Skip to content

v1.0.0 - Electron-Native Evolution: HUD Revolution, Privacy Protection & High-Fidelity Engine Rebirth

Choose a tag to compare

View all tags
@github-actions github-actions released this 04 Feb 07:30
· 178 commits to master since this release
v1.0.0
fc9a5fe

Important

v1.0.0 是 Akagi-NG 发展历程中的重要里程碑。我们正式告别了纯 Web 模式，全面转向 Electron 原生桌面架构，并引入了极具现代感的毛玻璃 HUD 界面以取代旧有的 PiP 方案。不仅如此，本版本在底层安全性、多平台适配稳健性以及代码测试覆盖率等方面都取得了突破性进展，为您提供更加稳定、安全和视觉效果卓越的 AI 辅助体验。

简体中文 (Simplified Chinese)

✨ 新功能

  • 深度集成的桌面端核心 (Electron-Native Core)：从 Web 迁移至 Electron 架构，后端完全受控于 Electron 生命周期，提供更稳定的运行体验。
  • 全新毛玻璃设计的 HUD 界面：引入了高质感的 HUD 界面，全面取代了原有的 PiP (画中画) 方案，提供更现代化交互体验。
  • 隐私特征混淆与安全性增强：全面优化了 User-Agent 策略并抹除了 Electron 专有指纹特征，显著提升了在各个游戏中的隐蔽性与安全性。

🚀 性能提升

  • 轻量化依赖：移除了不必要的依赖，安装包体积显著减少 100MB+。
  • 极致动态解析：全面采用动态协议解析方案，彻底移除 .pb2 文件污染，提升运行效率并解决环境兼容性问题。
  • 组件延迟加载：对 Dashboard 和 HUD 等关键组件实现延迟加载，优化启动速度。

🎨 UI/UX 增强

  • 毛玻璃 (Glassmorphism) 视觉重构：全组件库采用高阶毛玻璃设计系统，界面更加通透精致。
  • 视觉一致性优化：重构了主题切换逻辑与国际化机制，修复了语言切换时的翻译滞后问题。

⚡ 改进与优化

  • 模型引擎与重连增强：优化了 AI 模型引擎的切换逻辑，并将高性能的 Replay Engine 也应用于断线重连机制，显著增强了系统恢复与初始同步的健壮性。
  • 立直前瞻深度优化：重构了立直前瞻逻辑并与底层 Replay Engine 深度对齐。在保持提供全量候选方案的同时，大幅提升了仿真效率与多平台数据回放的响应速度。

🛠️ 维护与重构

  • 项目目录大重构：为了适配 Electron 架构，对前后端目录结构进行了深度重构。
  • 构建链现代化：废弃旧的 Python 脚本，采用 TS 驱动的构建工具链，提升流水线可维护性。

🐛 问题修复

  • 精准决策回归：修复了多个“吃”操作置信度显示完全一致的严重 Bug，确保决策百分比精确可靠。
  • 多平台适配修复：修复了立直一番街与天月麻雀的平台适配问题，并针对雀魂新协议进行了动态解析优化。
  • 天凤平台适配修正：修复了天凤平台中的一些边缘情况 Bug。

🧪 测试与质量

  • 单元测试覆盖率提升：经过深度重构与补全，后端代码测试覆盖率已提升至 80%+，显著提升了软件的可靠性。
  • Ruff 规范化审计：全代码库通过 Ruff 严格扫描，消除了大量代码异味与低性能写法。
  • 测试架构重构：对单元测试与集成测试进行了全面更新，以匹配新的引擎架构。

English

Important

v1.0.0 marks a significant milestone in the evolution of Akagi-NG. We have officially transitioned from a pure Web model to a full Electron-native desktop architecture and introduced a futuristic Glassmorphism HUD interface to replace the legacy PiP solution. Furthermore, this release brings breakthroughs in underlying security, multi-platform robustness, and code coverage (80%+), providing you with a more stable, secure, and visually stunning AI assistance experience.

✨ Major Features

  • Electron-Native Core Integration: Migrated to Electron architecture for a more stable desktop experience, with the backend managed by the Electron lifecycle.
  • New HUD Interface with Glassmorphism: Introduced a high-fidelity HUD interface, replacing the legacy PiP (Picture-in-Picture) solution with a more modern interaction model.
  • Enhanced Privacy & Fingerprint Obfuscation: Comprehensive sanitization of Electron-specific signatures and User-Agent strategies to reduce detection risks on game platforms.

🚀 Performance Improvements

  • Slimmer Footprint: Removed unnecessary dependencies, significantly reducing build size by 100MB+.
  • Dynamic Protocol Parsing: Fully transitioned to dynamic Protobuf parsing, removing bulky .pb2 files and resolving environment mismatches.
  • Component Lazy Loading: Implemented lazy loading for Dashboard and HUD to boost initial render performance.

🎨 UI/UX Enhancements

  • Glassmorphism Overhaul: Rewrote components with a high-order glassmorphism design system for a premium aesthetic.
  • Translation & Theme Persistence: Fixed UI lag in translation switching and ensured consistent theme rendering across different operating systems.

⚡ Improvements & Refinements

  • Engine & Reconnection Refinement: Optimized AI model engine switching and extended the high-performance ReplayEngine to the reconnection mechanism, significantly enhancing system recovery and initial sync robustness.
  • Advanced Riichi Lookahead: Refactored the lookahead logic to align with the core ReplayEngine. Improves simulation efficiency and replay speed while maintaining full candidate coverage.

🛠️ Maintenance & Refactoring

  • Architectural Refactor: Deeply reorganized directory structures to support the new Electron-centric model.
  • Modern Build Toolchain: Replaced legacy Python build scripts with a modern TypeScript implementation.

🐛 Bug Fixes

  • Decision Accuracy Fix: Resolved the issue where different "Chi" options displayed identical confidence scores.
  • Multi-Platform Support: Resolved compatibility issues for Riichi City and Amatsuki Mahjong, along with optimized Majsoul protocol handling.
  • Tenhou Compatibility: Fixed edge cases in Tenhou platform.

🧪 Testing & Quality

  • Unit Test Coverage Boost: Core backend code coverage has been increased to 80%+ through deep refactoring and completion, ensuring higher reliability.
  • Ruff Linting: Achieved full codebase compliance with Ruff, fixing dead logic and sub-optimal patterns.
  • Test Suite Refactor: Updated unit and integration tests to align with the new modular architecture.

Full Changelog: v0.5.0...v1.0.0

Assets 3
Loading