Skip to content

docs(profile): 主页 v5 结构性熵减——事实对齐、七类重复收敛与折叠层重组 - #7

Merged
ThreeFish-AI merged 7 commits into
masterfrom
ThreeFish-AI/preening-substrate-readme
Sep 30, 2026
Merged

docs(profile): 主页 v5 结构性熵减——事实对齐、七类重复收敛与折叠层重组#7
ThreeFish-AI merged 7 commits into
masterfrom
ThreeFish-AI/preening-substrate-readme

Conversation

@ThreeFish-AI

@ThreeFish-AI ThreeFish-AI commented Sep 30, 2026

Copy link
Copy Markdown
Owner

变更概览

对主页双语 README 执行 preening-substrate 激进档梳理清减（基线 33a64fe，7 个原子提交 B1–B7）：

批次 内容
B1 修 3 处事实漂移：coding-proxy 卡补齐第九家厂商 Xiaomi/小米（对齐折叠区「九家」）；「a quarter fix」→「a fifth」（fix 实为 19.0%）；年度时间线补 2021=198（原图 alt 有值、散文缺失）
B2 技术栈汇总行收敛「一条命令五容器」至 negentropy 权威位置
B3 七类重复叙事按权威位置收敛：归档映射、两个 rc、无 Redis、可插拔后端枚举、上游比例、平方根标度、镜像声明
B4 删 Volume/Cadence/Discipline 汇总块（每项数字仍存于对应图注；16 个 DATA 出现点/文件对称移除）
B5 details 折叠块 10→7：诚实性+省略、纯内核+规约、方法+普查三组合并（内容零删除）
B6 「一个输入框，三篇文献」转入折叠层；8 块重排为 诚实性→规约→项目深潜→FAQ→方法 分层；Notes 引言去重
B7 语义终检：修正 FAQ 对诚实性说明条目序号的失准表述；术语与数字全页对账

量化对照（基线 → 终态）

指标 基线 终态
行数（EN=ZH 逐行镜像） 335 313（−6.6%）
字节 EN / ZH 43,767 / 42,660 42,113 / 41,016
DATA 标记 / 文件 91 75（16 个纯复述出现点移除；pct_hour 归零）
FIG 图表 / 文件 13 13（区间体逐字节恒等）
details 折叠块 10 8（信息零丢失）
事实不一致 4 处 0 处

护栏与核验

  • 管线契约全程绿：--lint 双语标记序列逐位一致（75 DATA + 13 FIG … identical set and order）；FIG 区间体、DATA 值、mermaid、脚注、外链与基线逐字节恒等；scripts/、.github/、assets/ 零变更
  • 逐行镜像传统保持（EN 第 n 行 ↔ ZH 第 n 行，313=313）
  • 独立 Verifier 四门盲审（护栏防篡改 / 契约零断裂 / 渲染卫生 / 度量对账）18 项全 PASS
  • 分支渲染与 master 同抓取器对照等价

Follow-up 候选（未纳入本次）

  • 脚本侧 3 个未被引用的 FACT_KEYS（peak_n/peak_pct/rel_repos）可在脚本域单独清减
  • ZH 版 13 图使用 raw.githubusercontent master 绝对 URL，默认分支更名时会整批断链

🤖 Generated with Claude Code

ThreeFish-AI and others added 7 commits September 30, 2026 22:15
@ThreeFish-AI @claude
docs(profile): 修复厂商名单、类型占比与年度时间线三处事实不一致;
b40cabf 
- coding-proxy 卡补齐第九家厂商 Xiaomi/小米，与折叠区「九家」口径对齐
- Conventional Commits 类型占比措辞 quarter→fifth（fix 实为 19.0%）
- 年度时间线补齐 2021 年 198 条提交（原图 alt 有值，散文缺失）

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@ThreeFish-AI @claude
docs(profile): 技术栈汇总行收敛「一条命令五容器」至 negentropy 权威位置;
b235418 
🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@ThreeFish-AI @claude
docs(profile): 七类重复叙事按权威位置收敛——归档映射、rc、无 Redis、后端枚举、上游比例、平方根、镜像声明;
d7e2aa1 
🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@ThreeFish-AI @claude
docs(profile): 删除 Volume/Cadence/Discipline 汇总块——纯聚合复述，各数字仍存于对应图注;
c0bd554 
🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@ThreeFish-AI @claude
docs(profile): details 折叠区 10 块合并为 7 块——诚实性+省略、纯内核+规约、方法+普查，内容零删除;
b778b43 
🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@ThreeFish-AI @claude
docs(profile): R3 章节重排——「一个输入框」移入折叠层，八块按 诚实性→规约→深潜→FAQ→方法 序分层，Notes 引…
b91c025 
…言去重;

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@ThreeFish-AI @claude
docs(profile): 语义终检——修正 FAQ 对诚实性说明条目序号的失准表述，术语与数字全页对账通过;
9623b81 
🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
@ThreeFish-AI
ThreeFish-AI merged commit 993de20 into master Sep 30, 2026
2 checks passed
@ThreeFish-AI ThreeFish-AI mentioned this pull request Oct 18, 2026
Merged
ThreeFish-AI added a commit that referenced this pull request Oct 18, 2026
@ThreeFish-AI
docs(profile): 主页手写重复熵减 + 护栏上收——双清单收敛、结构测试与 CI 强制门 (#9)
896c841 
* docs(profile): 主页手写重复收敛——九厂商清单与 no-Redis 短语各归唯一权威位;

- coding-proxy 九厂商清单曾两处全文手写且已实际漂移过一次（PR #7 仅单侧补 Xiaomi）：项目卡定为唯一权威清单，深潜块收敛为回指，消除双清单结构；
- 故障转移 doctrine 句去掉第三处 no Redis/no queue 点名（权威位保留项目卡与总栈行），因果链改由 SQLite-WAL 本地性直接引出；
- 中英两份 README 逐行镜像同步；13 FIG + 75 DATA 标记流、锚点三方契约、全部冻结面零触碰，31 用例与 --lint 全绿。

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>

* docs(agents): 新建知识索引与 Issue 台账，沉淀双语镜像编辑契约;

- knowledge-map.md：文档索引 + 双语镜像编辑契约卡（机器领地/标记即契约/锚点三方同名同序/冻结面绕行/编辑后验证/重复收敛原则）；
- issue.md：登记 I-001 九厂商清单单侧漂移（本次结构性收敛）与 I-002 语义级镜像漂移、锚点死链无机器门（lab gate 护航，上收 tests 为后续路径）。

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>

* test(readme): 锚点三方一致与双语行级镜像上收为结构测试;

- 新增 tests/test_readme_structure.py：锚点序 == FIG 簇序、导航链接集 == 锚点集（两份 README 各查）、marker 与结构骨架行号镜像、行数相等、cur_year/cur_total 单调性状态源存续；
- 此前三项承重面仅靠人工契约（破坏即 GitHub 静默死链），现并入零依赖特征化套件（31 → 35 用例）。

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>

* ci(lint): PR lint job 接入全套 unittest 护栏;

- refresh-profile-data.yml 的 lint job 新增静态步骤 python -m unittest discover tests；
- 35 用例（Golden Master、结构守卫、SVG 闸门金样）由本地自觉变 PR 强制门，零网络、亚秒级。

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>

* refactor(pipeline): cadence_alt 移出 refresh 至模块级，与其余 alt 对齐;

- 原嵌套于 refresh() 内（唯一非模块级的 *_alt），闭包变量 rel_total 改为显式参数穿入；
- 函数体除缩进与签名外逐 token 不变（独立 Verifier 经 AST 比对确认），EN/ZH 合成输入行为冒烟一致；
- 附带收益：tests/_bps.py 按名加载器现可直接加载 cadence_alt，为后续可测铺路。

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>

* docs(agents): 契约卡与 Issue 台账现行态同步;

- knowledge-map：锚点三方契约标注机器强制、编辑后验证命令更新为 35 用例双命令（与 CI 同款）；
- issue：I-002 根治闭环（结构测试上收 + CI 接入），状态置已完成。

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>

* test(readme): 删除臆造分区号并强化导航链接顺序断言;

- docstring 去掉「Section B2」——全仓仅存在 Section A/C/C/D/A/C/E 的非正式引用，B/B2 无出处；
- 簇导航链接断言由 sorted 集合比较强化为「文档序 == FIG 簇序」，与 knowledge-map 契约卡第 4 条「集合与顺序一致」宣称同口径；变异自证换序可判红。

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>

* ci(lint): paths 增补 tests/**，消除纯测试 PR 的门禁盲区;

- 此前仅改 tests/ 的 PR 会整体跳过 lint job，新接入的 unittest 步骤对其不生效，与 issue.md I-002「全套护栏变 PR 强制门」宣称相悖；
- 坏测试可无 CI 合入的缺口闭合。

🤖 Generated with [Claude Code](https://github.com/claude), [CodeX](https://openai.com), [Gemini](https://github.com/apps/gemini-code-assist)
Co-Authored-By: Aurelius Huang<threefish.ai@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@ThreeFish-AI