private RegexNode mReservedPath; //The child node that must be chosen.
//If this is not null then the node must repeat at least once
public RegexRepeatNode(RegexNode refNode, int minRepeat, int maxRepeat, bool sameValue)
{
//if this does not cover zero to infinity, then this node can be invalidated
if (RegexCompiler.IsInvalidSection && (minRepeat > 0 || maxRepeat != -1))
{
RegexCompiler.InvalidableNodes.Add(this);
}
mMinRepeat = minRepeat;
mMaxRepeat = maxRepeat;
mSameValue = sameValue;
mReservedPath = null;
mRefNode = refNode;
mRefNode.Parent = this;
}
/// <summary>
/// Returns a random string, conforming to the provided regular expression pattern.
/// </summary>
/// <param name="regex">The regular expression, which the generated string should conform to.</param>
/// <param name="seed">The random number generator seed.</param>
/// <returns>A string, conforming to the provided regular expression pattern.</returns>
///
/// <example>
/// The following example demonstrates how to generate a random string from a Regex pattern:
/// <code lang="C#" >
/// // Using a custom regex pattern...
/// Regex emailAddress1 = new Regex(@"^([0-9a-zA-Z]([-.\w]*[0-9a-zA-Z])*@([0-9a-zA-Z][-\w]*[0-9a-zA-Z]\.)+[a-zA-Z]{2,9})$");
/// string s1 = StringFactory.GenerateRandomString(emailAddress1, 1234);
///
/// // Using a standard regex pattern...
/// Regex emailAddress2 = CommonRegexPatterns.EmailAddress;
/// string s2 = StringFactory.GenerateRandomString(emailAddress2, 1234);
/// </code>
/// </example>
public static string GenerateRandomString(Regex regex, int seed)
{
Random random = new Random(seed);
//reset the static variables
RegexCompiler.IsInvalidSection = false;
RegexCompiler.InvalidNode = null;
RegexCompiler.InvalidableNodes.Clear();
//construct the RegEx tree
RegexCompiler compiler = new RegexCompiler();
RegexNode node = compiler.Compile(regex.ToString());
//search for a signal to invalidate a node
if (regex.ToString().IndexOf("\\i") != -1)
{
//something should have been invalidated
//select a node to invalidate
if (RegexCompiler.InvalidableNodes.Count == 0)
{
throw new ArgumentException("Asked to generate invalid: Impossible to invalidate");
}
RegexCompiler.InvalidNode = RegexCompiler.InvalidableNodes[random.Next(RegexCompiler.InvalidableNodes.Count)];
//Mark REOrNodes and RERepeatNodes to ensure that the invalid node will be part of the string
RegexCompiler.InvalidNode.ReservePath(null);
}
//generate and return the string
string result = node.Generate(random);
if (RegexCompiler.InvalidNode != null)
{
//confirm that the generated string is invalid (e.g. [a-z]|[^a-z] will always fail)
Regex compare = new Regex("^" + regex.Replace("\\i", "") + "$");
if (compare.IsMatch(result))
{
throw new ArgumentException(regex + ": Did not generate invalid string: " + result);
}
}
return(result);
}
public override string Generate(Random random)
{
if (this == RegexCompiler.InvalidNode)
{
RegexNode.AssertParse(mNumChoices > 0, "No valid range specified in char set");
//select from the elements that are not available (elements that are invalid)
int randIndex = random.Next(mMapSize - mNumChoices);
int i = -1;
while (randIndex >= 0) //seek to the available element
{
i++;
//invert positive and negative sets
if ((mPositiveSet && mMap[i] == 0) || (!mPositiveSet && mMap[i] == 1))
{
randIndex--;
}
}
return(Convert.ToChar(i).ToString());
}
else
{
RegexNode.AssertParse(mNumChoices > 0, "No valid range specified in char set");
//select from the elements that are available
int randIndex = random.Next(mNumChoices);
int i = -1;
while (randIndex >= 0) //seek to the available element
{
i++;
if ((mPositiveSet && mMap[i] == 1) || (!mPositiveSet && mMap[i] == 0))
{
randIndex--;
}
}
return(Convert.ToChar(i).ToString());
}
}
//Compile node starting with |
public RegexNode CompileBranch()
{
RegexNode piece = CompilePiece();
if (mParseDone || mCurrent == '|' || mCurrent == ')')
{
return(piece);
}
RegexAndNode andNode = new RegexAndNode();
andNode.Children.Add(piece);
piece.Parent = andNode;
while (!(mParseDone || mCurrent == '|' || mCurrent == ')'))
{
RegexNode nextPiece = CompilePiece();
andNode.Children.Add(nextPiece);
nextPiece.Parent = andNode;
}
return(andNode);
}
//Compile the expression i.e. main body or expr in paranthesis
public RegexNode CompileExpr()
{
RegexNode branch = CompileBranch();
if (mCurrent != '|')
{
return(branch);
}
RegexOrNode expr = new RegexOrNode();
expr.Children.Add(branch);
branch.Parent = expr;
while (mCurrent == '|')
{
NextChar();
RegexNode nextBranch = CompileBranch();
expr.Children.Add(nextBranch);
nextBranch.Parent = expr;
}
return(expr);
}
//Compile \d \D \s \S etc.
public RegexNode CompileSimpleMacro(char c)
{
RegexNode node = null;
RegexSetNode set = null;
if (@"[]{}()*-+.?\|".Contains(c.ToString()))
{
return(new RegexTextNode(c.ToString()));
}
switch (c)
{
case 'd': // [0-9]
node = set = new RegexSetNode(true);
set.AddRange('0', '9');
break;
case 'D': // [^0-9]
node = set = new RegexSetNode(false);
set.AddRange('0', '9');
break;
case 's':
node = set = new RegexSetNode(true);
set.AddChars(" \r\n\f\v\t");
break;
case 'S':
node = set = new RegexSetNode(false);
set.AddChars(" \r\n\f\v\t");
break;
case 'w': // [a-zA-Z0-9_]
node = set = new RegexSetNode(true);
set.AddRange('a', 'z');
set.AddRange('A', 'Z');
set.AddRange('0', '9');
set.AddChars("_");
break;
case 'W': // [^a-zA-Z0-9_]
node = set = new RegexSetNode(false);
set.AddRange('a', 'z');
set.AddRange('A', 'Z');
set.AddRange('0', '9');
set.AddChars("_");
break;
case 'f':
node = new RegexTextNode("\f");
break;
case 'n':
node = new RegexTextNode("\n");
break;
case 'r':
node = new RegexTextNode("\r");
break;
case 't':
node = new RegexTextNode("\t");
break;
case 'v':
node = new RegexTextNode("\v");
break;
case 'A':
case 'Z':
case 'z':
node = new RegexTextNode(String.Empty);
break;
default:
AssertParse(false, "Invalid escape.");
break;
}
return(node);
}
//Compile token
public RegexNode CompileAtom()
{
RegexNode atom = null;
RegexSetNode set = null;
int start = 0;
int end = 0;
AssertParse(!mParseDone, "Reached end of string. No element found.");
AssertParse(!("|)?+*{}".Contains(mCurrent.ToString())), "No element found.");
switch (mCurrent)
{
case '.': //Any single char
atom = set = new RegexSetNode(true);
set.AddRange(Convert.ToChar(0), Convert.ToChar(127));
NextChar();
break;
case '[': //Positive or negative set
NextChar();
atom = CompileSet();
break;
case '(': //Sub expression
int refIndex = 0; //-2 -> don't capture, -1 -> named capture, 0-> indexed capture
NextChar();
//By default, subexpressions must be captured for future reference,
if (mCurrent == '?')
{
NextChar();
if (mCurrent == ':') //If sub expression begins with ?: it means don't store reference
{
NextChar();
refIndex = -2;
}
else //Named backreference, extract backreference name
{
ExtractBackrefName(ref start, ref end);
refIndex = -1;
}
} //else use indexed backreference
atom = new RegexSubExpressionNode(CompileExpr());
AssertParse(mCurrent == ')', "Expected ')'");
NextChar();
if (refIndex == -1) //Named backreference
{
(atom as RegexSubExpressionNode).Name = mRegex.ToString().Substring(start, end - start + 1);
mNamedBackRefs.Add(atom);
}
else if (refIndex == 0) //Indexed backreference
{
mBackRefs.Add(atom);
}
break;
case '^':
case '$':
atom = new RegexTextNode(String.Empty);
NextChar();
break;
case '\\':
NextChar();
if (Char.ToLower(mCurrent) == 'x' || Char.ToLower(mCurrent) == 'u' || mCurrent == '0')
{
atom = new RegexTextNode(EscapeValue().ToString());
}
else if (Char.IsDigit(mCurrent))
{
atom = GetBackRef((int)EscapeValue());
AssertParse(atom != null, "Couldn't find back reference");
atom = new RegexSubExpressionNode(atom);
}
else if (mCurrent == 'k') //referencing a backreference by name
{
NextChar();
ExtractBackrefName(ref start, ref end);
atom = GetBackRef(mRegex.ToString().Substring(start, end - start + 1));
AssertParse(atom != null, "Couldn't find back reference");
atom = new RegexSubExpressionNode(atom); //Create a copy of the referenced node
}
else
{
atom = CompileSimpleMacro(mCurrent);
NextChar();
}
break;
default:
int closeIndex = mRegex.ToString().IndexOfAny("-*+?(){}\\[]^$.|".ToCharArray(), mIndex + 1);
if (closeIndex == -1)
{
mParseDone = true;
closeIndex = mRegex.Length - 1;
atom = new RegexTextNode(mRegex.ToString().Substring(mIndex, closeIndex - mIndex + 1));
}
else
{
atom = new RegexTextNode(mRegex.ToString().Substring(mIndex, closeIndex - mIndex));
}
mIndex = closeIndex;
mCurrent = mRegex[mIndex];
break;
}
return(atom);
}
//Compile token followed by *+?{}
public RegexNode CompilePiece()
{
RegexNode node = null;
//store the old invalidating state for restoring after this node
bool oldInvalidState = RegexCompiler.IsInvalidSection;
//check if we want to invalidate the 'atom' node and subnodes
if (mCurrent == '\\' && mRegex[mIndex + 1] == 'i') //entering invalidating nodes section
{
NextChar();
NextChar();
//invalidate the following node and subnodes
RegexCompiler.IsInvalidSection = true;
}
RegexNode atom = CompileAtom();
//revert the invalidating state
RegexCompiler.IsInvalidSection = oldInvalidState;
//check special case of invalidating a repeating node
//have to confirm with "*+?{" to verify that it's not another type of node (that parses elsewhere)
if (mCurrent == '\\' && mRegex[mIndex + 1] == 'i' && "*+?{".Contains(mRegex[mIndex + 2].ToString()))
{
NextChar();
NextChar();
//invalidate the repeating node
RegexCompiler.IsInvalidSection = true;
}
const int MAXREPEAT = -1; //value representing infinity
switch (mCurrent)
{
case '*': //zero or more repetition
node = new RegexRepeatNode(atom, 0, MAXREPEAT, false);
NextChar();
break;
case '+': //one or more repetition
node = new RegexRepeatNode(atom, 1, MAXREPEAT, false);
NextChar();
break;
case '?': //zero or one repetition
node = new RegexRepeatNode(atom, 0, 1, false);
NextChar();
break;
case '{': //Min and max repetition limits defined
int nMin = 0;
int nMax = 0;
bool sameChar = false;
NextChar();
if (mCurrent == '=')
{
sameChar = true;
NextChar();
}
int closeIndex = mRegex.ToString().IndexOf('}', mIndex);
AssertParse(closeIndex != -1, "Expected '}'");
string[] repeatTokens = mRegex.ToString().Substring(mIndex, closeIndex - mIndex).
Split(new char[] { ',' });
if (repeatTokens.Length == 1)
{
nMin = nMax = int.Parse(repeatTokens[0]);
}
else if (repeatTokens.Length == 2)
{
nMin = int.Parse(repeatTokens[0]);
//check for {n,} case
if (repeatTokens[1].Length > 0)
{
nMax = int.Parse(repeatTokens[1]);
}
else
{
nMax = MAXREPEAT; //only lower bound specified
}
}
else
{
AssertParse(false, "Repeat values cannot be parsed");
}
AssertParse(nMin <= nMax || repeatTokens[1].Length == 0, "Max repeat is less than min repeat");
mIndex = closeIndex;
NextChar();
node = new RegexRepeatNode(atom, nMin, nMax, sameChar);
break;
default:
node = atom;
break;
}
//revert invalidation after generating the repeating node
RegexCompiler.IsInvalidSection = oldInvalidState;
return(node);
}
private RegexNode mReservedPath = null; //The child node that this Or Node must choose
//Chosen node is random if this is null
public override void ReservePath(RegexNode child)
{
//this child (in Children) must be called when generating the string
mReservedPath = child;
base.ReservePath(child);
}
public override void ReservePath(RegexNode child)
{
//this child (mRefNode) must be called when generating the string (cannot repeat zero times)
mReservedPath = child;
base.ReservePath(child);
}
public string Name; //Identifies subexpression by name, used for named backreferences
public RegexSubExpressionNode(RegexNode subExpr)
{
mRefNode = subExpr;
mRefNode.Parent = this;
}