miloignis/kraken
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Extended the Regular Expression class to now support * and ? as well as +. Next up, perens and alternation

Browse Source
This commit is contained in:
Nathan Braswell
2013-07-03 23:40:36 -04:00
parent cc6ff21986
commit 84566c4ff6
2 changed files with 68 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

61
src/RegEx.cpp
View File

@@ -2,14 +2,65 @@
RegEx::RegEx(std::string inPattern) { RegEx::RegEx(std::string inPattern) {
pattern = inPattern; pattern = inPattern;
RegExState* current; std::vector<RegExState*> previousStates;
std::vector<RegExState*> currentStates;
begin = new RegExState(); begin = new RegExState();
current = begin; currentStates.push_back(begin);
for (int i = 0; i < pattern.length(); i++) { for (int i = 0; i < pattern.length(); i++) {
RegExState* next = new RegExState(pattern[i]); switch (pattern[i]) {
current->addNext(next); case '*':
current = next; {
std::cout << "Star at " << i << " in " << pattern << std::endl;
for (std::vector<RegExState*>::size_type j = 0; j < currentStates.size(); j++)
for (std::vector<RegExState*>::size_type k = 0; k < currentStates.size(); k++)
currentStates[j]->addNext(currentStates[k]);
//add all previous states to current states to enable skipping over the starred item
currentStates.insert(currentStates.end(), previousStates.begin(), previousStates.end());
}
break;
case '+':
{
std::cout << "Plus at " << i << " in " << pattern << std::endl;
//OtherThingy
//current->addNext(current);
for (std::vector<RegExState*>::size_type j = 0; j < currentStates.size(); j++)
for (std::vector<RegExState*>::size_type k = 0; k < currentStates.size(); k++)
currentStates[j]->addNext(currentStates[k]);
}
break;
case '?':
{
std::cout << "Question at " << i << " in " << pattern << std::endl;
//add all previous states to current states to enable skipping over the questioned item
currentStates.insert(currentStates.end(), previousStates.begin(), previousStates.end());
}
break;
case '|':
std::cout << "Alternation at " << i << " in " << pattern << std::endl;
//alternation
break;
case '(':
std::cout << "Begin peren at " << i << " in " << pattern << std::endl;
//perentheses
break;
default:
{
std::cout << "Regular" << std::endl;
//Ahh, it's regular
RegExState* next = new RegExState(pattern[i]);
for (std::vector<RegExState*>::size_type j = 0; j < currentStates.size(); j++)
currentStates[j]->addNext(next);
previousStates.clear();
previousStates.insert(previousStates.begin(), currentStates.begin(), currentStates.end());
currentStates.clear();
currentStates.push_back(next);
}
}
} }
//last one is goal state
for (std::vector<RegExState*>::size_type i = 0; i < currentStates.size(); i++)
currentStates[i]->addNext(NULL);
} }
RegEx::~RegEx() { RegEx::~RegEx() {

15
src/RegExState.cpp
View File

@@ -29,21 +29,30 @@ bool RegExState::characterIs(char inCharacter) {
std::vector<RegExState*>* RegExState::advance(char advanceCharacter) { std::vector<RegExState*>* RegExState::advance(char advanceCharacter) {
std::vector<RegExState*>* advanceStates = new std::vector<RegExState*>(); std::vector<RegExState*>* advanceStates = new std::vector<RegExState*>();
for (std::vector<RegExState*>::size_type i = 0; i < nextStates.size(); i++) { for (std::vector<RegExState*>::size_type i = 0; i < nextStates.size(); i++) {
if (nextStates[i]->characterIs(advanceCharacter)) if (nextStates[i] != NULL && nextStates[i]->characterIs(advanceCharacter))
advanceStates->push_back(nextStates[i]); advanceStates->push_back(nextStates[i]);
} }
return advanceStates; return advanceStates;
} }
bool RegExState::isGoal() { bool RegExState::isGoal() {
return inner == NULL && nextStates.size() == 0; //return inner == NULL && nextStates.size() == 0;
for (std::vector<RegExState*>::size_type i = 0; i < nextStates.size(); i++)
if (nextStates[i] == NULL)
return true;
return false;
} }
std::string RegExState::toString() { std::string RegExState::toString() {
std::string string = ""; std::string string = "";
string += std::string("\"") + character + "\""; string += std::string("\"") + character + "\"";
for (std::vector<RegExState*>::size_type i = 0; i < nextStates.size(); i++) for (std::vector<RegExState*>::size_type i = 0; i < nextStates.size(); i++)
string += "->" + nextStates[i]->toString() + " EC "; if (nextStates[i] != this && nextStates[i] != NULL)
string += "->" + nextStates[i]->toString() + " EC ";
else if (nextStates[i] == NULL)
string += "-> GOAL ";
else
string += "->this";
//std::cout << "inner = " << inner << " nextStates size = " << nextStates.size() <<std::endl; //std::cout << "inner = " << inner << " nextStates size = " << nextStates.size() <<std::endl;
return string; return string;
} }