fix: models with multiple parts not being normalized properly (#1825)

Tested that models are migrated to the new format a-ok
Tested that the new logic works as expected.
Old code needs to be kept so that models in both states can be brought to modern standards
This commit is contained in:
David Markowitz
2025-06-23 00:08:16 -07:00
committed by GitHub
parent 48e3471831
commit 9f92f48a0f
9 changed files with 251 additions and 4 deletions

210
dCommon/LxfmlBugged.cpp Normal file
View File

@@ -0,0 +1,210 @@
#include "Lxfml.h"
#include "GeneralUtils.h"
#include "StringifiedEnum.h"
#include "TinyXmlUtils.h"
#include <ranges>
// this file should not be touched
Lxfml::Result Lxfml::NormalizePositionOnlyFirstPart(const std::string_view data) {
Result toReturn;
tinyxml2::XMLDocument doc;
const auto err = doc.Parse(data.data());
if (err != tinyxml2::XML_SUCCESS) {
LOG("Failed to parse xml %s.", StringifiedEnum::ToString(err).data());
return toReturn;
}
TinyXmlUtils::DocumentReader reader(doc);
std::map<std::string/* refID */, std::string> transformations;
auto lxfml = reader["LXFML"];
if (!lxfml) {
LOG("Failed to find LXFML element.");
return toReturn;
}
// First get all the positions of bricks
for (const auto& brick : lxfml["Bricks"]) {
const auto* part = brick.FirstChildElement("Part");
if (part) {
const auto* bone = part->FirstChildElement("Bone");
if (bone) {
auto* transformation = bone->Attribute("transformation");
if (transformation) {
auto* refID = bone->Attribute("refID");
if (refID) transformations[refID] = transformation;
}
}
}
}
// These points are well out of bounds for an actual player
NiPoint3 lowest{ 10'000.0f, 10'000.0f, 10'000.0f };
NiPoint3 highest{ -10'000.0f, -10'000.0f, -10'000.0f };
// Calculate the lowest and highest points on the entire model
for (const auto& transformation : transformations | std::views::values) {
auto split = GeneralUtils::SplitString(transformation, ',');
if (split.size() < 12) {
LOG("Not enough in the split?");
continue;
}
auto x = GeneralUtils::TryParse<float>(split[9]).value();
auto y = GeneralUtils::TryParse<float>(split[10]).value();
auto z = GeneralUtils::TryParse<float>(split[11]).value();
if (x < lowest.x) lowest.x = x;
if (y < lowest.y) lowest.y = y;
if (z < lowest.z) lowest.z = z;
if (highest.x < x) highest.x = x;
if (highest.y < y) highest.y = y;
if (highest.z < z) highest.z = z;
}
auto delta = (highest - lowest) / 2.0f;
auto newRootPos = lowest + delta;
// Clamp the Y to the lowest point on the model
newRootPos.y = lowest.y;
// Adjust all positions to account for the new origin
for (auto& transformation : transformations | std::views::values) {
auto split = GeneralUtils::SplitString(transformation, ',');
if (split.size() < 12) {
LOG("Not enough in the split?");
continue;
}
auto x = GeneralUtils::TryParse<float>(split[9]).value() - newRootPos.x;
auto y = GeneralUtils::TryParse<float>(split[10]).value() - newRootPos.y;
auto z = GeneralUtils::TryParse<float>(split[11]).value() - newRootPos.z;
std::stringstream stream;
for (int i = 0; i < 9; i++) {
stream << split[i];
stream << ',';
}
stream << x << ',' << y << ',' << z;
transformation = stream.str();
}
// Finally write the new transformation back into the lxfml
for (auto& brick : lxfml["Bricks"]) {
auto* part = brick.FirstChildElement("Part");
if (part) {
auto* bone = part->FirstChildElement("Bone");
if (bone) {
auto* transformation = bone->Attribute("transformation");
if (transformation) {
auto* refID = bone->Attribute("refID");
if (refID) {
bone->SetAttribute("transformation", transformations[refID].c_str());
}
}
}
}
}
tinyxml2::XMLPrinter printer;
doc.Print(&printer);
toReturn.lxfml = printer.CStr();
toReturn.center = newRootPos;
return toReturn;
}
Lxfml::Result Lxfml::NormalizePositionAfterFirstPart(const std::string_view data, const NiPoint3& position) {
Result toReturn;
tinyxml2::XMLDocument doc;
const auto err = doc.Parse(data.data());
if (err != tinyxml2::XML_SUCCESS) {
LOG("Failed to parse xml %s.", StringifiedEnum::ToString(err).data());
return toReturn;
}
TinyXmlUtils::DocumentReader reader(doc);
std::map<std::string/* refID */, std::string> transformations;
auto lxfml = reader["LXFML"];
if (!lxfml) {
LOG("Failed to find LXFML element.");
return toReturn;
}
// First get all the positions of bricks
for (const auto& brick : lxfml["Bricks"]) {
const auto* part = brick.FirstChildElement("Part");
bool firstPart = true;
while (part) {
if (firstPart) {
firstPart = false;
} else {
LOG("Found extra bricks");
const auto* bone = part->FirstChildElement("Bone");
if (bone) {
auto* transformation = bone->Attribute("transformation");
if (transformation) {
auto* refID = bone->Attribute("refID");
if (refID) transformations[refID] = transformation;
}
}
}
part = part->NextSiblingElement("Part");
}
}
auto newRootPos = position;
// Adjust all positions to account for the new origin
for (auto& transformation : transformations | std::views::values) {
auto split = GeneralUtils::SplitString(transformation, ',');
if (split.size() < 12) {
LOG("Not enough in the split?");
continue;
}
auto x = GeneralUtils::TryParse<float>(split[9]).value() - newRootPos.x;
auto y = GeneralUtils::TryParse<float>(split[10]).value() - newRootPos.y;
auto z = GeneralUtils::TryParse<float>(split[11]).value() - newRootPos.z;
std::stringstream stream;
for (int i = 0; i < 9; i++) {
stream << split[i];
stream << ',';
}
stream << x << ',' << y << ',' << z;
transformation = stream.str();
}
// Finally write the new transformation back into the lxfml
for (auto& brick : lxfml["Bricks"]) {
auto* part = brick.FirstChildElement("Part");
bool firstPart = true;
while (part) {
if (firstPart) {
firstPart = false;
} else {
auto* bone = part->FirstChildElement("Bone");
if (bone) {
auto* transformation = bone->Attribute("transformation");
if (transformation) {
auto* refID = bone->Attribute("refID");
if (refID) {
bone->SetAttribute("transformation", transformations[refID].c_str());
}
}
}
}
part = part->NextSiblingElement("Part");
}
}
tinyxml2::XMLPrinter printer;
doc.Print(&printer);
toReturn.lxfml = printer.CStr();
toReturn.center = newRootPos;
return toReturn;
}