feat(tools): 代码定位保守修复脚本与校验产物 (#4)#61
Open
cgflag wants to merge 3 commits into
Open
Conversation
Deliver location repair tooling, fixed artifacts, verifier, and high-confidence subset without modifying data/entries.jsonl. Co-authored-by: Cursor <cursoragent@cursor.com>
Produce entries.fixed.desc_synced.jsonl (10 anchors on entry-00458), export needs_human.for_49.csv with fetch_failed dropped, and document COMPAT with PR Tencent#49 init smoke — without modifying data/entries.jsonl. Co-authored-by: Cursor <cursoragent@cursor.com>
Enable tested --repo-wide path and document zero unique incremental hits; keep default off and leave official entries untouched.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
关联 Issue:#4
问题是什么?
data/entries.jsonl中部分节点的line/file与对应 commit 上的code对不齐。不确定时不能猜,否则会引入新的错误标注。做了什么?
tools/fix_locations/:按官方顺序原位置 → 邻域 ±5 → 原文件全文 → 全仓库唯一命中(
--repo-wide)。默认关闭全仓搜索;开启后仅当全仓库唯一命中才改
file/line。file/line;多命中 / 找不到 / 退化片段(如单独})写入needs_human.csv。entries.fixed.jsonl、fix_diff.csv、needs_human.csv、README、单元测试与校验脚本。data/entries.jsonl(官方要求提交修复后的 JSONL,未要求直接改主数据;是否合入由维护者决定)。另附高置信小子集,便于择优合入。和已有 PR 的关系
code、原先只标起始行」扩成N-M也算自动修复。needs_human.for_49.csv+COMPAT_WITH_49.md,并对 #49init做过冒烟(见COMPAT_49_SMOKE.md)。数字解读: 自动修复 102 处 = 90 处「扩成行范围」+ 12 处「行号真的偏了」。
与 #35/#39 修复集合相交约 12 处;与 #41 相交约 94 处。详见
tools/fix_locations/out/OVERLAP_ANALYSIS.md。本 PR 额外提供的实用件
out/OVERLAP_ANALYSIS.md)scripts/verify_location_outputs.py):schema、与 diff 一致性、抽查幂等out/entries.high_confidence.jsonl(5 条 entry)out/FETCH_FAILED_NOTE.mdentries.fixed.desc_synced.jsonl(本批 10 处,全在entry-00458;不覆盖entries.fixed.jsonl)export_queue_for_49.py默认剔除fetch_failed;与 feat(tools): 添加可审计的代码定位人工复核闭环 #49 apply 内 desc 同步可并存(产物分文件)--repo-wide):已全量跑通;增量唯一命中 0;报告见out/repo_wide/(DELTA.md/needs_human.csv等)结果摘要
needs_human.csv(导出needs_human.for_49.csv供人审)--repo-wide全量:与上列 102 同集合;repo_wide_unique增量 0;needs_human 细化后为 18×degenerate_snippet+ 55×repo_wide_no_match(退化未被「修没」)out/SPOTCHECK_AUTO.md复现
热缓存下全量(默认不开全仓):
可选:开启第 3 级全仓唯一搜索(更慢;产物建议隔离目录):