| Name |
Last commit
|
Last update |
|---|---|---|
| .. | ||
| admin | ||
| agent | ||
| ai | ||
| auth | ||
| clinical-gap | ||
| facts | ||
| patient | ||
| persona | ||
| plan | ||
| plan-aggregate | ||
| realtime-coach | ||
| sync |
回答"宿主改了表结构/字段/枚举没通知,PAC 能不能察觉":在已有 mappingMiss(枚举漂移)基础上加
"字段/列漂移"检测,两路:
- assembler 层(column_absent):映射字段对应 host 列在装配行里整列缺席 → 进 stats.suspectFields。
对直连原生表的资源有效;transform 产出表会补 key 掩盖,故补第二路 ↓
- raw 层(form A,column_removed/added,最可靠):ingestRawTables 在 transform 前快照推送列集,
对比"历史同源 rawPayload 列集"(基线抽样 50 条,排除本 run 防自污染,基线<5 跳过)→ 删列/加列。
两路都并入 SyncLog.metadata.suspectFields,/admin/mapping-miss 端点一并透出(报告含 misses+suspectFields)。
本地直连验证:删 doctor_name + 加 new_weird_col 推送 →
suspectFields=[{doctor_name,column_removed},{new_weird_col,column_added}]。
不建表(复用 SyncLog.metadata)。只检测+可见化,不告警/不阻断(按需后续接 webhook)。
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
| Name |
Last commit
|
Last update |
|---|---|---|
| .. | ||
| admin | Loading commit data... | |
| agent | Loading commit data... | |
| ai | Loading commit data... | |
| auth | Loading commit data... | |
| clinical-gap | Loading commit data... | |
| facts | Loading commit data... | |
| patient | Loading commit data... | |
| persona | Loading commit data... | |
| plan | Loading commit data... | |
| plan-aggregate | Loading commit data... | |
| realtime-coach | Loading commit data... | |
| sync | Loading commit data... |