Replace MPQ runtime with loose file asset system

Extract assets from MPQ archives into organized loose files indexed by
manifest.json, enabling fully parallel reads without StormLib serialization.
Add asset_extract and blp_convert tools, PNG texture override support.
This commit is contained in:
Kelsi 2026-02-12 20:32:14 -08:00
parent 5fda1a3157
commit aa16a687c2
16 changed files with 1427 additions and 101 deletions

View file

@ -3,8 +3,11 @@
#include "core/memory_monitor.hpp"
#include <algorithm>
#include <cstdlib>
#include <filesystem>
#include <limits>
#include "stb_image.h"
namespace wowee {
namespace pipeline {
@ -40,26 +43,36 @@ bool AssetManager::initialize(const std::string& dataPath_) {
dataPath = dataPath_;
LOG_INFO("Initializing asset manager with data path: ", dataPath);
// Initialize MPQ manager
if (!mpqManager.initialize(dataPath)) {
LOG_ERROR("Failed to initialize MPQ manager");
setupFileCacheBudget();
std::string manifestPath = dataPath + "/manifest.json";
if (!std::filesystem::exists(manifestPath)) {
LOG_ERROR("manifest.json not found in: ", dataPath);
LOG_ERROR("Run asset_extract to extract MPQ archives first");
return false;
}
// Set dynamic file cache budget based on available RAM.
// Bias toward MPQ decompressed-file caching to minimize runtime read/decompress stalls.
if (!manifest_.load(manifestPath)) {
LOG_ERROR("Failed to load manifest");
return false;
}
initialized = true;
LOG_INFO("Asset manager initialized: ", manifest_.getEntryCount(),
" files indexed (file cache: ", fileCacheBudget / (1024 * 1024), " MB)");
return true;
}
void AssetManager::setupFileCacheBudget() {
auto& memMonitor = core::MemoryMonitor::getInstance();
size_t recommendedBudget = memMonitor.getRecommendedCacheBudget();
size_t dynamicBudget = (recommendedBudget * 3) / 4; // 75% of global cache budget
size_t dynamicBudget = (recommendedBudget * 3) / 4;
// The MemoryMonitor recommendation is intentionally aggressive; without a cap, large patch MPQs
// can cause the decompressed-file cache to balloon into tens/hundreds of GB and push the OS into swap.
// Provide env overrides and clamp to a safe default.
const size_t envFixedMB = parseEnvSizeMB("WOWEE_FILE_CACHE_MB");
const size_t envMaxMB = parseEnvSizeMB("WOWEE_FILE_CACHE_MAX_MB");
const size_t minBudgetBytes = 256ull * 1024ull * 1024ull; // 256 MB
const size_t defaultMaxBudgetBytes = 32768ull * 1024ull * 1024ull; // 32 GB
const size_t minBudgetBytes = 256ull * 1024ull * 1024ull;
const size_t defaultMaxBudgetBytes = 32768ull * 1024ull * 1024ull;
const size_t maxBudgetBytes = (envMaxMB > 0)
? (envMaxMB * 1024ull * 1024ull)
: defaultMaxBudgetBytes;
@ -74,13 +87,6 @@ bool AssetManager::initialize(const std::string& dataPath_) {
} else {
fileCacheBudget = std::clamp(dynamicBudget, minBudgetBytes, maxBudgetBytes);
}
initialized = true;
LOG_INFO("Asset manager initialized (file cache: ",
fileCacheBudget / (1024 * 1024), " MB, recommended=",
recommendedBudget / (1024 * 1024), " MB, max=",
maxBudgetBytes / (1024 * 1024), " MB)");
return true;
}
void AssetManager::shutdown() {
@ -90,7 +96,6 @@ void AssetManager::shutdown() {
LOG_INFO("Shutting down asset manager");
// Log cache statistics
if (fileCacheHits + fileCacheMisses > 0) {
float hitRate = (float)fileCacheHits / (fileCacheHits + fileCacheMisses) * 100.0f;
LOG_INFO("File cache stats: ", fileCacheHits, " hits, ", fileCacheMisses, " misses (",
@ -98,8 +103,6 @@ void AssetManager::shutdown() {
}
clearCache();
mpqManager.shutdown();
initialized = false;
}
@ -109,19 +112,22 @@ BLPImage AssetManager::loadTexture(const std::string& path) {
return BLPImage();
}
// Normalize path
std::string normalizedPath = normalizePath(path);
LOG_DEBUG("Loading texture: ", normalizedPath);
// Route through readFile() so decompressed MPQ bytes participate in file cache.
// Check for PNG override
BLPImage pngImage = tryLoadPngOverride(normalizedPath);
if (pngImage.isValid()) {
return pngImage;
}
std::vector<uint8_t> blpData = readFile(normalizedPath);
if (blpData.empty()) {
LOG_WARNING("Texture not found: ", normalizedPath);
return BLPImage();
}
// Load BLP
BLPImage image = BLPLoader::load(blpData);
if (!image.isValid()) {
LOG_ERROR("Failed to load texture: ", normalizedPath);
@ -132,13 +138,47 @@ BLPImage AssetManager::loadTexture(const std::string& path) {
return image;
}
BLPImage AssetManager::tryLoadPngOverride(const std::string& normalizedPath) const {
if (normalizedPath.size() < 4) return BLPImage();
std::string ext = normalizedPath.substr(normalizedPath.size() - 4);
if (ext != ".blp") return BLPImage();
std::string fsPath = manifest_.resolveFilesystemPath(normalizedPath);
if (fsPath.empty()) return BLPImage();
// Replace .blp/.BLP extension with .png
std::string pngPath = fsPath.substr(0, fsPath.size() - 4) + ".png";
if (!LooseFileReader::fileExists(pngPath)) {
return BLPImage();
}
int w, h, channels;
unsigned char* pixels = stbi_load(pngPath.c_str(), &w, &h, &channels, 4);
if (!pixels) {
LOG_WARNING("PNG override exists but failed to load: ", pngPath);
return BLPImage();
}
BLPImage image;
image.width = w;
image.height = h;
image.channels = 4;
image.format = BLPFormat::BLP2;
image.compression = BLPCompression::ARGB8888;
image.data.assign(pixels, pixels + (w * h * 4));
stbi_image_free(pixels);
LOG_INFO("PNG override loaded: ", pngPath, " (", w, "x", h, ")");
return image;
}
std::shared_ptr<DBCFile> AssetManager::loadDBC(const std::string& name) {
if (!initialized) {
LOG_ERROR("AssetManager not initialized");
return nullptr;
}
// Check cache first
auto it = dbcCache.find(name);
if (it != dbcCache.end()) {
LOG_DEBUG("DBC already loaded (cached): ", name);
@ -147,28 +187,20 @@ std::shared_ptr<DBCFile> AssetManager::loadDBC(const std::string& name) {
LOG_DEBUG("Loading DBC: ", name);
// Construct DBC path (DBFilesClient directory)
std::string dbcPath = "DBFilesClient\\" + name;
// Read DBC file from MPQ (must hold readMutex — StormLib is not thread-safe)
std::vector<uint8_t> dbcData;
{
std::lock_guard<std::mutex> lock(readMutex);
dbcData = mpqManager.readFile(dbcPath);
}
std::vector<uint8_t> dbcData = readFile(dbcPath);
if (dbcData.empty()) {
LOG_WARNING("DBC not found: ", dbcPath);
return nullptr;
}
// Load DBC
auto dbc = std::make_shared<DBCFile>();
if (!dbc->load(dbcData)) {
LOG_ERROR("Failed to load DBC: ", dbcPath);
return nullptr;
}
// Cache the DBC
dbcCache[name] = dbc;
LOG_INFO("Loaded DBC: ", name, " (", dbc->getRecordCount(), " records)");
@ -187,57 +219,45 @@ bool AssetManager::fileExists(const std::string& path) const {
if (!initialized) {
return false;
}
std::lock_guard<std::mutex> lock(readMutex);
return mpqManager.fileExists(normalizePath(path));
return manifest_.hasEntry(normalizePath(path));
}
std::vector<uint8_t> AssetManager::readFile(const std::string& path) const {
if (!initialized) {
return std::vector<uint8_t>();
return {};
}
std::string normalized = normalizePath(path);
// Check cache first
{
std::lock_guard<std::mutex> cacheLock(cacheMutex);
// Check cache first
auto it = fileCache.find(normalized);
if (it != fileCache.end()) {
// Cache hit - update access time and return cached data
it->second.lastAccessTime = ++fileCacheAccessCounter;
fileCacheHits++;
return it->second.data;
}
}
// Cache miss path: serialize MPQ reads. Before reading, re-check cache while holding
// readMutex so only one thread performs decompression per hot path at a time.
std::vector<uint8_t> data;
{
std::lock_guard<std::mutex> readLock(readMutex);
{
std::lock_guard<std::mutex> cacheLock(cacheMutex);
auto it = fileCache.find(normalized);
if (it != fileCache.end()) {
it->second.lastAccessTime = ++fileCacheAccessCounter;
fileCacheHits++;
return it->second.data;
}
fileCacheMisses++;
}
data = mpqManager.readFile(normalized);
// Read from filesystem (fully parallel, no serialization needed)
std::string fsPath = manifest_.resolveFilesystemPath(normalized);
if (fsPath.empty()) {
return {};
}
auto data = LooseFileReader::readFile(fsPath);
if (data.empty()) {
return data; // File not found
LOG_WARNING("Manifest entry exists but file unreadable: ", fsPath);
return data;
}
// Add to cache if within budget
size_t fileSize = data.size();
if (fileSize > 0 && fileSize < fileCacheBudget / 2) { // Don't cache files > 50% of budget (very aggressive)
if (fileSize > 0 && fileSize < fileCacheBudget / 2) {
std::lock_guard<std::mutex> cacheLock(cacheMutex);
// Evict old entries if needed (LRU)
while (fileCacheTotalBytes + fileSize > fileCacheBudget && !fileCache.empty()) {
// Find least recently used entry
auto lru = fileCache.begin();
for (auto it = fileCache.begin(); it != fileCache.end(); ++it) {
if (it->second.lastAccessTime < lru->second.lastAccessTime) {
@ -248,7 +268,6 @@ std::vector<uint8_t> AssetManager::readFile(const std::string& path) const {
fileCache.erase(lru);
}
// Add new entry
CachedFile cached;
cached.data = data;
cached.lastAccessTime = ++fileCacheAccessCounter;
@ -261,18 +280,16 @@ std::vector<uint8_t> AssetManager::readFile(const std::string& path) const {
std::vector<uint8_t> AssetManager::readFileOptional(const std::string& path) const {
if (!initialized) {
return std::vector<uint8_t>();
return {};
}
// Avoid MPQManager missing-file warnings for expected probe misses.
if (!fileExists(path)) {
return std::vector<uint8_t>();
return {};
}
return readFile(path);
}
void AssetManager::clearCache() {
std::scoped_lock lock(readMutex, cacheMutex);
std::lock_guard<std::mutex> lock(cacheMutex);
dbcCache.clear();
fileCache.clear();
fileCacheTotalBytes = 0;
@ -282,13 +299,9 @@ void AssetManager::clearCache() {
std::string AssetManager::normalizePath(const std::string& path) const {
std::string normalized = path;
// Convert forward slashes to backslashes (WoW uses backslashes)
std::replace(normalized.begin(), normalized.end(), '/', '\\');
// Lowercase for case-insensitive cache keys (improves hit rate across mixed-case callers).
std::transform(normalized.begin(), normalized.end(), normalized.begin(),
[](unsigned char c) { return static_cast<char>(std::tolower(c)); });
return normalized;
}

View file

@ -0,0 +1,97 @@
#include "pipeline/asset_manifest.hpp"
#include "core/logger.hpp"
#include <nlohmann/json.hpp>
#include <fstream>
#include <filesystem>
#include <chrono>
namespace wowee {
namespace pipeline {
bool AssetManifest::load(const std::string& manifestPath) {
auto startTime = std::chrono::steady_clock::now();
std::ifstream file(manifestPath);
if (!file.is_open()) {
LOG_ERROR("Failed to open manifest: ", manifestPath);
return false;
}
nlohmann::json doc;
try {
doc = nlohmann::json::parse(file);
} catch (const nlohmann::json::parse_error& e) {
LOG_ERROR("Failed to parse manifest JSON: ", e.what());
return false;
}
// Read header
int version = doc.value("version", 0);
if (version != 1) {
LOG_ERROR("Unsupported manifest version: ", version);
return false;
}
basePath_ = doc.value("basePath", "assets");
manifestDir_ = std::filesystem::path(manifestPath).parent_path().string();
// If basePath is relative, resolve against manifest directory
if (!basePath_.empty() && basePath_[0] != '/') {
basePath_ = manifestDir_ + "/" + basePath_;
}
// Parse entries
auto& entriesObj = doc["entries"];
if (!entriesObj.is_object()) {
LOG_ERROR("Manifest missing 'entries' object");
return false;
}
entries_.reserve(entriesObj.size());
for (auto& [key, val] : entriesObj.items()) {
Entry entry;
entry.filesystemPath = val.value("p", "");
entry.size = val.value("s", uint64_t(0));
// Parse CRC32 from hex string
std::string hexHash = val.value("h", "");
if (!hexHash.empty()) {
entry.crc32 = static_cast<uint32_t>(std::strtoul(hexHash.c_str(), nullptr, 16));
} else {
entry.crc32 = 0;
}
entries_[key] = std::move(entry);
}
loaded_ = true;
auto elapsed = std::chrono::steady_clock::now() - startTime;
auto ms = std::chrono::duration_cast<std::chrono::milliseconds>(elapsed).count();
LOG_INFO("Loaded asset manifest: ", entries_.size(), " entries in ", ms, "ms (base: ", basePath_, ")");
return true;
}
const AssetManifest::Entry* AssetManifest::lookup(const std::string& normalizedWowPath) const {
auto it = entries_.find(normalizedWowPath);
if (it != entries_.end()) {
return &it->second;
}
return nullptr;
}
std::string AssetManifest::resolveFilesystemPath(const std::string& normalizedWowPath) const {
auto it = entries_.find(normalizedWowPath);
if (it == entries_.end()) {
return {};
}
return basePath_ + "/" + it->second.filesystemPath;
}
bool AssetManifest::hasEntry(const std::string& normalizedWowPath) const {
return entries_.find(normalizedWowPath) != entries_.end();
}
} // namespace pipeline
} // namespace wowee

View file

@ -0,0 +1,47 @@
#include "pipeline/loose_file_reader.hpp"
#include "core/logger.hpp"
#include <fstream>
#include <filesystem>
namespace wowee {
namespace pipeline {
std::vector<uint8_t> LooseFileReader::readFile(const std::string& filesystemPath) {
std::ifstream file(filesystemPath, std::ios::binary | std::ios::ate);
if (!file.is_open()) {
return {};
}
auto size = file.tellg();
if (size <= 0) {
return {};
}
std::vector<uint8_t> data(static_cast<size_t>(size));
file.seekg(0, std::ios::beg);
file.read(reinterpret_cast<char*>(data.data()), size);
if (!file.good()) {
LOG_WARNING("Incomplete read of: ", filesystemPath);
data.resize(static_cast<size_t>(file.gcount()));
}
return data;
}
bool LooseFileReader::fileExists(const std::string& filesystemPath) {
std::error_code ec;
return std::filesystem::exists(filesystemPath, ec);
}
uint64_t LooseFileReader::getFileSize(const std::string& filesystemPath) {
std::error_code ec;
auto size = std::filesystem::file_size(filesystemPath, ec);
if (ec) {
return 0;
}
return static_cast<uint64_t>(size);
}
} // namespace pipeline
} // namespace wowee

View file

@ -267,7 +267,7 @@ void Minimap::parseTRS() {
if (trsParsed || !assetManager) return;
trsParsed = true;
auto data = assetManager->getMPQManager().readFile("Textures\\Minimap\\md5translate.trs");
auto data = assetManager->readFile("Textures\\Minimap\\md5translate.trs");
if (data.empty()) {
LOG_WARNING("Failed to load md5translate.trs");
return;