/// <summary>
/// Crteate a Boolean decision tree.
/// References to solver and domain are not saved in the resulting decision tree.
/// </summary>
/// <param name="solver">character alberbra</param>
/// <param name="domain">elements that map to true</param>
/// <param name="precomputeLimit">upper limit for block ids for characters to be precomputed in an array (default is 0xFF, i.e. extended ASCII)</param>
/// <returns></returns>
internal static BooleanDecisionTree Create(CharSetSolver solver, BDD domain, ushort precomputeLimit = 0xFF)
{
BDD domain_compl = solver.MkNot(domain);
var partition = new BDD[] { domain_compl, domain };
if (precomputeLimit == 0)
{
return(new BooleanDecisionTree(new bool[] { }, MkBST(new DecisionTree.PartitionCut(solver, partition), 0, 0xFFFF)));
}
bool[] precomp = Precompute(solver, domain, precomputeLimit);
DecisionTree.BST bst = null;
if (precomputeLimit < ushort.MaxValue)
{
bst = MkBST(new DecisionTree.PartitionCut(solver, partition), precomputeLimit + 1, ushort.MaxValue);
}
return(new BooleanDecisionTree(precomp, bst));
}
/// <summary>
/// Assumes that set is a union of some minterms (or empty).
/// If null then null is returned.
/// </summary>
public BV ConvertFromCharSet(BDD set)
{
if (set == null)
{
return(null);
}
var alg = set.algebra;
BV res = this.zero;
for (int i = 0; i < partition.Length; i++)
{
BDD bdd_i = partition[i].AsBDD(alg);
var conj = alg.MkAnd(bdd_i, set);
if (alg.IsSatisfiable(conj))
{
res = res | atoms[i];
}
}
return(res);
}
private static bool[] Precompute(CharSetSolver solver, BDD domain, int precomputeLimit)
{
bool[] precomp = new bool[precomputeLimit + 1];
Func <int, bool> F = i =>
{
var bdd = solver.MkCharConstraint((char)i);
if (solver.IsSatisfiable(solver.MkAnd(bdd, domain)))
{
return(true);
}
else
{
return(false);
}
};
for (int c = 0; c <= precomputeLimit; c++)
{
precomp[c] = F(c);
}
return(precomp);
}
internal static string ToRegexCharSet(BDD label, IUnicodeCategoryTheory <BDD> categorizer, CharSetSolver solver)
{
if (categorizer.CategoryCondition(8) == label)
{
return(@"\d");
}
if (solver.MkNot(categorizer.CategoryCondition(8)) == label)
{
return(@"\D");
}
if (categorizer.WordLetterCondition == label)
{
return(@"\w");
}
if (solver.MkNot(categorizer.WordLetterCondition) == label)
{
return(@"\W");
}
if (categorizer.WhiteSpaceCondition == label)
{
return(@"\s");
}
if (solver.MkNot(categorizer.WhiteSpaceCondition) == label)
{
return(@"\S");
}
for (int i = 0; i < categorizer.UnicodeCategoryStandardAbbreviations.Length; i++)
{
if (categorizer.CategoryCondition(i) == label)
{
return(@"\P{" + categorizer.UnicodeCategoryStandardAbbreviations[i] + "}");
}
}
var ranges = solver.ToRanges(label);
if (ranges.Length == 1 && ranges[0].Item1 == ranges[0].Item2)
{
return(StringUtility.Escape((char)ranges[0].Item1));
}
var res = new StringBuilder("[");
for (int i = 0; i < ranges.Length; i++)
{
var range = ranges[i];
if (range.Item1 == range.Item2)
{
res.Append(StringUtility.EscapeWithNumericSpace((char)range.Item1));
}
else if (range.Item1 == range.Item2 - 1)
{
res.Append(StringUtility.EscapeWithNumericSpace((char)range.Item1));
res.Append(StringUtility.EscapeWithNumericSpace((char)range.Item2));
}
else
{
res.Append(StringUtility.EscapeWithNumericSpace((char)range.Item1));
res.Append("-");
res.Append(StringUtility.EscapeWithNumericSpace((char)range.Item2));
}
}
res.Append("]");
return(res.ToString());
}
public IgnoreCaseTransformer(CharSetSolver charSetSolver)
{
this.solver = charSetSolver;
IgnoreCaseRel = charSetSolver.Deserialize(Microsoft.SRM.Generated.IgnoreCaseRelation.ignorecase);
domain = IgnoreCaseRel >> 16;
}
private static void WriteRangeFields(BitWidth encoding, StreamWriter sw, string field)
{
int bits = (int)encoding;
int maxChar = (1 << bits) - 1;
var catMap = new Dictionary <UnicodeCategory, Ranges>();
for (int c = 0; c < 30; c++)
{
catMap[(UnicodeCategory)c] = new Ranges();
}
Ranges whitespace = new Ranges();
Ranges wordcharacter = new Ranges();
for (int i = 0; i <= maxChar; i++)
{
char ch = (char)i;
if (char.IsWhiteSpace(ch))
{
whitespace.Add(i);
}
UnicodeCategory cat = char.GetUnicodeCategory(ch);
catMap[cat].Add(i);
int catCode = (int)cat;
//in .NET 3.5
if (bits == 7)
{
if (catCode == 0 || catCode == 1 || catCode == 2 || catCode == 3 || catCode == 4 || catCode == 5 || catCode == 8 || catCode == 18)
{
wordcharacter.Add(i);
}
}
}
//generate bdd reprs for each of the category ranges
BDD[] catBDDs = new BDD[30];
CharSetSolver bddb = new CharSetSolver(encoding);
for (int c = 0; c < 30; c++)
{
catBDDs[c] = bddb.MkBddForIntRanges(catMap[(UnicodeCategory)c].ranges);
}
BDD whitespaceBdd = bddb.MkBddForIntRanges(whitespace.ranges);
//in .NET 3.5 category 5 was NOT a word character
//union of categories 0,1,2,3,4,8,18
BDD wordCharBdd = bddb.MkOr(catBDDs[0],
bddb.MkOr(catBDDs[1],
bddb.MkOr(catBDDs[2],
bddb.MkOr(catBDDs[3],
bddb.MkOr(catBDDs[4],
bddb.MkOr(catBDDs[5],
bddb.MkOr(catBDDs[8], catBDDs[18])))))));
if (bits == 7)
{
sw.WriteLine(@"/// <summary>
/// Array of 30 UnicodeCategory ranges. Each entry is a pair of integers.
/// corresponding to the lower and upper bounds of the unicodes of the characters
/// that have the given UnicodeCategory code (between 0 and 29).
/// </summary>");
sw.WriteLine("public static int[][][] " + field + " = new int[][][]{");
foreach (UnicodeCategory c in catMap.Keys)
{
sw.WriteLine("//{0}({1}):", c, (int)c);
if (catMap[c].Count == 0)
{
sw.WriteLine("null,");
}
else
{
sw.WriteLine("new int[][]{");
foreach (int[] range in catMap[c].ranges)
{
sw.WriteLine(" new int[]{" + string.Format("{0},{1}", range[0], range[1]) + "},");
}
sw.WriteLine("},");
}
}
sw.WriteLine("};");
}
sw.WriteLine(@"/// <summary>
/// Compact BDD encodings of the categories.
/// </summary>");
sw.WriteLine("public static int[][] " + field + "Bdd = new int[][]{");
foreach (UnicodeCategory c in catMap.Keys)
{
sw.WriteLine("//{0}({1}):", c, (int)c);
BDD catBdd = catBDDs[(int)c];
if (catBdd == null || catBdd.IsEmpty)
{
sw.WriteLine("null, //false");
}
else if (catBdd.IsFull)
{
sw.WriteLine("new int[]{0,0}, //true");
}
else
{
sw.WriteLine("new int[]{");
foreach (var arc in bddb.SerializeCompact(catBdd))
{
sw.WriteLine("{0},", arc);
}
sw.WriteLine("},");
}
}
sw.WriteLine("};");
if (bits == 7)
{
sw.WriteLine(@"/// <summary>
/// Whitespace character ranges.
/// </summary>");
sw.WriteLine("public static int[][] " + field + "Whitespace = new int[][]{");
foreach (int[] range in whitespace.ranges)
{
sw.WriteLine(" new int[]{" + string.Format("{0},{1}", range[0], range[1]) + "},");
}
sw.WriteLine("};");
sw.WriteLine(@"/// <summary>
/// Word character ranges.
/// </summary>");
sw.WriteLine("public static int[][] " + field + "WordCharacter = new int[][]{");
foreach (int[] range in wordcharacter.ranges)
{
sw.WriteLine(" new int[]{" + string.Format("{0},{1}", range[0], range[1]) + "},");
}
sw.WriteLine("};");
}
sw.WriteLine(@"/// <summary>
/// Compact BDD encoding of the whitespace characters.
/// </summary>");
sw.WriteLine("public static int[] " + field + "WhitespaceBdd = new int[]{");
foreach (var arc in bddb.SerializeCompact(whitespaceBdd))
{
sw.WriteLine("{0},", arc);
}
sw.WriteLine("};");
sw.WriteLine(@"/// <summary>
/// Compact BDD encoding of word characters
/// </summary>");
sw.WriteLine("public static int[] " + field + "WordCharacterBdd = new int[]{");
foreach (var arc in bddb.SerializeCompact(wordCharBdd))
{
sw.WriteLine("{0},", arc);
}
sw.WriteLine("};");
}
Tuple <uint, uint>[] ToRanges1(BDD set)
{
Tuple <uint, uint>[] ranges;
if (!rangeCache.TryGetValue(set, out ranges))
{
int b = set.Ordinal;
uint mask = (uint)1 << b;
if (set.Zero.IsEmpty)
{
#region 0-case is empty
if (set.One.IsFull)
{
var range = new Tuple <uint, uint>(mask, (mask << 1) - 1);
ranges = new Tuple <uint, uint>[] { range };
}
else //1-case is neither full nor empty
{
var ranges1 = LiftRanges(b, (b - set.One.Ordinal) - 1, ToRanges1(set.One));
ranges = new Tuple <uint, uint> [ranges1.Length];
for (int i = 0; i < ranges1.Length; i++)
{
ranges[i] = new Tuple <uint, uint>(ranges1[i].Item1 | mask, ranges1[i].Item2 | mask);
}
}
#endregion
}
else if (set.Zero.IsFull)
{
#region 0-case is full
if (set.One.IsEmpty)
{
var range = new Tuple <uint, uint>(0, mask - 1);
ranges = new Tuple <uint, uint>[] { range };
}
else
{
var rangesR = LiftRanges(b, (b - set.One.Ordinal) - 1, ToRanges1(set.One));
var range = rangesR[0];
if (range.Item1 == 0)
{
ranges = new Tuple <uint, uint> [rangesR.Length];
ranges[0] = new Tuple <uint, uint>(0, range.Item2 | mask);
for (int i = 1; i < rangesR.Length; i++)
{
ranges[i] = new Tuple <uint, uint>(rangesR[i].Item1 | mask, rangesR[i].Item2 | mask);
}
}
else
{
ranges = new Tuple <uint, uint> [rangesR.Length + 1];
ranges[0] = new Tuple <uint, uint>(0, mask - 1);
for (int i = 0; i < rangesR.Length; i++)
{
ranges[i + 1] = new Tuple <uint, uint>(rangesR[i].Item1 | mask, rangesR[i].Item2 | mask);
}
}
}
#endregion
}
else
{
#region 0-case is neither full nor empty
var rangesL = LiftRanges(b, (b - set.Zero.Ordinal) - 1, ToRanges1(set.Zero));
var last = rangesL[rangesL.Length - 1];
if (set.One.IsEmpty)
{
ranges = rangesL;
}
else if (set.One.IsFull)
{
var ranges1 = new List <Tuple <uint, uint> >();
for (int i = 0; i < rangesL.Length - 1; i++)
{
ranges1.Add(rangesL[i]);
}
if (last.Item2 == (mask - 1))
{
ranges1.Add(new Tuple <uint, uint>(last.Item1, (mask << 1) - 1));
}
else
{
ranges1.Add(last);
ranges1.Add(new Tuple <uint, uint>(mask, (mask << 1) - 1));
}
ranges = ranges1.ToArray();
}
else //general case: neither 0-case, not 1-case is full or empty
{
var rangesR0 = ToRanges1(set.One);
var rangesR = LiftRanges(b, (b - set.One.Ordinal) - 1, rangesR0);
var first = rangesR[0];
if (last.Item2 == (mask - 1) && first.Item1 == 0) //merge together the last and first ranges
{
ranges = new Tuple <uint, uint> [rangesL.Length + rangesR.Length - 1];
for (int i = 0; i < rangesL.Length - 1; i++)
{
ranges[i] = rangesL[i];
}
ranges[rangesL.Length - 1] = new Tuple <uint, uint>(last.Item1, first.Item2 | mask);
for (int i = 1; i < rangesR.Length; i++)
{
ranges[rangesL.Length - 1 + i] = new Tuple <uint, uint>(rangesR[i].Item1 | mask, rangesR[i].Item2 | mask);
}
}
else
{
ranges = new Tuple <uint, uint> [rangesL.Length + rangesR.Length];
for (int i = 0; i < rangesL.Length; i++)
{
ranges[i] = rangesL[i];
}
for (int i = 0; i < rangesR.Length; i++)
{
ranges[rangesL.Length + i] = new Tuple <uint, uint>(rangesR[i].Item1 | mask, rangesR[i].Item2 | mask);
}
}
}
#endregion
}
rangeCache[set] = ranges;
}
return(ranges);
}
private static Dictionary <char, BDD> ComputeIgnoreCaseDistionary(CharSetSolver solver)
{
var ignoreCase = new Dictionary <char, BDD>();
for (uint i = 0; i <= 0xFFFF; i++)
{
char c = (char)i;
char cU = char.ToUpper(c); // (char.IsLetter(char.ToUpper(c)) ? char.ToUpper(c) : c);
char cL = char.ToLower(c); // (char.IsLetter(char.ToLower(c)) ? char.ToLower(c) : c);
if (c != cU || c != cL || cU != cL)
{
//make sure that the regex engine considers c as being equivalent to cU and cL, else ignore c
//in some cases c != cU but the regex engine does not consider the chacarters equivalent wrt the ignore-case option.
//These characters are:
//c=\xB5,cU=\u039C
//c=\u0131,cU=I
//c=\u017F,cU=S
//c=\u0345,cU=\u0399
//c=\u03C2,cU=\u03A3
//c=\u03D0,cU=\u0392
//c=\u03D1,cU=\u0398
//c=\u03D5,cU=\u03A6
//c=\u03D6,cU=\u03A0
//c=\u03F0,cU=\u039A
//c=\u03F1,cU=\u03A1
//c=\u03F5,cU=\u0395
//c=\u1E9B,cU=\u1E60
//c=\u1FBE,cU=\u0399
if (System.Text.RegularExpressions.Regex.IsMatch(cU.ToString() + cL.ToString(), "^(?i:" + StringUtility.Escape(c) + ")+$"))
{
BDD equiv = solver.False;
if (ignoreCase.ContainsKey(c))
{
equiv = equiv | ignoreCase[c];
}
if (ignoreCase.ContainsKey(cU))
{
equiv = equiv | ignoreCase[cU];
}
if (ignoreCase.ContainsKey(cL))
{
equiv = equiv | ignoreCase[cL];
}
equiv = equiv | solver.MkCharSetFromRange(c, c) | solver.MkCharSetFromRange(cU, cU) | solver.MkCharSetFromRange(cL, cL);
foreach (char d in solver.GenerateAllCharacters(equiv))
{
ignoreCase[d] = equiv;
}
}
//else
//{
// outp += "c=" + StringUtility.Escape(c) + "," + "cU=" + StringUtility.Escape(cU);
// Console.WriteLine("c=" + StringUtility.Escape(c) + "," + "cL=" + StringUtility.Escape(cL) + "," + "cU=" + StringUtility.Escape(cU));
//}
}
}
return(ignoreCase);
}
