PP 预测与更快的刷新🎉
PP prediction & a faster refresh🎉
🔍更新内容 What's in v2.1.0
-
PP 预测(准确率预览) 点击任意 ranked 回放的 PP 值,在下方弹出准确率预览小窗:查看该谱面在不同准确率下的预估 PP,公式复刻自 ScoreSaber 官方(误差 ±0.1%)。
PP prediction (accuracy preview) Click a ranked replay's PP value to open an accuracy-preview popover: read the estimated PP at any accuracy, using ScoreSaber's official formula (±0.1% deviation). -
逐场 PP 计算 总览 / 历史 / 详情里的 PP 现在按每条本地回放的实际准确率独立计算;NF 场次按有效分折算。
Per-play PP Overview / history / detail PP is now computed per local replay from its actual accuracy; NF plays use the effective score. -
更快的一键刷新 批量分析不再逐条调用 AI 生成报告;AI 报告改为在详情页手动生成,一键刷新速度大幅提升。
Faster one-click refresh Batch analysis no longer calls the AI per replay; reports are now generated on demand from the detail page , making the refresh much faster. -
启动自动替换旧实例 重新启动 SaberLab 时会自动顶掉 6980 上残留的旧实例并重新绑定 6980,窗口不再「消失」;其他程序占用端口时依旧安全顺延,不会误杀。
Launcher replaces stale instances Relaunching SaberLab replaces the previous instance on 6980 (verified identity), so its window never gets lost; unrelated port occupants are never touched and still trigger the safe fallback. -
LocalLeaderboard 副本恢复 回放文件被外部删除时会明确标注「文件缺失」;若安装了 LocalLeaderboard mod,一键刷新会自动用其副本找回并修复这类回放。
LocalLeaderboard safety copy If a replay file was deleted externally, a clear "File missing" mark appears; with LocalLeaderboard installed, one-click refresh auto-recovers such replays from its copy. -
保存设置不再清空分析数据 修复此前保存任意设置都会清空分析缓存的 bug。
Settings save no longer wipes analysis Fixed a bug where saving any setting cleared the analysis cache. -
更多细节打磨 toast 改为顶部居中(带进度条)、多语言补全。
More polish Top-center toasts with progress bars, and i18n completions.
🚀 快速开始 Quick Start
-
解压
SaberLab-v2.1.0-win64.zip到任意目录,双击SaberLab.exe启动。
Unzip and runSaberLab.exe. -
打开「设置」→「游戏路径」,选择 Beat Saber 游戏根目录。
In Settings → Game Path, select your Beat Saber root folder. -
点击「⚡ 一键刷新」导入回放;在「云端数据」页选择数据源并拉取数据计算动态水平。
Click ⚡ Quick Refresh to import replays; on the Cloud Data page pick your source and fetch your dynamic level. -
(可选)配置 AI 报告:在程序目录的
.env中填入DEEPSEEK_API_KEY(参考自带.env.example)。未配置时自动使用内置规则报告。
Optional: addDEEPSEEK_API_KEYto the.envnext toSaberLab.exe(see.env.example); the built-in rule-based report is used otherwise.