/*****************************************************************************
 * Copyright (c) 2014-2023 OpenRCT2 developers
 *
 * For a complete list of all authors, please refer to contributors.md
 * Interested in contributing? Visit https://github.com/OpenRCT2/OpenRCT2
 *
 * OpenRCT2 is licensed under the GNU General Public License version 3.
 *****************************************************************************/

#include <algorithm>
#include <cctype>
#include <cwctype>
#include <iomanip>
#include <sstream>
#include <stdexcept>
#include <vector>
#ifndef _WIN32
#    include <unicode/ucnv.h>
#    include <unicode/unistr.h>
#    include <unicode/utypes.h>
#endif

#ifdef _WIN32
#    include <windows.h>
#endif

#include "../common.h"
#include "../localisation/ConversionTables.h"
#include "../localisation/FormatCodes.h"
#include "../localisation/Language.h"
#include "../util/Util.h"
#include "Memory.hpp"
#include "String.hpp"
#include "StringBuilder.h"

namespace String
{
    std::string ToStd(const utf8* str)
    {
        if (str == nullptr)
            return std::string();

        return std::string(str);
    }

    std::string ToUtf8(std::wstring_view src)
    {
#ifdef _WIN32
        int srcLen = static_cast<int>(src.size());
        int sizeReq = WideCharToMultiByte(OpenRCT2::CodePage::UTF8, 0, src.data(), srcLen, nullptr, 0, nullptr, nullptr);
        auto result = std::string(sizeReq, 0);
        WideCharToMultiByte(OpenRCT2::CodePage::UTF8, 0, src.data(), srcLen, result.data(), sizeReq, nullptr, nullptr);
        return result;
#else
// Which constructor to use depends on the size of wchar_t...
// UTF-32 is the default on most POSIX systems; Windows uses UTF-16.
// Unfortunately, we'll have to help the compiler here.
#    if U_SIZEOF_WCHAR_T == 4
        icu::UnicodeString str = icu::UnicodeString::fromUTF32(reinterpret_cast<const UChar32*>(src.data()), src.length());
#    elif U_SIZEOF_WCHAR_T == 2
        std::wstring wstr = std::wstring(src);
        icu::UnicodeString str = icu::UnicodeString(static_cast<const wchar_t*>(wstr.c_str()));
#    else
#        error Unsupported U_SIZEOF_WCHAR_T size
#    endif

        std::string result;
        str.toUTF8String(result);

        return result;
#endif
    }

    std::wstring ToWideChar(std::string_view src)
    {
#ifdef _WIN32
        int srcLen = static_cast<int>(src.size());
        int sizeReq = MultiByteToWideChar(OpenRCT2::CodePage::UTF8, 0, src.data(), srcLen, nullptr, 0);
        auto result = std::wstring(sizeReq, 0);
        MultiByteToWideChar(OpenRCT2::CodePage::UTF8, 0, src.data(), srcLen, result.data(), sizeReq);
        return result;
#else
        icu::UnicodeString str = icu::UnicodeString::fromUTF8(std::string(src));

// Which constructor to use depends on the size of wchar_t...
// UTF-32 is the default on most POSIX systems; Windows uses UTF-16.
// Unfortunately, we'll have to help the compiler here.
#    if U_SIZEOF_WCHAR_T == 4
        size_t length = static_cast<size_t>(str.length());
        std::wstring result(length, '\0');

        UErrorCode status = U_ZERO_ERROR;
        str.toUTF32(reinterpret_cast<UChar32*>(&result[0]), str.length(), status);

#    elif U_SIZEOF_WCHAR_T == 2
        const char16_t* buffer = str.getBuffer();
        std::wstring result = static_cast<wchar_t*>(buffer);

#    else
#        error Unsupported U_SIZEOF_WCHAR_T size
#    endif

        return result;
#endif
    }

    std::string_view ToStringView(const char* ch, size_t maxLen)
    {
        size_t len{};
        for (size_t i = 0; i < maxLen; i++)
        {
            if (ch[i] == '\0')
            {
                break;
            }

            len++;
        }
        return std::string_view(ch, len);
    }

    bool IsNullOrEmpty(const utf8* str)
    {
        return str == nullptr || str[0] == '\0';
    }

    int32_t Compare(const std::string& a, const std::string& b, bool ignoreCase)
    {
        return Compare(a.c_str(), b.c_str(), ignoreCase);
    }

    int32_t Compare(const utf8* a, const utf8* b, bool ignoreCase)
    {
        if (a == b)
            return 0;
        if (a == nullptr)
            a = "";
        if (b == nullptr)
            b = "";
        if (ignoreCase)
        {
            return _stricmp(a, b);
        }

        return strcmp(a, b);
    }

    bool Equals(std::string_view a, std::string_view b, bool ignoreCase)
    {
        if (ignoreCase)
        {
            if (a.size() == b.size())
            {
                for (size_t i = 0; i < a.size(); i++)
                {
                    if (tolower(static_cast<unsigned char>(a[i])) != tolower(static_cast<unsigned char>(b[i])))
                    {
                        return false;
                    }
                }
                return true;
            }

            return false;
        }

        return a == b;
    }

    bool Equals(const std::string& a, const std::string& b, bool ignoreCase)
    {
        if (a.size() != b.size())
            return false;

        if (ignoreCase)
        {
            for (size_t i = 0; i < a.size(); i++)
            {
                auto ai = a[i];
                auto bi = b[i];

                // Only do case insensitive comparison on ASCII characters
                if ((ai & 0x80) != 0 || (bi & 0x80) != 0)
                {
                    if (a[i] != b[i])
                    {
                        return false;
                    }
                }
                else if (tolower(static_cast<unsigned char>(ai)) != tolower(static_cast<unsigned char>(bi)))
                {
                    return false;
                }
            }
        }
        else
        {
            for (size_t i = 0; i < a.size(); i++)
            {
                if (a[i] != b[i])
                {
                    return false;
                }
            }
        }

        return true;
    }

    bool Equals(const utf8* a, const utf8* b, bool ignoreCase)
    {
        if (a == b)
            return true;
        if (a == nullptr || b == nullptr)
            return false;

        if (ignoreCase)
        {
            return _stricmp(a, b) == 0;
        }

        return strcmp(a, b) == 0;
    }

    bool StartsWith(std::string_view str, std::string_view match, bool ignoreCase)
    {
        if (str.size() >= match.size())
        {
            auto view = str.substr(0, match.size());
            return Equals(view, match, ignoreCase);
        }
        return false;
    }

    bool EndsWith(std::string_view str, std::string_view match, bool ignoreCase)
    {
        if (str.size() >= match.size())
        {
            auto view = str.substr(str.size() - match.size());
            return Equals(view, match, ignoreCase);
        }
        return false;
    }

    bool Contains(std::string_view haystack, std::string_view needle, bool ignoreCase)
    {
        if (needle.size() > haystack.size())
            return false;

        if (!ignoreCase)
            return haystack.find(needle) != std::string_view::npos;

        auto end = haystack.size() - needle.size();
        for (size_t start = 0; start <= end; start++)
        {
            auto sub = haystack.substr(start, needle.size());
            if (Equals(sub, needle, ignoreCase))
            {
                return true;
            }
        }
        return false;
    }

    size_t IndexOf(const utf8* str, utf8 match, size_t startIndex)
    {
        const utf8* ch = str + startIndex;
        for (; *ch != '\0'; ch++)
        {
            if (*ch == match)
            {
                return static_cast<size_t>(ch - str);
            }
        }
        return SIZE_MAX;
    }

    ptrdiff_t LastIndexOf(const utf8* str, utf8 match)
    {
        const utf8* lastOccurance = nullptr;
        const utf8* ch = str;
        for (; *ch != '\0'; ch++)
        {
            if (*ch == match)
            {
                lastOccurance = ch;
            }
        }

        if (lastOccurance == nullptr)
        {
            return -1;
        }

        return lastOccurance - str;
    }

    size_t LengthOf(const utf8* str)
    {
        return UTF8Length(str);
    }

    size_t SizeOf(const utf8* str)
    {
        return strlen(str);
    }

    utf8* Set(utf8* buffer, size_t bufferSize, const utf8* src)
    {
        return SafeStrCpy(buffer, src, bufferSize);
    }

    utf8* Set(utf8* buffer, size_t bufferSize, const utf8* src, size_t srcSize)
    {
        utf8* dst = buffer;
        size_t minSize = std::min(bufferSize - 1, srcSize);
        for (size_t i = 0; i < minSize; i++)
        {
            *dst++ = *src;
            if (*src == '\0')
                break;
            src++;
        }
        *dst = '\0';
        return buffer;
    }

    utf8* Append(utf8* buffer, size_t bufferSize, const utf8* src)
    {
        return SafeStrCat(buffer, src, bufferSize);
    }

    utf8* Format(utf8* buffer, size_t bufferSize, const utf8* format, ...)
    {
        va_list args;

        va_start(args, format);
        vsnprintf(buffer, bufferSize, format, args);
        va_end(args);

        // Terminate buffer in case formatted string overflowed
        buffer[bufferSize - 1] = '\0';

        return buffer;
    }

    u8string StdFormat(const utf8* format, ...)
    {
        va_list args;
        va_start(args, format);
        auto result = Format_VA(format, args);
        va_end(args);
        return result;
    }

    u8string Format_VA(const utf8* format, va_list args)
    {
        // When passing no buffer and a size of 0, vsnprintf returns the numbers of chars it would have writte, excluding the
        // null terminator.
        va_list copy;
        va_copy(copy, args);
        auto len = vsnprintf(nullptr, 0, format, copy);
        va_end(copy);

        if (len >= 0)
        {
            // The length returned by vsnprintf does not include the null terminator, but this byte is accounted for when
            // writing to a buffer, so we need to allocate one additional byte to fit the entire string in.
            len++;
            auto buffer = static_cast<utf8*>(alloca(len));

            len = vsnprintf(buffer, len, format, args);

            return u8string(buffer, buffer + len);
        }

        LOG_WARNING("Encoding error occured");
        return u8string{};
    }

    utf8* AppendFormat(utf8* buffer, size_t bufferSize, const utf8* format, ...)
    {
        utf8* dst = buffer;
        size_t i;
        for (i = 0; i < bufferSize; i++)
        {
            if (*dst == '\0')
                break;
            dst++;
        }

        if (i < bufferSize - 1)
        {
            va_list args;
            va_start(args, format);
            vsnprintf(dst, bufferSize - i - 1, format, args);
            va_end(args);

            // Terminate buffer in case formatted string overflowed
            buffer[bufferSize - 1] = '\0';
        }

        return buffer;
    }

    std::vector<std::string> Split(std::string_view s, std::string_view delimiter)
    {
        if (delimiter.empty())
        {
            throw std::invalid_argument(nameof(delimiter) " can not be empty.");
        }

        std::vector<std::string> results;
        if (!s.empty())
        {
            size_t index = 0;
            size_t nextIndex;
            do
            {
                nextIndex = s.find(delimiter, index);
                if (nextIndex == std::string::npos)
                {
                    results.emplace_back(s.substr(index));
                }
                else
                {
                    results.emplace_back(s.substr(index, nextIndex - index));
                }
                index = nextIndex + delimiter.size();
            } while (nextIndex != SIZE_MAX);
        }
        return results;
    }

    utf8* SkipBOM(utf8* buffer)
    {
        return const_cast<utf8*>(SkipBOM(static_cast<const utf8*>(buffer)));
    }

    const utf8* SkipBOM(const utf8* buffer)
    {
        if (static_cast<uint8_t>(buffer[0]) == 0xEF && static_cast<uint8_t>(buffer[1]) == 0xBB
            && static_cast<uint8_t>(buffer[2]) == 0xBF)
        {
            return buffer + 3;
        }
        return buffer;
    }

    size_t GetCodepointLength(codepoint_t codepoint)
    {
        return UTF8GetCodepointLength(codepoint);
    }

    codepoint_t GetNextCodepoint(utf8* ptr, utf8** nextPtr)
    {
        return GetNextCodepoint(static_cast<const utf8*>(ptr), const_cast<const utf8**>(nextPtr));
    }

    codepoint_t GetNextCodepoint(const utf8* ptr, const utf8** nextPtr)
    {
        return UTF8GetNext(ptr, nextPtr);
    }

    utf8* WriteCodepoint(utf8* dst, codepoint_t codepoint)
    {
        return UTF8WriteCodepoint(dst, codepoint);
    }

    void AppendCodepoint(std::string& str, codepoint_t codepoint)
    {
        char buffer[8]{};
        UTF8WriteCodepoint(buffer, codepoint);
        str.append(buffer);
    }

    bool IsWhiteSpace(codepoint_t codepoint)
    {
        // 0x3000 is the 'ideographic space', a 'fullwidth' character used in CJK languages.
        return iswspace(static_cast<wchar_t>(codepoint)) || codepoint == 0x3000;
    }

    utf8* Trim(utf8* str)
    {
        utf8* firstNonWhitespace = nullptr;

        codepoint_t codepoint;
        utf8* ch = str;
        utf8* nextCh;
        while ((codepoint = GetNextCodepoint(ch, &nextCh)) != '\0')
        {
            if (codepoint <= WCHAR_MAX && !IsWhiteSpace(codepoint))
            {
                if (firstNonWhitespace == nullptr)
                {
                    firstNonWhitespace = ch;
                }
            }
            ch = nextCh;
        }

        if (firstNonWhitespace != nullptr && firstNonWhitespace != str)
        {
            // Take multibyte characters into account: use the last byte of the
            // current character.
            size_t newStringSize = (nextCh - 1) - firstNonWhitespace;

#ifdef DEBUG
            size_t currentStringSize = String::SizeOf(str);
            Guard::Assert(newStringSize < currentStringSize, GUARD_LINE);
#endif

            std::memmove(str, firstNonWhitespace, newStringSize);
            str[newStringSize] = '\0';
        }
        else
        {
            *ch = '\0';
        }

        return str;
    }

    const utf8* TrimStart(const utf8* str)
    {
        codepoint_t codepoint;
        const utf8* ch = str;
        const utf8* nextCh;
        while ((codepoint = GetNextCodepoint(ch, &nextCh)) != '\0')
        {
            if (codepoint <= WCHAR_MAX && !IsWhiteSpace(codepoint))
            {
                return ch;
            }
            ch = nextCh;
        }
        // String is all whitespace
        return ch;
    }

    utf8* TrimStart(utf8* buffer, size_t bufferSize, const utf8* src)
    {
        return String::Set(buffer, bufferSize, TrimStart(src));
    }

    std::string TrimStart(const std::string& s)
    {
        const utf8* trimmed = TrimStart(s.c_str());
        return std::string(trimmed);
    }

    std::string Trim(const std::string& s)
    {
        codepoint_t codepoint;
        const utf8* ch = s.c_str();
        const utf8* nextCh;
        const utf8* startSubstr = nullptr;
        const utf8* endSubstr = nullptr;
        while ((codepoint = GetNextCodepoint(ch, &nextCh)) != '\0')
        {
            bool isWhiteSpace = codepoint <= WCHAR_MAX && IsWhiteSpace(codepoint);
            if (!isWhiteSpace)
            {
                if (startSubstr == nullptr)
                {
                    startSubstr = ch;
                }

                // Take multibyte characters into account: move pointer towards
                // the last byte of the current character.
                endSubstr = nextCh - 1;
            }
            ch = nextCh;
        }

        if (startSubstr == nullptr)
        {
            // String is all whitespace
            return std::string();
        }

        size_t stringLength = endSubstr - startSubstr + 1;
        return std::string(startSubstr, stringLength);
    }

#ifndef _WIN32
    static const char* GetIcuCodePage(int32_t codePage)
    {
        switch (codePage)
        {
            case OpenRCT2::CodePage::CP_932:
                return "windows-932";

            case OpenRCT2::CodePage::CP_936:
                return "GB2312";

            case OpenRCT2::CodePage::CP_949:
                return "windows-949";

            case OpenRCT2::CodePage::CP_950:
                return "big5";

            case OpenRCT2::CodePage::CP_1252:
                return "windows-1252";

            case OpenRCT2::CodePage::UTF8:
                return "utf-8";

            default:
                throw std::runtime_error("Unsupported code page: " + std::to_string(codePage));
        }
    }
#endif

    std::string ConvertToUtf8(std::string_view src, int32_t srcCodePage)
    {
#ifdef _WIN32
        // Convert from source code page to UTF-16
        std::wstring u16;
        {
            int srcLen = static_cast<int>(src.size());
            int sizeReq = MultiByteToWideChar(srcCodePage, 0, src.data(), srcLen, nullptr, 0);
            u16 = std::wstring(sizeReq, 0);
            MultiByteToWideChar(srcCodePage, 0, src.data(), srcLen, u16.data(), sizeReq);
        }

        // Convert from UTF-16 to destination code page
        std::string dst;
        {
            int srcLen = static_cast<int>(u16.size());
            int sizeReq = WideCharToMultiByte(OpenRCT2::CodePage::UTF8, 0, u16.data(), srcLen, nullptr, 0, nullptr, nullptr);
            dst = std::string(sizeReq, 0);
            WideCharToMultiByte(OpenRCT2::CodePage::UTF8, 0, u16.data(), srcLen, dst.data(), sizeReq, nullptr, nullptr);
        }

        return dst;
#else
        const char* codepage = GetIcuCodePage(srcCodePage);
        icu::UnicodeString convertString(src.data(), codepage);

        std::string result;
        convertString.toUTF8String(result);

        return result;
#endif
    }

    std::string ToUpper(std::string_view src)
    {
#ifdef _WIN32
        auto srcW = ToWideChar(src);

        // Measure how long the destination needs to be
        auto requiredSize = LCMapStringEx(
            LOCALE_NAME_USER_DEFAULT, LCMAP_UPPERCASE | LCMAP_LINGUISTIC_CASING, srcW.c_str(), static_cast<int>(srcW.length()),
            nullptr, 0, nullptr, nullptr, 0);

        auto dstW = std::wstring();
        dstW.resize(requiredSize);

        // Transform the string
        auto result = LCMapStringEx(
            LOCALE_NAME_USER_DEFAULT, LCMAP_UPPERCASE | LCMAP_LINGUISTIC_CASING, srcW.c_str(), static_cast<int>(srcW.length()),
            dstW.data(), static_cast<int>(dstW.length()), nullptr, nullptr, 0);
        if (result == 0)
        {
            // Check the error
            auto error = GetLastError();
            LOG_WARNING("LCMapStringEx failed with %d", error);
            return std::string(src);
        }

        return String::ToUtf8(dstW);
#else
        icu::UnicodeString str = icu::UnicodeString::fromUTF8(std::string(src));
        str.toUpper();

        std::string res;
        str.toUTF8String(res);

        return res;
#endif
    }

    std::string_view UTF8Truncate(std::string_view v, size_t size)
    {
        auto trunc = v.substr(0, size);
        for (size_t i = 0; i < trunc.size();)
        {
            auto length = UTF8GetCodePointSize(trunc.substr(i, trunc.size()));
            if (!length.has_value())
            {
                return trunc.substr(0, i);
            }
            i += length.value();
        }

        return trunc;
    }

    std::string URLEncode(std::string_view value)
    {
        std::ostringstream escaped;
        escaped.fill('0');
        escaped << std::hex;

        for (auto c : value)
        {
            // Keep alphanumeric and other accepted characters intact
            if (std::isalnum(static_cast<unsigned char>(c)) || c == '-' || c == '_' || c == '.' || c == '~')
            {
                escaped << c;
            }
            else
            {
                // Any other characters are percent-escaped
                escaped << std::uppercase;
                escaped << '%' << std::setw(2) << int32_t(static_cast<unsigned char>(c));
                escaped << std::nouppercase;
            }
        }

        return escaped.str();
    }
} // namespace String

char32_t CodepointView::iterator::GetNextCodepoint(const char* ch, const char** next)
{
    return UTF8GetNext(ch, next);
}