Parser.cpp

#include "Parser.h"
#include "CommonRegexes.h"
#include "Log.h"
#include "ParserHelpers.h"
#include "StringUtils.h"
#include <filesystem>
#include <fstream>
namespace fs = std::filesystem;

namespace commonItems
{
std::string getNextLexeme(std::istream& theStream);
} // namespace commonItems


void commonItems::absorbBOM(std::istream& theStream)
{
	if (const auto firstChar = static_cast<char>(theStream.peek()); firstChar == '\xEF')
	{
		char bitBucket[3];
		theStream.read(bitBucket, sizeof bitBucket);
	}
}


void commonItems::parser::registerKeyword(const std::string& keyword, const parsingFunction& function)
{
	registeredKeywordStrings.insert(std::make_pair(keyword, function));
}


void commonItems::parser::registerKeyword(const std::string& keyword, const parsingFunctionStreamOnly& function)
{
	registeredKeywordStringsStreamOnly.insert(std::make_pair(keyword, function));
}


void commonItems::parser::registerRegex(const std::string& keyword, const parsingFunction& function)
{
	generatedRegexes.emplace_back(std::regex(keyword), function);
}


void commonItems::parser::IgnoreUnregisteredItems()
{
	registerRegex(catchallRegex, ignoreItem);
}


void commonItems::parser::IgnoreAndLogUnregisteredItems()
{
	registerRegex(catchallRegex, ignoreAndLogItem);
}


void commonItems::parser::IgnoreAndStoreUnregisteredItems(std::set<std::string>& ignored_tokens)
{
	registerRegex(catchallRegex, [&ignored_tokens](const std::string& key, std::istream& the_stream) {
		ignored_tokens.insert(key);
		ignoreItem(key, the_stream);
	});
}


void commonItems::parser::parseStream(std::istream& theStream)
{
	auto braceDepth = 0;
	auto value = false; // tracker to indicate whether we reached the value part of key=value pair
	std::string tokensSoFar;

	while (true)
	{
		if (auto token = getNextToken(theStream); token)
		{
			tokensSoFar += *token;
			if (*token == "=")
			{
				if (!value)
				{
					value = true; // swapping to value part.
					continue;
				}

				// value is positive, meaning we were at value, and now we're hitting an equal. This is bad. We need to
				// manually fast-forward to brace-lvl 0 and die.
				char inputChar;
				while (braceDepth)
				{
					theStream >> inputChar;
					if (inputChar == '{')
						braceDepth++;
					if (inputChar == '}')
						braceDepth--;
					if (!isspace(inputChar))
						tokensSoFar += inputChar;
				}
				Log(LogLevel::Warning) << "Broken token syntax at " << tokensSoFar;
				return;
			}
			if (*token == "{")
				braceDepth++;
			else if (*token == "}")
			{
				braceDepth--;
				if (braceDepth == 0)
					break;
			}
			else
				Log(LogLevel::Warning) << "Unknown token while parsing stream: " << *token;
		}
		else
			break;
	}
}


void commonItems::parser::parseFile(std::string_view filename)
{
	std::ifstream theFile(fs::u8path(filename));
	if (!theFile.is_open())
	{
		Log(LogLevel::Error) << "Could not open " << filename << " for parsing.";
		return;
	}

	absorbBOM(theFile);
	parseStream(theFile);
	theFile.close();
}

void commonItems::parser::clearRegisteredKeywords() noexcept
{
	std::map<std::string, parsingFunction>().swap(registeredKeywordStrings);
	std::map<std::string, parsingFunctionStreamOnly>().swap(registeredKeywordStringsStreamOnly);
	std::vector<std::pair<std::regex, parsingFunction>>().swap(generatedRegexes);
}

void commonItems::parser::operator()(std::istream& theStream)
{
	this->parseStream(theStream);
}

std::optional<std::string> commonItems::parser::getNextToken(std::istream& theStream)
{
	theStream >> std::noskipws;
	std::string toReturn;

	auto gotToken = false;
	while (!gotToken)
	{
		if (theStream.eof())
			return {};

		toReturn = getNextLexeme(theStream);

		const auto strippedLexeme = remQuotes(toReturn);
		const auto isLexemeQuoted = strippedLexeme.size() < toReturn.size();

		auto matched = tryToMatchAgainstKeywords(toReturn, strippedLexeme, isLexemeQuoted, theStream);

		if (!matched)
			matched = tryToMatchAgainstRegexes(toReturn, strippedLexeme, isLexemeQuoted, theStream);

		if (!matched)
			gotToken = true;
	}

	if (!toReturn.empty())
		return std::move(toReturn);
	return std::nullopt;
}


inline bool commonItems::parser::tryToMatchAgainstKeywords(const std::string& toReturn,
	 const std::string& strippedLexeme,
	 bool isLexemeQuoted,
	 std::istream& theStream)
{
	if (const auto& match = registeredKeywordStringsStreamOnly.find(toReturn); match != registeredKeywordStringsStreamOnly.end())
	{
		match->second(theStream);
		return true;
	}
	if (const auto& match = registeredKeywordStrings.find(toReturn); match != registeredKeywordStrings.end())
	{
		match->second(toReturn, theStream);
		return true;
	}
	if (isLexemeQuoted)
	{
		if (const auto& strippedMatch = registeredKeywordStringsStreamOnly.find(strippedLexeme); strippedMatch != registeredKeywordStringsStreamOnly.end())
		{
			strippedMatch->second(theStream);
			return true;
		}
		if (const auto& strippedMatch = registeredKeywordStrings.find(strippedLexeme); strippedMatch != registeredKeywordStrings.end())
		{
			strippedMatch->second(toReturn, theStream);
			return true;
		}
	}

	return false;
}

inline bool commonItems::parser::tryToMatchAgainstRegexes(const std::string& toReturn,
	 const std::string& strippedLexeme,
	 bool isLexemeQuoted,
	 std::istream& theStream)
{
	for (const auto& [regex, parsingFunction]: generatedRegexes)
	{
		if (std::smatch match; std::regex_match(toReturn, match, regex))
		{
			parsingFunction(toReturn, theStream);
			return true;
		}
	}
	if (isLexemeQuoted)
	{
		for (const auto& [regex, parsingFunction]: generatedRegexes)
		{
			if (std::smatch match; std::regex_match(strippedLexeme, match, regex))
			{
				parsingFunction(toReturn, theStream);
				return true;
			}
		}
	}
	return false;
}


std::optional<std::string> commonItems::parser::getNextTokenWithoutMatching(std::istream& theStream)
{
	theStream >> std::noskipws;

	std::string toReturn;

	auto gotToken = false;
	while (!gotToken)
	{
		if (theStream.eof())
			return std::nullopt;

		toReturn = getNextLexeme(theStream);
		gotToken = true;
	}

	if (!toReturn.empty())
		return std::move(toReturn);
	return std::nullopt;
}


std::string commonItems::getNextLexeme(std::istream& theStream)
{
	std::string toReturn;

	auto inQuotes = false;
	auto inLiteralQuote = false;
	unsigned char previousCharacter = '\0';

	while (true)
	{
		char inputChar;
		theStream >> inputChar;
		if (theStream.eof())
			break;
		if (!inQuotes && inputChar == '#')
		{
			std::string bitBucket;
			std::getline(theStream, bitBucket);
			if (!toReturn.empty())
				break;
		}
		else if (inputChar == '\n')
		{
			if (!inQuotes)
			{
				if (!toReturn.empty())
					break;
			}
			else
			{
				// fix paradox' mistake and don't break proper names in half
				toReturn += " ";
			}
		}
		else if (inputChar == '\"' && !inQuotes && toReturn.empty())
		{
			inQuotes = true;
			toReturn += inputChar;
		}
		else if (inputChar == '\"' && !inQuotes && toReturn.size() == 1 && toReturn.back() == 'R')
		{
			inLiteralQuote = true;
			toReturn.pop_back();
			toReturn += inputChar;
		}
		else if (inputChar == '(' && inLiteralQuote && toReturn.size() == 1)
		{
			continue;
		}
		else if (inputChar == '\"' && inLiteralQuote && previousCharacter == ')')
		{
			toReturn.pop_back();
			toReturn += inputChar;
			break;
		}
		else if (inputChar == '\"' && inQuotes && previousCharacter != '\\')
		{
			toReturn += inputChar;
			break;
		}
		else if (!inQuotes && !inLiteralQuote && std::isspace(inputChar))
		{
			if (!toReturn.empty())
				break;
		}
		else if (!inQuotes && !inLiteralQuote && inputChar == '{')
		{
			if (toReturn.empty())
			{
				toReturn += inputChar;
			}
			else
			{
				theStream.putback('{');
			}
			break;
		}
		else if (!inQuotes && !inLiteralQuote && inputChar == '}')
		{
			if (toReturn.empty())
			{
				toReturn += inputChar;
			}
			else
			{
				theStream.putback('}');
			}
			break;
		}
		else if (!inQuotes && !inLiteralQuote && inputChar == '=')
		{
			if (toReturn.empty())
			{
				toReturn += inputChar;
			}
			else
			{
				theStream.putback('=');
			}
			break;
		}
		else
		{
			toReturn += inputChar;
		}

		previousCharacter = inputChar;
	}
	return toReturn;
}