bludgeon/Token.cs

   1
   2 using System;
   3 using System.Text;
   4
   5 namespace Bludgeon {
   6
   7         public class Token {
   8
   9                 public const int Count = 512;
  10
  11                 static public string IdToString (int id)
  12                 {
  13                         if (id < 0 || id >= Count)
  14                                 throw new ArgumentException ();
  15                         return token_table [id];
  16                 }
  17
  18                 static public int StringToId (string str)
  19                 {
  20                         int i;
  21                         i = Array.BinarySearch (token_table, str);
  22                         if (i < 0 || i >= Count)
  23                                 return -1;
  24                         return i;
  25                 }
  26
  27                 private static Random random = new Random ();
  28
  29                 static public string GetRandom ()
  30                 {
  31                         return token_table [random.Next (Count)];
  32                 }
  33
  34                 ///////////////////////////////////////////////////////////////////////
  35
  36                 static private string [] token_table;
  37
  38                 static Token ()
  39                 {
  40                         token_table = new string [Count];
  41
  42                         for (int i = 0; i < Count; ++i)
  43                                 token_table [i] = TokenFromSeed (i);
  44                         Array.Sort (token_table);
  45
  46                         // Paranoia is healthy.
  47
  48                         for (int i = 1; i < Count; ++i)
  49                                 if (token_table [i-1] == token_table [i])
  50                                         throw new Exception ("Duplicate tokens!");
  51                 }
  52
  53                 // This is a silly algorithm, but it is an easy way to
  54                 // reproducibly generate a bunch of strings that all stem to
  55                 // distinct values.
  56
  57                 static private char [] buffer = new char [8];
  58                 static private string TokenFromSeed (int seed)
  59                 {
  60                         const int first_char = 97; // lower case 'a'
  61                         const uint p = 23;
  62                         const uint q = 7;
  63
  64                         uint state = (uint) seed;
  65
  66                         for (int i = 0; i < buffer.Length; ++i) {
  67
  68                                 // Put 'z' in the last two characters,
  69                                 // to avoid unhappy accidents of stemming.
  70                                 if (i >= buffer.Length - 2) {
  71                                         buffer [i] = 'z';
  72                                         continue;
  73                                 }
  74
  75                                 buffer [i] = Convert.ToChar (first_char + (state % p));
  76                                 state *= state + q;
  77                         }
  78
  79                         return new string (buffer);
  80                 }
  81
  82         }
  83
  84 }