Files
HSAP/docs/DATA_LAKE_GAP.md
Chengfang Lu e72bc061c5 feat: HSAP platform v2 — modular navigation, quality review, audit log, world model simulation
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)
2026-06-03 11:40:21 +08:00

2.2 KiB
Raw Blame History

数据入湖清单 vs HSAP 实现差距

对照 DATA_LAKE_CHECKLIST.md 阶段 AE 与当前 as_platform 能力。

阶段 清单要求 HSAP 现状 差距
A 上传接入 zip/目录上传、进度、candidate_id POST /api/v1/data/upload/fileDatasetCandidate 表;analyzed 后可 POST .../promote-inbox 无统一 lake/staging/ 路径约定;进度条依赖前端 upload
A staging 区隔离 候选写入 DB + 磁盘路径 未强制 lake/staging/<project>/<candidate_id>/ 目录规范
B 自动分析 上传后异步 quality worker inspect-upload、部分 catalog 刷新 无独立 QualityWorker JobDMS/Lane 报告未统一落 quality.json
B DMS/Lane 指标 Catalog、catalogDms、validate 脚本 Catalog 已展示采样指标(条/饼/竖柱/雷达/划分柱/散点/密度);上传触发全自动
C 审核流 自动提交审核单 approvalssubmit API 已有;与送标 register 联动
C 通过/驳回规范 approve/reject 已有
D 版本入湖 审核后晋级 curated ingest_incrementalregister_batch stage 主路径在 ml.py/as.py,非 candidate→lake 闸门
D catalog 索引更新 GET /catalog refresh 已有
E 运维安全 失败可读、重试 Job 队列、approval 备注 部分;上传重试靠前端

已有可复用组件

  • 数据候选:platform/as_platform/db/models.pyDatasetCandidate
  • 上传 APIserver.pyupload/fileinspect-upload
  • 审核:audit/queue.py/api/v1/approvals/*
  • 入湖 CLIas.py build / add + ingest_incremental.py

建议下一里程碑(未在本汇总 plan 全量实现)

  1. 统一 staging 根目录与环境变量 AS_LAKE_STAGING_ROOT
  2. 上传完成 → 入队 quality_analyze Job → 写 quality.json
  3. 审核通过后调用现有 ingest_incremental 并更新 batch.meta stage

验收脚本

bash HSAP/scripts/smoke_manifest_alignment.sh
bash HSAP/scripts/smoke_platform_api.sh
curl -sS -H "Authorization: Bearer $TOKEN" http://127.0.0.1:8787/api/v1/pending/gates