public CipherDictionary MostLikely()
{
var ret = new CipherDictionary(this.Count);
foreach (CipherPossibilities poss in this)
{
CipherPair ml = poss.MostLikely();
if (ml != null)
{
ret.Add(ml);
}
}
return(ret);
}
public static ulong[] CalcPairFreqs(int[] data, int[] codelist, CipherPair eof = null)
{
//convert the data into a string, so we can use regex on it
char[] ctext = new char[data.Length * 4];
Buffer.BlockCopy(data, 0, ctext, 0, data.Length * 4);
string text = new string(ctext);
var count = new SortedList <int, ulong>(codelist.Length * codelist.Length, new DuplicateKeyComparer <int>());
for (int i = 0; i < codelist.Length; i++)
{
for (int j = 0; j < codelist.Length; j++)
{
count.Add(Regex.Matches(text, string.Concat(codelist[i], codelist[j])).Count, ((ulong)i << 32) | (uint)j);
}
}
return(count.Select(p => p.Value).ToArray());
}
public static int[] CalcSingleFreqs(int[] data, CipherPair eof = null)
{
Dictionary <int, int> freqs = new Dictionary <int, int>();
foreach (int point in data)
{
if (eof != null && point == eof.Code)
{
break;
}
int freq;
if (freqs.TryGetValue(point, out freq))
{
freqs[point]++;
}
else
{
freqs.Add(point, 1);
}
}
return(freqs.OrderByDescending(pair => pair.Value).Select(pair => pair.Key).ToArray());
}
public void BestGuess()
{
CipherPair eof = null;
if (data[data.Length - 1] == data[data.Length - 2] && data[data.Length - 1] == data[data.Length - 3])
{
eof = new CipherPair(data[data.Length - 1], (char)26);
}
var codelist = data.Distinct();
var dicts = new List <CipherPredictionDictionary>();
dicts.Add(CipherPredictionDictionary.SingleFrequencyPrediction(data, eof));
int spaceCode = dicts[0].Find(p => p.MostLikely().Value == ' ').MostLikely().Code;
dicts.Add(CipherPredictionDictionary.PairFrequencyPrediction(data, codelist, eof));
dicts.Add(CipherPredictionDictionary.FirstLetterPrediction(data, spaceCode));
dicts.Add(CipherPredictionDictionary.KnownWordPrediction(data, "<untranslatable>"));
preDict = CipherPredictionDictionary.AggregateProbabilities(dicts, codelist);
dict = preDict.MostLikely();
dict.AddMissingValues(codelist);
charsetView.DataSource = dict.BindingList;
}
/// <summary>
/// Initializes a new instance of <see cref="CipherPossibilities"/> with a code and one possibility specified by a <see cref="CipherPair"/>
/// </summary>
public CipherPossibilities(CipherPair pair, double probability) : this(pair.Code)
{
Probabilites.Add(pair.Value, probability);
}
public static CipherPredictionDictionary SingleFrequencyPrediction(int[] data, CipherPair eof = null)
{
const double PROBABILITY_COEFFICIENT = 0.2;
const int PRED_DISTANCE = 4;
int[] freqs = CalcSingleFreqs(data, eof);
var dict = new CipherPredictionDictionary(freqs.Length + 1);
for (int i = 0; i < freqs.Length; i++)
{
CipherPossibilities poss = dict.Add(freqs[i]);
//loop over indexes from (i - PRED_DISTANCE) to (i + PRED_DISTANCE), within the bounds of the array
for (int j = Math.Max(i - PRED_DISTANCE, 0); j < Math.Min(i + PRED_DISTANCE + 1, singleFreqOrder.Length); j++)
{
poss.Add(singleFreqOrder[j], (PRED_DISTANCE - Math.Abs(i - j)) / (double)PRED_DISTANCE * PROBABILITY_COEFFICIENT);
}
}
dict.Insert(dict.Count, new CipherPossibilities(eof, 1.0));
return(dict);
}
public static CipherPredictionDictionary PairFrequencyPrediction(int[] data, IEnumerable <int> codelist, CipherPair eof = null)
{
const double PROBABILITY_COEFFICIENT = 0.3;
const int PRED_DISTANCE = 6;
ulong[] freqs = CalcPairFreqs(data, codelist.ToArray(), eof);
var dict = new CipherPredictionDictionary(freqs.Length + 1);
for (int i = 0; i < freqs.Length; i++)
{
CipherPossibilities poss1 = dict.Add((int)(freqs[i] >> 32)); //first char
CipherPossibilities poss2 = dict.Add((int)(freqs[i] & uint.MaxValue)); //second char
//loop over indexes from (i - PRED_DISTANCE) to (i + PRED_DISTANCE), within the bounds of the array
for (int j = Math.Max(i - PRED_DISTANCE, 0); j < Math.Min(i + PRED_DISTANCE + 1, pairFreqOrder.Length); j++)
{
double prob = (PRED_DISTANCE - Math.Abs(i - j)) / (double)PRED_DISTANCE * PROBABILITY_COEFFICIENT;
poss1.Assimilate(pairFreqOrder[j][0], prob);
poss2.Assimilate(pairFreqOrder[j][1], prob);
}
}
return(dict);
}
public bool EncodingEquals(CipherPair other)
{
return(other != null && other.Code == this.Code && other.Value == this.Value);
}