正
            map["\uE506"] = "熴";
            map["\uF390"] = "廍";
            map["\uF393"] = "枬";
            var sb = new StringBuilder();
            sb.AppendLine(@"
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;

public class TdccEudcConverter 
{
");
            sb.AppendLine("\tstatic Dictionary<char, string> Eudc2Unicode = new Dictionary<char, string>()");
            sb.AppendLine("\t{");
            sb.AppendLine(
                string.Join(
                    ",\r\n",
                    map.OrderBy(o => o.Key).Select(o => $@"     ['{o.Key}'] = ""{o.Value}""").ToArray())
            );
            sb.AppendLine("\t};");
            sb.AppendLine("\tstatic HashSet<string> UndefChars = new HashSet<string>()");
            sb.AppendLine("\t{");
            sb.AppendLine(string.Join(
                    ",\r\n",
                    undefChars.Select(o => $@"     ""{o}""").ToArray())
            );
            sb.AppendLine("\t};");
            sb.AppendLine("\tstatic string NoConvChars = \"賚祈禛熲...略....\";");

            sb.AppendLine(@"
    static Dictionary<char, string> SngCharMapping = new Dictionary<char, string>();
    static Dictionary<char, Dictionary<char, string>> DblCharMapping = new Dictionary<char, Dictionary<char, string>>();

    static TdccEudcConverter()
    {
        var big5 = Encoding.GetEncoding(950);
        Func<string, string> convB5Bytes = (c) =>
            BitConverter.ToString(big5.GetBytes(c));
        var dict = new Dictionary<string, string>();
        foreach (var kv in Eudc2Unicode)
        {
            if (dict.ContainsKey(kv.Value) || NoConvChars.Contains(kv.Value))
            {
                continue;
            }
            dict.Add(kv.Value, kv.Key.ToString());
        }
        foreach (var uc in UndefChars)
        {

            if (!dict.ContainsKey(uc)) dict.Add(uc, ""■"");
        }
        foreach (var kv in dict)
        {
            if (kv.Key.Length == 1)
                SngCharMapping.Add(kv.Key[0], kv.Value);
            else if (kv.Key.Length == 2)
            {
                var c1 = kv.Key[0];
        var c2 = kv.Key[1];
                if (!DblCharMapping.ContainsKey(c1))
                    DblCharMapping.Add(c1, new Dictionary<char, string>());
                DblCharMapping[c1].Add(c2, kv.Value);
    }
            else
                throw new NotSupportedException();
}
//特例規則： F3 90 對映到 廍，但反向則是 BF DB B4 DF 對映 F3 90
DblCharMapping['\udbbf']['\udfb4'] = ""\uf390"";
    }

    public static string ConvEudcToUnicode(string src)
{
    var sb = new StringBuilder();
    foreach (char c in src)
    {
        if (Eudc2Unicode.ContainsKey(c))
            sb.Append(Eudc2Unicode[c]);
        else
            sb.Append(c);
    }
    return sb.ToString();
}

public static string ConvUnicodeToEudc(string src)
{
    var sb = new StringBuilder();
    var i = 0;
    while (i < src.Length)
    {
        var c = src[i];
        if (SngCharMapping.ContainsKey(c))
            sb.Append(SngCharMapping[c]);
        else if (DblCharMapping.ContainsKey(c))
        {
            if (i + 1 < src.Length && DblCharMapping[c].ContainsKey(src[i + 1]))
            {
                sb.Append(DblCharMapping[c][src[i + 1]]);
                i++;
            }
            else
                sb.Append(c);
        }
        else
            sb.Append(c);
        i++;
    }
    return sb.ToString();
}
");
            sb.AppendLine("}");
            File.WriteAllText("TdccEudcConverter.cs", sb.ToString(), Encoding.UTF8);
        }