Skip to content

fix(record): retry unparsable L1 extraction output#190

Open
RerankerGuo wants to merge 1 commit into
TencentCloud:mainfrom
RerankerGuo:fix/l1-extraction-parse-retry
Open

fix(record): retry unparsable L1 extraction output#190
RerankerGuo wants to merge 1 commit into
TencentCloud:mainfrom
RerankerGuo:fix/l1-extraction-parse-retry

Conversation

@RerankerGuo

Copy link
Copy Markdown

Description | 描述

Retry L1 extraction once when the model returns output that cannot be parsed as the required JSON array.

  • Keeps valid empty extraction results ([]) as-is and does not retry them.
  • Reuses the same extraction prompt with an additional retry instruction that asks for JSON only.
  • Adds focused coverage for parse-failure retry success and for not retrying a valid empty array.

Related Issue | 关联 Issue

Fix #115

Change Type | 修改类型

  • Bug fix | Bug 修复
  • New feature | 新功能
  • Documentation update | 文档更新
  • Code optimization | 代码优化

Self-test Checklist | 自测清单

  • Verified locally | 本地验证通过
  • No existing features affected | 无影响现有功能

Additional Notes | 其他说明

Verified with npx vitest run src/core/record/l1-extractor.test.ts, npm test, and npm run build using Node v24.15.0.

Signed-off-by: Ziyang Guo <121015044+RerankerGuo@users.noreply.github.com>
@Maxwell-Code07

Copy link
Copy Markdown
Collaborator

Retrying on parse failure is a pragmatic improvement for L1 extraction reliability. We appreciate the care taken to not retry valid empty results. We'll review it internally.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

bug: seed command fails silently — L1 extraction JSON parse errors, L2/L3 timer conflicts, no retry

2 participants