AdministratorandClaude Opus 4.8 76320ca5a3 fix(knowledge): 模型判定 should_ingest=false 时直接跳过,不再落草稿
日志显示寒暄/无效会话并非 LLM 调用失败,而是模型正确返回 should_ingest=false。
上一版把这种情况也当失败回退、堆成草稿,反而把无价值对话存进了库。

- llm_extractor: 模型显式 should_ingest=false 且无 notes → 抛 LLMRejectedIngest
- capture_thread: 捕获后抛 EmptyThreadError → 整条跳过、不入库(寒暄/总结此页这类不再进库)
- 仅当真正失败(API 报错 / 非 JSON)才回退规则抽取并落草稿
- 兼容包装 extract_thread_llm 吞掉该异常返回 None
- 更新测试(should_ingest=false 跳过)+ CLAUDE.md

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-08 22:49:00 +08:00
2026-06-03 10:01:10 +08:00
2026-06-01 14:16:13 +08:00
2026-05-31 16:48:43 +08:00
2026-05-22 20:40:36 +08:00
2026-05-28 17:14:01 +08:00
2026-06-05 20:06:55 +08:00
2026-05-27 19:43:15 +08:00
2026-06-03 09:54:11 +08:00
2026-05-13 22:22:59 +08:00
2026-05-28 17:14:01 +08:00
2026-05-28 17:14:01 +08:00
2026-05-22 20:40:36 +08:00
S
Description
deerflow-code
238 MiB
0 Stars 1 Watchers 0 Forks
Languages
Python 39.2%
TypeScript 35.9%
HTML 22.2%
CSS 1.5%
JavaScript 0.9%
Other 0.2%