/*
  OpenMW - The completely unofficial reimplementation of Morrowind
  Copyright (C) 2008-2010  Nicolay Korslund
  Email: < korslund@gmail.com >
  WWW: http://openmw.sourceforge.net/

  This file (compressedbsafile.cpp) is part of the OpenMW package.

  OpenMW is distributed as free software: you can redistribute it
  and/or modify it under the terms of the GNU General Public License
  version 3, as published by the Free Software Foundation.

  This program is distributed in the hope that it will be useful, but
  WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  General Public License for more details.

  You should have received a copy of the GNU General Public License
  version 3 along with this program. If not, see
  http://www.gnu.org/licenses/ .

  Compressed BSA stuff added by cc9cii 2018

 */
#include "compressedbsafile.hpp"

#include <cassert>
#include <filesystem>
#include <fstream>

#include <lz4frame.h>

#include <boost/iostreams/copy.hpp>
#include <boost/iostreams/filtering_streambuf.hpp>

#if defined(_MSC_VER)
#pragma warning(push)
#pragma warning(disable : 4706)
#include <boost/iostreams/filter/zlib.hpp>
#pragma warning(pop)
#else
#include <boost/iostreams/filter/zlib.hpp>
#endif

#include <boost/iostreams/device/array.hpp>
#include <components/bsa/memorystream.hpp>
#include <components/files/constrainedfilestream.hpp>
#include <components/files/conversion.hpp>
#include <components/misc/strings/lower.hpp>

namespace Bsa
{
    // special marker for invalid records,
    // equal to max uint32_t value
    const uint32_t CompressedBSAFile::sInvalidOffset = std::numeric_limits<uint32_t>::max();

    // bit marking compression on file size
    const uint32_t CompressedBSAFile::sCompressedFlag = 1u << 30u;

    CompressedBSAFile::FileRecord::FileRecord()
        : size(0)
        , offset(sInvalidOffset)
    {
    }

    bool CompressedBSAFile::FileRecord::isValid() const
    {
        return offset != sInvalidOffset;
    }

    bool CompressedBSAFile::FileRecord::isCompressed(bool bsaCompressedByDefault) const
    {
        bool recordCompressionFlagEnabled = ((size & sCompressedFlag) == sCompressedFlag);

        // record is compressed when:
        //- bsaCompressedByDefault flag is set and 30th bit is NOT set, or
        //- bsaCompressedByDefault flag is NOT set and 30th bit is set
        // record is NOT compressed when:
        //- bsaCompressedByDefault flag is NOT set and 30th bit is NOT set, or
        //- bsaCompressedByDefault flag is set and 30th bit is set
        return (bsaCompressedByDefault != recordCompressionFlagEnabled);
    }

    std::uint32_t CompressedBSAFile::FileRecord::getSizeWithoutCompressionFlag() const
    {
        return size & (~sCompressedFlag);
    }

    void CompressedBSAFile::getBZString(std::string& str, std::istream& filestream)
    {
        char size = 0;
        filestream.read(&size, 1);

        auto buf = std::vector<char>(size);
        filestream.read(buf.data(), size);

        if (buf[size - 1] != 0)
        {
            str.assign(buf.data(), size);
            if (str.size() != ((size_t)size))
            {
                fail("getBZString string size mismatch");
            }
        }
        else
        {
            str.assign(buf.data(), size - 1); // don't copy null terminator
            if (str.size() != ((size_t)size - 1))
            {
                fail("getBZString string size mismatch (null terminator)");
            }
        }
    }

    CompressedBSAFile::CompressedBSAFile()
        : mCompressedByDefault(false)
        , mEmbeddedFileNames(false)
    {
    }

    CompressedBSAFile::~CompressedBSAFile() = default;

    /// Read header information from the input source
    void CompressedBSAFile::readHeader()
    {
        assert(!mIsLoaded);

        std::ifstream input(mFilepath, std::ios_base::binary);

        // Total archive size
        std::streamoff fsize = 0;
        if (input.seekg(0, std::ios_base::end))
        {
            fsize = input.tellg();
            input.seekg(0);
        }

        if (fsize < 36) // header is 36 bytes
            fail("File too small to be a valid BSA archive");

        // Get essential header numbers
        // size_t dirsize, filenum;
        std::uint32_t archiveFlags, folderCount, totalFileNameLength;
        {
            // First 36 bytes
            std::uint32_t header[9];

            input.read(reinterpret_cast<char*>(header), 36);

            if (header[0] != 0x00415342) /*"BSA\x00"*/
                fail("Unrecognized compressed BSA format");
            mVersion = header[1];
            if (mVersion != 0x67 /*TES4*/ && mVersion != 0x68 /*FO3, FNV, TES5*/ && mVersion != 0x69 /*SSE*/)
                fail("Unrecognized compressed BSA version");

            // header[2] is offset, should be 36 = 0x24 which is the size of the header

            // Oblivion - Meshes.bsa
            //
            // 0111 1000 0111 = 0x0787
            //  ^^^ ^     ^^^
            //  ||| |     ||+-- has names for dirs  (mandatory?)
            //  ||| |     |+--- has names for files (mandatory?)
            //  ||| |     +---- files are compressed by default
            //  ||| |
            //  ||| +---------- unknown (TES5: retain strings during startup)
            //  ||+------------ unknown (TES5: embedded file names)
            //  |+------------- unknown
            //  +-------------- unknown
            //
            archiveFlags = header[3];
            folderCount = header[4];
            // header[5] - fileCount
            // totalFolderNameLength = header[6];
            totalFileNameLength = header[7];
            // header[8]; // fileFlags : an opportunity to optimize here

            mCompressedByDefault = (archiveFlags & 0x4) != 0;
            if (mVersion == 0x68 || mVersion == 0x69) /*FO3, FNV, TES5, SSE*/
                mEmbeddedFileNames = (archiveFlags & 0x100) != 0;
        }

        // folder records
        std::uint64_t hash;
        FolderRecord fr;
        for (std::uint32_t i = 0; i < folderCount; ++i)
        {
            input.read(reinterpret_cast<char*>(&hash), 8);
            input.read(reinterpret_cast<char*>(&fr.count), 4); // not sure purpose of count
            if (mVersion == 0x69) // SSE
            {
                std::uint32_t unknown;
                input.read(reinterpret_cast<char*>(&unknown), 4);
                input.read(reinterpret_cast<char*>(&fr.offset), 8);
            }
            else
                input.read(reinterpret_cast<char*>(&fr.offset), 4); // not sure purpose of offset

            auto lb = mFolders.lower_bound(hash);
            if (lb != mFolders.end() && !(mFolders.key_comp()(hash, lb->first)))
                fail("Archive found duplicate folder name hash");
            else
                mFolders.insert(lb, std::pair<std::uint64_t, FolderRecord>(hash, fr));
        }

        // file record blocks
        std::uint64_t fileHash;
        FileRecord file;

        std::string folder;
        std::uint64_t folderHash;
        if ((archiveFlags & 0x1) == 0)
            folderCount = 1; // TODO: not tested - unit test necessary

        mFiles.clear();
        std::vector<std::string> fullPaths;

        for (std::uint32_t i = 0; i < folderCount; ++i)
        {
            if ((archiveFlags & 0x1) != 0)
                getBZString(folder, input);

            folderHash = generateHash(folder, {});

            auto iter = mFolders.find(folderHash);
            if (iter == mFolders.end())
                fail("Archive folder name hash not found");

            for (std::uint32_t j = 0; j < iter->second.count; ++j)
            {
                input.read(reinterpret_cast<char*>(&fileHash), 8);
                input.read(reinterpret_cast<char*>(&file.size), 4);
                input.read(reinterpret_cast<char*>(&file.offset), 4);

                auto lb = iter->second.files.lower_bound(fileHash);
                if (lb != iter->second.files.end() && !(iter->second.files.key_comp()(fileHash, lb->first)))
                    fail("Archive found duplicate file name hash");

                iter->second.files.insert(lb, std::pair<std::uint64_t, FileRecord>(fileHash, file));

                FileStruct fileStruct{};
                fileStruct.fileSize = file.getSizeWithoutCompressionFlag();
                fileStruct.offset = file.offset;
                mFiles.push_back(fileStruct);

                fullPaths.push_back(folder);
            }
        }

        // file record blocks
        if ((archiveFlags & 0x2) != 0)
        {
            mStringBuf.resize(totalFileNameLength);
            input.read(mStringBuf.data(), mStringBuf.size()); // TODO: maybe useful in building a lookup map?
        }

        size_t mStringBuffOffset = 0;
        size_t totalStringsSize = 0;
        for (std::uint32_t fileIndex = 0; fileIndex < mFiles.size(); ++fileIndex)
        {

            if (mStringBuffOffset >= totalFileNameLength)
            {
                fail("Corrupted names record in BSA file");
            }

            // The vector guarantees that its elements occupy contiguous memory
            mFiles[fileIndex].setNameInfos(mStringBuffOffset, &mStringBuf);

            fullPaths.at(fileIndex) += "\\" + std::string(mStringBuf.data() + mStringBuffOffset);

            while (mStringBuffOffset < totalFileNameLength)
            {
                if (mStringBuf[mStringBuffOffset] != '\0')
                {
                    mStringBuffOffset++;
                }
                else
                {
                    mStringBuffOffset++;
                    break;
                }
            }
            // we want to keep one more 0 character at the end of each string
            totalStringsSize += fullPaths.at(fileIndex).length() + 1u;
        }
        mStringBuf.resize(totalStringsSize);

        mStringBuffOffset = 0;
        for (std::uint32_t fileIndex = 0u; fileIndex < mFiles.size(); fileIndex++)
        {
            size_t stringLength = fullPaths.at(fileIndex).length();

            std::copy(fullPaths.at(fileIndex).c_str(),
                // plus 1 because we also want to copy 0 at the end of the string
                fullPaths.at(fileIndex).c_str() + stringLength + 1u, mStringBuf.data() + mStringBuffOffset);

            mFiles[fileIndex].setNameInfos(mStringBuffOffset, &mStringBuf);

            mStringBuffOffset += stringLength + 1u;
        }

        if (mStringBuffOffset != mStringBuf.size())
        {
            fail("Could not resolve names of files in BSA file");
        }

        convertCompressedSizesToUncompressed();
        mIsLoaded = true;
    }

    CompressedBSAFile::FileRecord CompressedBSAFile::getFileRecord(const std::string& str) const
    {
        for (const auto c : str)
        {
            if (((static_cast<unsigned>(c) >> 7U) & 1U) != 0U)
            {
                fail("File record " + str + " contains unicode characters, refusing to load.");
            }
        }

#ifdef _WIN32
        const auto& path = str;
#else
        // Force-convert the path into something UNIX can handle first
        // to make sure std::filesystem::path doesn't think the entire path is the filename on Linux
        // and subsequently purge it to determine the file folder.
        std::string path = str;
        std::replace(path.begin(), path.end(), '\\', '/');
#endif

        const auto p = std::filesystem::path{ path }; // Purposefully damage Unicode strings.
        const auto stem = p.stem();
        const auto ext = p.extension().string(); // Purposefully damage Unicode strings.

        std::uint64_t folderHash = generateHash(p.parent_path(), {});

        auto it = mFolders.find(folderHash);
        if (it == mFolders.end())
            return FileRecord(); // folder not found, return default which has offset of sInvalidOffset

        std::uint64_t fileHash = generateHash(stem, ext);
        auto iter = it->second.files.find(fileHash);
        if (iter == it->second.files.end())
            return FileRecord(); // file not found, return default which has offset of sInvalidOffset

        return iter->second;
    }

    Files::IStreamPtr CompressedBSAFile::getFile(const FileStruct* file)
    {
        FileRecord fileRec = getFileRecord(file->name());
        if (!fileRec.isValid())
        {
            fail("File not found: " + std::string(file->name()));
        }
        return getFile(fileRec);
    }

    void CompressedBSAFile::addFile(const std::string& filename, std::istream& file)
    {
        assert(false); // not implemented yet
        fail("Add file is not implemented for compressed BSA: " + filename);
    }

    Files::IStreamPtr CompressedBSAFile::getFile(const char* file)
    {
        FileRecord fileRec = getFileRecord(file);
        if (!fileRec.isValid())
        {
            fail("File not found: " + std::string(file));
        }
        return getFile(fileRec);
    }

    Files::IStreamPtr CompressedBSAFile::getFile(const FileRecord& fileRecord)
    {
        size_t size = fileRecord.getSizeWithoutCompressionFlag();
        size_t uncompressedSize = size;
        bool compressed = fileRecord.isCompressed(mCompressedByDefault);
        Files::IStreamPtr streamPtr = Files::openConstrainedFileStream(mFilepath, fileRecord.offset, size);
        std::istream* fileStream = streamPtr.get();
        if (mEmbeddedFileNames)
        {
            // Skip over the embedded file name
            char length = 0;
            fileStream->read(&length, 1);
            fileStream->ignore(length);
            size -= length + sizeof(char);
        }
        if (compressed)
        {
            fileStream->read(reinterpret_cast<char*>(&uncompressedSize), sizeof(uint32_t));
            size -= sizeof(uint32_t);
        }
        auto memoryStreamPtr = std::make_unique<MemoryInputStream>(uncompressedSize);

        if (compressed)
        {
            if (mVersion != 0x69) // Non-SSE: zlib
            {
                boost::iostreams::filtering_streambuf<boost::iostreams::input> inputStreamBuf;
                inputStreamBuf.push(boost::iostreams::zlib_decompressor());
                inputStreamBuf.push(*fileStream);

                boost::iostreams::basic_array_sink<char> sr(memoryStreamPtr->getRawData(), uncompressedSize);
                boost::iostreams::copy(inputStreamBuf, sr);
            }
            else // SSE: lz4
            {
                auto buffer = std::vector<char>(size);
                fileStream->read(buffer.data(), size);
                LZ4F_decompressionContext_t context = nullptr;
                LZ4F_createDecompressionContext(&context, LZ4F_VERSION);
                LZ4F_decompressOptions_t options = {};
                LZ4F_errorCode_t errorCode = LZ4F_decompress(
                    context, memoryStreamPtr->getRawData(), &uncompressedSize, buffer.data(), &size, &options);
                if (LZ4F_isError(errorCode))
                    fail("LZ4 decompression error (file " + Files::pathToUnicodeString(mFilepath)
                        + "): " + LZ4F_getErrorName(errorCode));
                errorCode = LZ4F_freeDecompressionContext(context);
                if (LZ4F_isError(errorCode))
                    fail("LZ4 decompression error (file " + Files::pathToUnicodeString(mFilepath)
                        + "): " + LZ4F_getErrorName(errorCode));
            }
        }
        else
        {
            fileStream->read(memoryStreamPtr->getRawData(), size);
        }

        return std::make_unique<Files::StreamWithBuffer<MemoryInputStream>>(std::move(memoryStreamPtr));
    }

    // mFiles used by OpenMW expects uncompressed sizes
    void CompressedBSAFile::convertCompressedSizesToUncompressed()
    {
        for (auto& mFile : mFiles)
        {
            const FileRecord& fileRecord = getFileRecord(mFile.name());
            if (!fileRecord.isValid())
            {
                fail("Could not find file " + std::string(mFile.name()) + " in BSA");
            }

            if (!fileRecord.isCompressed(mCompressedByDefault))
            {
                // no need to fix fileSize in mFiles - uncompressed size already set
                continue;
            }

            Files::IStreamPtr dataBegin = Files::openConstrainedFileStream(
                mFilepath, fileRecord.offset, fileRecord.getSizeWithoutCompressionFlag());

            if (mEmbeddedFileNames)
            {
                std::string embeddedFileName;
                getBZString(embeddedFileName, *(dataBegin.get()));
            }

            dataBegin->read(reinterpret_cast<char*>(&(mFile.fileSize)), sizeof(mFile.fileSize));
        }
    }

    std::uint64_t CompressedBSAFile::generateHash(const std::filesystem::path& stem, std::string extension)
    {
        auto str = stem.u8string();
        size_t len = str.length();
        if (len == 0)
            return 0;
        std::replace(str.begin(), str.end(), '/', '\\');
        Misc::StringUtils::lowerCaseInPlace(str);
        uint64_t result = str[len - 1] | (len >= 3 ? (str[len - 2] << 8) : 0) | (len << 16) | (str[0] << 24);
        if (len >= 4)
        {
            uint32_t hash = 0;
            for (size_t i = 1; i <= len - 3; ++i)
                hash = hash * 0x1003f + str[i];
            result += static_cast<uint64_t>(hash) << 32;
        }
        if (extension.empty())
            return result;
        Misc::StringUtils::lowerCaseInPlace(extension);
        if (extension == ".kf")
            result |= 0x80;
        else if (extension == ".nif")
            result |= 0x8000;
        else if (extension == ".dds")
            result |= 0x8080;
        else if (extension == ".wav")
            result |= 0x80000000;
        uint32_t hash = 0;
        for (const auto& c : extension)
            hash = hash * 0x1003f + c;
        result += static_cast<uint64_t>(hash) << 32;
        return result;
    }

} // namespace Bsa