private void EvaluateLook(
int start,
int end,
FSA <TValue> fsa,
LookAheadQuantifier quantifier,
AstConcatNode concatNode,
ORegexOptions options)
{
bool isBehind = options.HasFlag(ORegexOptions.ReversePattern) ? !quantifier.IsBehind : quantifier.IsBehind;
bool isNegative = quantifier.IsNegative;
var condOptions = isBehind ? ORegexOptions.RightToLeft : ORegexOptions.None;
var concat = new AstConcatNode(concatNode.Children, concatNode.Range);
var root = new AstRootNode(concat,
true,
false,
concat.Range,
new[]
{
ORegexAstFactory <TValue> .MainCaptureName
});
var fa = Create(root, condOptions);
var oregex = new ORegex <TValue>(fa, condOptions);
var func = new ORegexPredicateEdge <TValue>("#look", oregex, isNegative, isBehind);
EvaluateCondition(start, end, fsa, func);
}
private void EvaluateRepeat(int start, int end, FSA <TValue> fsa, AstRepeatNode astRepeatNode, ORegexOptions options)
{
var toRepeat = astRepeatNode.Argument;
var prev = start;
for (int i = 0; i < astRepeatNode.MinCount; i++)
{
var next = CreateNewState(fsa);
Evaluate(prev, next, fsa, toRepeat, options);
prev = next;
}
if (astRepeatNode.MaxCount == int.MaxValue)
{
RepeatZeroOrInfinite(prev, end, fsa, toRepeat, astRepeatNode.IsLazy, options);
}
else
{
int count = astRepeatNode.MaxCount - astRepeatNode.MinCount - 1;
int next;
for (int i = 0; i < count; i++)
{
next = CreateNewState(fsa);
RepeatZeroOrOne(prev, next, fsa, toRepeat, astRepeatNode.IsLazy, options);
prev = next;
}
next = end;
RepeatZeroOrOne(prev, next, fsa, toRepeat, astRepeatNode.IsLazy, options);
}
}
public void Evaluate(int start, int end, FSA <TValue> fsa, AstNodeBase node, ORegexOptions options)
{
// ReSharper disable once CanBeReplacedWithTryCastAndCheckForNull
if (node is AstAtomNode <TValue> )
{
EvaluateAtom(start, end, fsa, (AstAtomNode <TValue>)node);
}
else if (node is AstConcatNode)
{
EvaluateConcat(start, end, fsa, (AstConcatNode)node, options);
}
else if (node is AstOrNode)
{
EvaluateOr(start, end, fsa, (AstOrNode)node, options);
}
else if (node is AstRepeatNode)
{
EvaluateRepeat(start, end, fsa, (AstRepeatNode)node, options);
}
else if (node is AstRootNode)
{
EvaluateRoot(start, end, fsa, (AstRootNode)node, options);
}
else
{
throw new NotImplementedException(node.GetType().Name);
}
}
private void EvaluateOr(int start, int end, FSA <TValue> fsa, AstOrNode node, ORegexOptions options)
{
foreach (var child in node.GetChildren())
{
Evaluate(start, end, fsa, child, options);
}
}
/// <summary>
/// Reverse any FSA.
/// Warning: Type of automaton can change from NFA to DFA or DFA to NFA.
/// </summary>
/// <param name="fsa"></param>
/// <returns></returns>
public FSA <TValue> ReverseFsa(FSA <TValue> fsa)
{
return
(new FSA <TValue>(fsa.Name,
fsa.Transitions.Select(x => new FSATransition <TValue>(x.EndState, x.Condition, x.BeginState)),
fsa.F, fsa.Q0)
{
ExactBegin = fsa.ExactBegin,
ExactEnd = fsa.ExactEnd,
CaptureNames = fsa.CaptureNames
});
}
private void RepeatZeroOrOne(int start, int end, FSA <TValue> fsa, AstNodeBase node, bool isLasy, ORegexOptions options)
{
if (isLasy)
{
fsa.AddEpsilonTransition(start, end);
Evaluate(start, end, fsa, node, options);
}
else
{
Evaluate(start, end, fsa, node, options);
fsa.AddEpsilonTransition(start, end);
}
}
public FSA <TValue> CreateRawFsa(AstRootNode root, ORegexOptions options)
{
var result = new FSA <TValue>(root.CaptureGroupNames[0])
{
CaptureNames = root.CaptureGroupNames
};
var start = result.NewState();
var end = result.NewState();
Evaluate(start, end, result, root, options);
result.AddFinal(end);
result.AddStart(start);
return(result);
}
private void EvaluateConcat(int start, int end, FSA <TValue> fsa, AstConcatNode node, ORegexOptions options)
{
var group = node as AstGroupNode;
if (group != null)
{
if (group.Quantifier != null)
{
// ReSharper disable once CanBeReplacedWithTryCastAndCheckForNull
if (group.Quantifier is CaptureQuantifier)
{
var captureQ = (CaptureQuantifier)group.Quantifier;
var sys = new SystemPredicateEdge <TValue>("#capture")
{
IsCapture = true,
CaptureName = captureQ.CaptureName,
CaptureId = captureQ.CaptureId
};
var startTmp = CreateNewState(fsa);
fsa.AddTransition(start, sys, startTmp);
start = startTmp;
var endTmp = CreateNewState(fsa);
fsa.AddTransition(endTmp, sys, end);
end = endTmp;
}
else if (group.Quantifier is LookAheadQuantifier)
{
var lookQ = (LookAheadQuantifier)group.Quantifier;
EvaluateLook(start, end, fsa, lookQ, group, options);
return;
}
}
}
var prev = start;
int next;
var children = node.GetChildren().ToArray();
for (int i = 0; i < children.Length - 1; i++)
{
next = CreateNewState(fsa);
Evaluate(prev, next, fsa, children[i], options);
prev = next;
}
next = end;
Evaluate(prev, next, fsa, children[children.Length - 1], options);
}
public CFSA(FSA <TValue> fsa)
{
ExactBegin = fsa.ExactBegin;
ExactEnd = fsa.ExactEnd;
Name = fsa.Name;
CaptureNames = fsa.CaptureNames;
_transitionMatrix = new IFSATransition <TValue> [fsa.StateCount][];
foreach (var look in fsa.Transitions.ToLookup(x => x.BeginState, x => x))
{
_transitionMatrix[look.Key] = look.ToArray();
}
_startState = fsa.Q0.First();
_finalsLookup = new bool[_transitionMatrix.Length];
foreach (var f in fsa.F)
{
_finalsLookup[f] = true;
}
}
private void RepeatZeroOrInfinite(int start, int end, FSA <TValue> fsa, AstNodeBase predicate, bool isLasy, ORegexOptions options)
{
var tmp = CreateNewState(fsa);
if (isLasy)
{
fsa.AddEpsilonTransition(start, end);
fsa.AddEpsilonTransition(tmp, end);
Evaluate(tmp, tmp, fsa, predicate, options);
fsa.AddEpsilonTransition(start, tmp);
}
else
{
Evaluate(tmp, tmp, fsa, predicate, options);
fsa.AddEpsilonTransition(tmp, end);
fsa.AddEpsilonTransition(start, tmp);
fsa.AddEpsilonTransition(start, end);
}
}
/// <summary>
/// Builds the Epsilon closure of states for the given NFA
/// </summary>
/// <param name="nfa"></param>
/// <param name="states"></param>
/// <returns></returns>
private static Set <int> EpsilonClosure(FSA <TValue> nfa, Set <int> states)
{
// Push all states onto a stack
Stack <int> uncheckedStack = new Stack <int>(states);
// Initialize EpsilonClosure(states) to states
Set <int> epsilonClosure = states;
while (uncheckedStack.Count != 0)
{
// Pop state t, the top element, off the stack
var t = uncheckedStack.Pop();
// For each state u with an edge from t to u labeled Epsilon
IEnumerable <FSATransition <TValue> > transitions;
if (nfa.TryGetTransitionsFrom(t, out transitions))
{
foreach (var input in transitions)
{
if (PredicateEdgeBase <TValue> .IsEpsilon(input.Condition))
{
int u = input.EndState;
// If u is not already in epsilonClosure, add it and push it onto stack
if (!epsilonClosure.Contains(u))
{
epsilonClosure.Add(u);
uncheckedStack.Push(u);
}
}
}
}
}
return(epsilonClosure);
}
/// <summary>
/// Convert any FSA to reversed DFA.
/// Warning: elminates any epsilon transition.
/// </summary>
/// <param name="fsa"></param>
/// <returns></returns>
public FSA <TValue> RotateFsa(FSA <TValue> fsa)
{
return(ToDfa(ReverseFsa(fsa)));
}
/// <summary>
/// Convert any FSA to minimized DFA.
/// Warning: elminates any epsilon transition.
/// </summary>
/// <param name="fsa"></param>
/// <returns></returns>
public FSA <TValue> MinimizeFsa(FSA <TValue> fsa)
{
fsa = RotateFsa(fsa);
fsa = RotateFsa(fsa);
return(fsa);
}
private void EvaluateRoot(int start, int end, FSA <TValue> fsa, AstRootNode astRootNode, ORegexOptions options)
{
fsa.ExactBegin = astRootNode.MatchBegin;
fsa.ExactEnd = astRootNode.MatchEnd;
Evaluate(start, end, fsa, astRootNode.Regex, options);
}
private void EvaluateCondition(int start, int end, FSA <TValue> fsa, PredicateEdgeBase <TValue> condition)
{
fsa.AddTransition(start, condition, end);
}
private void EvaluateAtom(int start, int end, FSA <TValue> fsa, AstAtomNode <TValue> node)
{
EvaluateCondition(start, end, fsa, node.Condition);
}
/// <summary>
/// Subset machine that employs the powerset construction or subset construction algorithm.
/// It creates a DFA that recognizes the same language as the given NFA.
/// </summary>
private static FSA <TValue> ToDfa(FSA <TValue> fsa)
{
FSA <TValue> dfa = new FSA <TValue>(fsa.Name)
{
ExactBegin = fsa.ExactBegin,
ExactEnd = fsa.ExactEnd,
CaptureNames = fsa.CaptureNames,
};
// Sets of NFA states which is represented by some DFA state
var markedStates = new HashSet <Set <int> >();
var unmarkedStates = new HashSet <Set <int> >();
// Gives a number to each state in the DFA
var dfaStateNum = new Dictionary <Set <int>, int>();
var nfaInitial = fsa.Q0.ToSet();
// Initially, EpsilonClosure(nfa.initial) is the only state in the DFAs states
// and it's unmarked.
var first = EpsilonClosure(fsa, nfaInitial);
unmarkedStates.Add(first);
// The initial dfa state
int dfaInitial = dfa.NewState();
dfaStateNum[first] = dfaInitial;
dfa.AddStart(dfaInitial);
while (unmarkedStates.Count != 0)
{
// Takes out one unmarked state and posteriorly mark it.
var aState = unmarkedStates.First();
// Removes from the unmarked set.
unmarkedStates.Remove(aState);
// Inserts into the marked set.
markedStates.Add(aState);
// If this state contains the NFA's final state, add it to the DFA's set of
// final states.
if (fsa.F.Any(x => aState.Contains(x)))
{
dfa.AddFinal(dfaStateNum[aState]);
}
// For each input symbol the NFA knows...
foreach (var current in fsa.Sigma)
{
// Next state
var next = EpsilonClosure(fsa, fsa.Move(aState, current));
if (next.Count > 0)
{
// If we haven't examined this state before, add it to the unmarkedStates,
// and make up a new number for it.
if (!unmarkedStates.Contains(next) && !markedStates.Contains(next))
{
unmarkedStates.Add(next);
dfaStateNum.Add(next, dfa.NewState());
}
var from = dfaStateNum[aState];
var to = dfaStateNum[next];
var condition = current;
dfa.AddTransition(from, condition, to);
}
}
}
return(dfa);
}
private int CreateNewState(FSA <TValue> fsa)
{
return(fsa.NewState());
}