Chengfang Lu
e72bc061c5
Major changes: - New frontend (platform/web/): Vite + React 18 + TypeScript + Tailwind - 4-module navigation: 数据送标 / 模型管理 / 车队管理 / 系统管理 - Data catalog with charts (DMS/ADAS/Lane 3-tab view) - Quality review workflow (标注质检): Good/Fine/Bad scoring with auto-advance - Audit enhancements: batch operations, rejection categories, Feishu notifications - Operation audit log (操作日志) - World model simulation studio (仿真工坊) - Dataset version management with snapshots and diff - ADAS 7-class dataset integration (138K images organized + compressed) - User management with Feishu integration and pagination - CRUD/search/filter on all pages, card layout redesign - PIL-optimized image overlay rendering - Auto-snapshot on build, in_review workflow stage - Removed embedded algorithm code (now in workspace)
1.2 KiB
1.2 KiB
DMS 标注分期验收（addw pilot）
P0：dam / ddaw（检测）
- 批次台账入湖 → inbox 有
images/train - 送标工作台「待送标」可见批次
- 进入标注：左侧工具箱（face 等）+ 保存
- 提交后
batch.metastage =labeling_submitted，「待入库」Tab 可见 - 导出 Job 成功后 stage =
returned - 提交入库审核 → build 通过
P1：addw / addw_face（当前 pilot）
Pilot：inbox/addw/20260525_pilot（12 张）
dms_detect.xml四类标签dms_pose.xml：face 框 + 37 关键点（kp_00–kp_36）export_ls_to_yolo.py：LS JSON → YOLO detect / YOLO pose- 标注保存至
labels/ls_annotations/*.json - 提交 / 导出 / 待入库全链路（见 P0 检查项）
- addw_face 导出 txt 每行 116 字段，通过
validate_pose_label
python3 datasets/dms/scripts/test_export_ls_to_yolo.py
P2：forward / lane
forward__detect/forward__classifyregistry 模板- lane
lane_culane.xml画布 smoke
自动化
bash scripts/smoke_labeling_api.sh
bash scripts/smoke_pending_gate.sh
python3 scripts/lake_checklist_audit.py