KSG-Framework/Regex.cpp

#include "Regex.h"

using namespace Framework;
using namespace Regex;

int findEndPos(Text* regex, int startPos, char endChar, char startChar)
{
    int length = regex->getLength();
    int i = startPos;
    int opened = 0;
    while (i < length)
    {
        if (regex->getText()[i] == '\\')
        {
            i++;
        }
        else if (regex->getText()[i] == startChar)
        {
            opened++;
        }
        else if (regex->getText()[i] == endChar)
        {
            if (opened == 0)
            {
                return i;
            }
            opened--;
        }
        i++;
    }
    return -1;
}

Automata<char>* parseCharacterList(Text* regex, RegexConfig& config)
{
    bool negated = regex->hatAt(0, "^");
    Automata<char>* result = new Automata<char>();
    State<char>* start = result->addState();
    State<char>* end = result->addState();
    end->setFinal(true);
    int length = regex->getLength();
    bool escaped = false;
    bool minusValid = false;
    Text characterList;
    std::function<bool(char, int)> negatedFunction
        = [](char c, int flags) { return true; };
    for (int i = 0; i < length; i++)
    {
        switch (regex->getText()[i])
        {
        case '\\':
            {
                escaped = true;
                minusValid = false;
                break;
            }
        case '-':
            if (escaped || !minusValid)
            {
                characterList += '-';
            }
            else if (minusValid)
            {
                if (i == length - 1)
                {
                    characterList += '-';
                }
                else
                {
                    unsigned char before = (unsigned char)
                        characterList[characterList.getLength() - 1];
                    characterList.remove(characterList.getLength() - 1, 1);
                    unsigned char after
                        = (unsigned char)regex->getText()[i + 1];
                    if (before > after)
                    {
                        unsigned char temp = before;
                        before = after;
                        after = temp;
                    }
                    for (unsigned char c = before; c <= after; c++)
                    {
                        characterList += c;
                    }
                    i++;
                }
                break;
            }
        default:
            {
                if (escaped)
                {
                    if (regex->getText()[i] == 'w')
                    {
                        if (negated)
                        {
                            negatedFunction
                                = [config, negatedFunction](char c, int flags) {
                                      return negatedFunction(c, flags)
                                          && !config.getWordChars().hat(c);
                                  };
                        }
                        else
                        {
                            characterList += config.getWordChars();
                        }
                        minusValid = false;
                    }
                    else if (regex->getText()[i] == 'W')
                    {
                        if (negated)
                        {
                            characterList += config.getWordChars();
                        }
                        else
                        {
                            start->addTransition(Transition<char>(
                                [config](char c, int flags) {
                                    return !config.getWordChars().hat(c);
                                },
                                end,
                                0));
                        }
                        minusValid = false;
                    }
                    else if (regex->getText()[i] == 'd')
                    {
                        if (negated)
                        {
                            negatedFunction
                                = [config, negatedFunction](char c, int flags) {
                                      return negatedFunction(c, flags)
                                          && (c < '0' || c > '9');
                                  };
                        }
                        else
                        {
                            characterList.append("0123456789");
                        }
                        minusValid = false;
                    }
                    else if (regex->getText()[i] == 's')
                    {
                        if (negated)
                        {
                            negatedFunction = [config, negatedFunction](
                                                  char c, int flags) {
                                return negatedFunction(c, flags)
                                    && !config.getWhitespaceChars().hat(c);
                            };
                        }
                        else
                        {
                            characterList.append(config.getWhitespaceChars());
                        }
                        minusValid = false;
                    }
                    else if (regex->getText()[i] == 'D')
                    {
                        if (negated)
                        {
                            negatedFunction
                                = [negatedFunction](char c, int flags) {
                                      return negatedFunction(c, flags)
                                          && (c >= '0' && c <= '9');
                                  };
                        }
                        else
                        {
                            start->addTransition(Transition<char>(
                                [](char c, int flags) {
                                    return c < '0' || c > '9';
                                },
                                end,
                                0));
                        }
                        minusValid = false;
                    }
                    else if (regex->getText()[i] == 'S')
                    {
                        if (negated)
                        {
                            negatedFunction
                                = [config, negatedFunction](char c, int flags) {
                                      return negatedFunction(c, flags)
                                          && config.getWhitespaceChars().hat(c);
                                  };
                        }
                        else
                        {
                            start->addTransition(Transition<char>(
                                [config](char c, int flags) {
                                    return !config.getWhitespaceChars().hat(c);
                                },
                                end,
                                0));
                        }
                        minusValid = false;
                    }
                    else if (regex->getText()[i] == 'n')
                    {
                        characterList.append('\n');
                    }
                    else if (regex->getText()[i] == 'r')
                    {
                        characterList.append('\r');
                    }
                    else if (regex->getText()[i] == 't')
                    {
                        characterList.append('\t');
                    }
                    else
                    {
                        characterList.append(regex->getText()[i]);
                        minusValid = true;
                    }
                }
                else
                {
                    characterList.append(regex->getText()[i]);
                    minusValid = true;
                }
                break;
            }
        }
    }
    if (negated)
    {
        negatedFunction = [characterList, negatedFunction](char c, int flags) {
            return negatedFunction(c, flags) && !characterList.hat(c);
        };
        start->addTransition(Transition<char>(negatedFunction, end, 0));
    }
    else
    {
        start->addTransition(Transition<char>(
            [characterList](char c, int flags) { return characterList.hat(c); },
            end,
            0));
    }
    regex->release();
    return result;
}

Automata<char>* concat(Array<Automata<char>*> parsers)
{
    if (parsers.getEintragAnzahl() == 0)
    {
        Automata<char>* result = new Automata<char>();
        State<char>* start = result->addState();
        State<char>* end = result->addState();
        end->setFinal(true);
        start->addTransition(Transition<char>(0, end, 0));
        return result;
    }
    Automata<char>* result = 0;
    for (Automata<char>* parser : parsers)
    {
        if (result == 0)
        {
            result = parser;
        }
        else
        {
            result = concat(result, parser);
        }
    }
    return result;
}

Automata<char>* parse(Text* regex, int& nextGroupId, RegexConfig& config)
{
    Array<Automata<char>*> results;
    int length = regex->getLength();
    for (int i = 0; i < length; i++)
    {
        bool escaped = false;
        switch (regex->getText()[i])
        {
        case '\\':
            {
                escaped = true;
                i++;
                break;
            }
        case '(':
            {
                int end = findEndPos(regex, i + 1, ')', '(');
                Text* subRegex = regex->getTeilText(i + 1, end);
                Automata<char>* subAutomata = 0;
                if (subRegex->hatAt(0, "?:"))
                {
                    subRegex->remove(0, 2);
                    subAutomata = parse(subRegex, nextGroupId, config);
                }
                else
                {
                    int groupId = nextGroupId++;
                    subAutomata
                        = group(parse(subRegex, nextGroupId, config), groupId);
                }
                results.add(subAutomata);
                i = end;
                break;
            }
        case '[':
            {
                int end = findEndPos(regex, i + 1, ']', 0);
                Text* subRegex = regex->getTeilText(i + 1, end);
                Automata<char>* subAutomata
                    = parseCharacterList(subRegex, config);
                results.add(subAutomata);
                i = end;
                break;
            }
        case '?':
            {
                Automata<char>* last
                    = results.get(results.getEintragAnzahl() - 1);
                results.remove(results.getEintragAnzahl() - 1);
                results.add(maybe(last));
                break;
            }
        case '*':
            {
                bool lazy = false;
                if (i < length - 1 && regex->getText()[i + 1] == '?')
                {
                    lazy = true;
                    i++;
                }
                Automata<char>* last
                    = results.get(results.getEintragAnzahl() - 1);
                results.remove(results.getEintragAnzahl() - 1);
                results.add(many(last, lazy));
                break;
            }
        case '+':
            {
                bool lazy = false;
                if (i < length - 1 && regex->getText()[i + 1] == '?')
                {
                    lazy = true;
                    i++;
                }
                Automata<char>* last
                    = results.get(results.getEintragAnzahl() - 1);
                results.remove(results.getEintragAnzahl() - 1);
                results.add(atLeastOnce(last, lazy));
                break;
            }
        case '.':
            {
                Automata<char>* any = new Automata<char>();
                State<char>* start = any->addState();
                State<char>* end = any->addState();
                end->setFinal(true);
                start->addTransition(Transition<char>(
                    [](char c, int flags) { return true; }, end, 0));
                results.add(any);
                break;
            }
        case '|':
            {
                Automata<char>* result = oneOf(concat(results),
                    parse(regex->getTeilText(i + 1), nextGroupId, config));
                regex->release();
                return result;
            }
        case '{':
            {
                int end = findEndPos(regex, i + 1, '}', 0);
                Text* subRegex = regex->getTeilText(i + 1, end);
                int min = 0;
                int max = 0;
                if (subRegex->hat(","))
                {
                    Text* minText
                        = subRegex->getTeilText(0, subRegex->positionVon(","));
                    Text* maxText
                        = subRegex->getTeilText(subRegex->positionVon(",") + 1);
                    min = (int)*minText;
                    max = (int)*maxText;
                    minText->release();
                    maxText->release();
                }
                else
                {
                    min = (int)*subRegex;
                    max = min;
                }
                Automata<char>* last
                    = results.get(results.getEintragAnzahl() - 1);
                results.remove(results.getEintragAnzahl() - 1);
                results.add(repeat(last, min, max));
                i = end;
                subRegex->release();
                break;
            }
        case '^':
            {
                Automata<char>* qutomata = new Automata<char>();
                State<char>* start = qutomata->addState();
                State<char>* end = qutomata->addState();
                end->setFinal(true);
                start->addTransition(Transition<char>(0, end, Flags::START));
                results.add(qutomata);
                break;
            }
        case '$':
            {
                Automata<char>* qutomata = new Automata<char>();
                State<char>* start = qutomata->addState();
                State<char>* end = qutomata->addState();
                end->setFinal(true);
                start->addTransition(Transition<char>(0, end, Flags::END));
                results.add(qutomata);
                break;
            }
        default:
            {
                Automata<char>* automata = new Automata<char>();
                State<char>* start = automata->addState();
                State<char>* end = automata->addState();
                end->setFinal(true);
                char current = regex->getText()[i];
                start->addTransition(Transition<char>(
                    [current](char c, int flags) { return c == current; },
                    end,
                    0));
                results.add(automata);
                break;
            }
        }
        if (escaped)
        {
            std::function<bool(char, int)> function;
            switch (regex->getText()[i])
            {
            case 'n':
                {
                    function = [](char c, int flags) { return c == '\n'; };
                    break;
                }
            case 'r':
                {
                    function = [](char c, int flags) { return c == '\r'; };
                    break;
                }
            case 't':
                {
                    function = [](char c, int flags) { return c == '\t'; };
                    break;
                }
            case 'w':
                {
                    function = [config](char c, int flags) {
                        return config.getWordChars().hat(c);
                    };
                    break;
                }
            case 'W':
                {
                    function = [config](char c, int flags) {
                        return !config.getWordChars().hat(c);
                    };
                    break;
                }
            case 'd':
                {
                    function = [](char c, int flags) {
                        return c >= '0' && c <= '9';
                    };
                    break;
                }
            case 'D':
                {
                    function
                        = [](char c, int flags) { return c < '0' || c > '9'; };
                    break;
                }
            case 's':
                {
                    function = [config](char c, int flags) {
                        return config.getWhitespaceChars().hat(c);
                    };
                    break;
                }
            case 'S':
                {
                    function = [config](char c, int flags) {
                        return !config.getWhitespaceChars().hat(c);
                    };
                    break;
                }
            default:
                {
                    char current = regex->getText()[i];
                    function
                        = [current](char c, int flags) { return c == current; };
                    break;
                }
            }
            Automata<char>* automata = new Automata<char>();
            State<char>* start = automata->addState();
            State<char>* end = automata->addState();
            end->setFinal(true);
            start->addTransition(Transition<char>(function, end, 0));
            results.add(automata);
        }
    }
    regex->release();
    return concat(results);
}

Framework::Regex::Group::Group(int start, int end, int id)
    : start(start),
      end(end),
      id(id)
{}

Framework::Regex::Group::Group()
    : start(-1),
      end(-1),
      id(-1)
{}

int Framework::Regex::Group::getStart()
{
    return start;
}

int Framework::Regex::Group::getEnd()
{
    return end;
}

int Framework::Regex::Group::getId()
{
    return id;
}

Framework::Regex::Result::Result(Array<Group> groups)
    : ReferenceCounter(),
      groups(groups)
{}

int Framework::Regex::Result::getGroupCount()
{
    return groups.getEintragAnzahl();
}

Group Framework::Regex::Result::getGroup(int index)
{
    return groups.get(index);
}

int Framework::Regex::Result::getStart()
{
    return groups.get(0).getStart();
}

int Framework::Regex::Result::getEnd()
{
    return groups.get(0).getEnd();
}

RegexConfig::RegexConfig()
{
    whitespaceChars = " \n\r\t";
    wordChars
        = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_";
}

void RegexConfig::setWhitespaceChars(Text whitespaceChars)
{
    this->whitespaceChars = whitespaceChars;
}

void RegexConfig::setWordChars(Text wordChars)
{
    this->wordChars = wordChars;
}

Text RegexConfig::getWhitespaceChars() const
{
    return whitespaceChars;
}

Text RegexConfig::getWordChars() const
{
    return wordChars;
}

Automata<char>* Framework::Regex::parse(Text regex)
{
    RegexConfig config;
    return parse(regex, config);
}

Automata<char>* Framework::Regex::parse(Text regex, RegexConfig& config)
{
    int nextGroupId = 1;
    return ::parse(new Text(regex), nextGroupId, config);
}