//This adds the transitions required to make a group trie correspond to the correct quantified regex
private RegexTrie <T> joinQuantifiedTrie(MinMaxResult result,
RegexTrie <T> startNodeOfCurrentComponent,
RegexTrie <T> endNodeOfCurrentComponent,
RegexTrie <T> endNodeOfPreviousComponent)
{
//This is the end node which will be reached by epsilon transition in case of zeroed quantifiers
//or quantifiers with a number range of more than one, start node is used for zeroing
RegexTrie <T> epsilonEndNode = new RegexTrie <T>();
RegexTrie <T> epsilonStartNode = new RegexTrie <T>();
//This holds the end node used during the loop
RegexTrie <T> loopEndNode = endNodeOfCurrentComponent;
//Loop from 1 (one occurrence already exists) to max occurrences. If loop index is greater than min,
//add epsilon from iteration trie end to final trie end.
if (result.minOccurs > 1 || result.maxOccurs > 1)
{
//This is used as a model for copying, only needs to be calculated if there's a set number of occurrences
KeyValuePair <RegexTrie <T>, RegexTrie <T> > trieModel = copyTrie(startNodeOfCurrentComponent, endNodeOfCurrentComponent);
//Result of trie copy
KeyValuePair <RegexTrie <T>, RegexTrie <T> > trieCopy;
for (int index = 2; index <= result.maxOccurs; index++)
{
if (index <= result.maxOccurs)
{
//Copy the trie and link it to the end of trie
trieCopy = copyTrie(trieModel);
loopEndNode.AddEpsilonTransition(trieCopy.Key);
//If the index is equal or greater than minoccurs, add epsilon to end node
if (result.minOccurs <= index)
{
trieCopy.Value.AddEpsilonTransition(epsilonEndNode);
}
//Assign the end node of the copy as the end node of the component
loopEndNode = trieCopy.Value;
}
}
}
//If there's less than two occurrences, connect the end of the first span of the trie to the epsilon end trie
//Do this after the copy loop in order to not disturb the copying.
if (result.minOccurs < 2)
{
endNodeOfCurrentComponent.AddEpsilonTransition(epsilonEndNode);
}
//If the quantifier is starred, add epsilon from end to start
if (result.starred)
{
endNodeOfCurrentComponent.AddEpsilonTransition(startNodeOfCurrentComponent);
}
//If the quantifier is zeroed, add an epsilon from epsilon start node
//newCurrentTrie
if (result.zeroInclusive)
{
epsilonStartNode.AddEpsilonTransition(startNodeOfCurrentComponent);
epsilonStartNode.AddEpsilonTransition(epsilonEndNode);
startNodeOfCurrentComponent = epsilonStartNode;
}
//Connect trie to previous trie
endNodeOfPreviousComponent.AddEpsilonTransition(startNodeOfCurrentComponent);
return(epsilonEndNode);
}
//This checks for a quantifier immediately after the character, and joins the current trie to the end of the
//previous trie. Returns a pair of int (to increment string index) and the node to which further tries will be connected
private KeyValuePair <int, RegexTrie <T> > checkForQuantifier(
string sourceTail,
RegexTrie <T> endNodeOfCurrentComponent,
RegexTrie <T> startNodeOfCurrentComponent,
RegexTrie <T> endNodeOfPreviousComponent,
Stack <RegexTrie <T> > commonDestination)
{
char nextChar;
try
{
nextChar = sourceTail[0];
}
catch
{
//assign a non-special character to trigger the default case
nextChar = 'n';
}
switch (nextChar)
{
case '*':
{
return(new KeyValuePair <int, RegexTrie <T> >(1, joinQuantifiedTrie(
new MinMaxResult(1, 1, true, true),
startNodeOfCurrentComponent,
endNodeOfCurrentComponent,
endNodeOfPreviousComponent)));
}
case '+':
{
return(new KeyValuePair <int, RegexTrie <T> >(1, joinQuantifiedTrie(
new MinMaxResult(1, 1, true, false),
startNodeOfCurrentComponent,
endNodeOfCurrentComponent,
endNodeOfPreviousComponent)));
}
case '?':
{
return(new KeyValuePair <int, RegexTrie <T> >(1, joinQuantifiedTrie(
new MinMaxResult(1, 1, false, true),
startNodeOfCurrentComponent,
endNodeOfCurrentComponent,
endNodeOfPreviousComponent)));
}
case '{':
{
int closeBraceIndex = sourceTail.IndexOf('}');
string quantString = sourceTail.Substring(1, closeBraceIndex - 1);
MinMaxResult minMax = findMinMax(quantString);
return(new KeyValuePair <int, RegexTrie <T> >(closeBraceIndex + 1, joinQuantifiedTrie(
minMax,
startNodeOfCurrentComponent,
endNodeOfCurrentComponent,
endNodeOfPreviousComponent)));
}
default:
{
endNodeOfPreviousComponent.AddEpsilonTransition(startNodeOfCurrentComponent);
return(new KeyValuePair <int, RegexTrie <T> >(0, endNodeOfCurrentComponent));
}
}
}
