style-fingerprint · diff
git:20260317.252a620 to git:20260830.05341dd
60 added, 105 removed. Audit A to A.
---
name: style-fingerprint
description: |
- Analyze and save Chinese writing style fingerprints.
- Extracts syntax patterns, lexical preferences, rhetorical features from text.
- Saves fingerprints to ./fingerprints/ for writing agents to use.
- No external dependencies - pure Python implementation.
+ Analyze and save Chinese writing style fingerprints (v2, zero dependencies).
+ Extracts rhythm, function-word habits, rhetoric patterns and exemplar sentences.
+ NEW in v2: compare (draft vs fingerprint deviation report), merge (multi-text),
+ selftest (regression), exemplar few-shot export, low-confidence marking.
- 分析和保存中文写作风格指纹。从文本中提取句法模式、词汇偏好、修辞特征。
- 将指纹保存到 ./fingerprints/ 目录供写作助手使用。
- 无外部依赖 - 纯 Python 实现。
+ 分析并保存中文写作风格指纹（v2，零依赖）。
+ 提取节奏、功能词习惯、句式修辞与范例句。
+ v2 新增：compare 新稿校验、merge 多文本合并、selftest 回归测试、
+ 范例句导出（few-shot）、小样本置信度标注。
metadata:
openclaw:
requires:
bins: ["python3"]
- emoji: 🎨
+ emoji: \U0001F3A8
category: writing
commands:
- name: analyze
description: Analyze text/file and save fingerprint | 分析文本/文件并保存指纹
args:
- - name: text
- flag: --text
- required: false
- description: Text content to analyze | 要分析的文本内容
- - name: file
- flag: --file
- required: false
- description: File path to analyze | 要分析的文件路径
- - name: name
- flag: --name
- required: true
- description: Fingerprint name (for saving) | 指纹名称（用于保存）
-
+ - {name: text, flag: "--text", required: false, description: "Text to analyze | 要分析的文本"}
+ - {name: file, flag: "--file", required: false, description: "File to analyze | 要分析的文件"}
+ - {name: name, flag: "--name", required: true, description: "Fingerprint name | 指纹名称"}
- name: list
- description: List all saved fingerprints | 列出所有保存的指纹
-
+ description: List all fingerprints | 列出所有指纹
- name: show
description: Show fingerprint details | 显示指纹详情
args:
- - name: name
- flag: --name
- required: true
-
+ - {name: name, flag: "--name", required: true}
- name: delete
description: Delete a fingerprint | 删除指纹
args:
- - name: name
- flag: --name
- required: true
-
+ - {name: name, flag: "--name", required: true}
- name: export
- description: Export style guide for writing agent | 导出写作助手指南
+ description: Export style guide (rules + exemplar sentences) | 导出风格指南（规则+范例句）
args:
- - name: name
- flag: --name
- required: true
- - name: output
- flag: --output
- required: false
- description: Output file path | 输出文件路径
+ - {name: name, flag: "--name", required: true}
+ - {name: output, flag: "--output", required: false, description: "Output file | 输出文件"}
+ - name: compare
+ description: Check a draft against a fingerprint (deviation report) | 新稿 vs 指纹偏差校验
+ args:
+ - {name: name, flag: "--name", required: true, description: "Fingerprint to compare against | 指纹名"}
+ - {name: text, flag: "--text", required: false}
+ - {name: file, flag: "--file", required: false}
+ - name: merge
+ description: Merge multiple fingerprints (weighted by char count) | 合并多个指纹（按字数加权）
+ args:
+ - {name: names, flag: "--names", required: true, description: "Comma-separated names | 逗号分隔"}
+ - {name: name, flag: "--name", required: true, description: "Merged fingerprint name | 合并后名称"}
+ - name: selftest
+ description: Run built-in regression tests | 运行内置回归测试
---
- # Style Fingerprint | 写作风格指纹
-
- **English**: Analyze Chinese writing style and save fingerprints for AI writing agents.
+ # Style Fingerprint v2 | 写作风格指纹
- **中文**: 分析中文写作风格并保存指纹，供 AI 写作助手使用。
+ ## What changed in v2 | v2 变更
- ---
+ | Fix/New | 说明 |
+ |---|---|
+ | Fixed | 反问句检测（v1 正则语序错误，永远返回 0） |
+ | Fixed | 被动语态误伤「由于/自由」 |
+ | Fixed | 省略主语字符类拆词误判 |
+ | Fixed | 隐喻误报（排除副词「好像/似乎」） |
+ | Rewritten | 词汇层：滑窗伪分词 → 封闭集合功能词指纹（疑问/连接/情态/程度/人称/语气词）+ 跨句口头禅候选 |
+ | New | `compare`：新稿 vs 指纹，15 维度偏差报告 + 修正建议（写作质量闸门） |
+ | New | `merge`：多文本合并指纹（按字数加权） |
+ | New | `export` 范例句层：few-shot 例句比规则对 LLM 更有效 |
+ | New | `selftest`：11 项内置回归用例 |
+ | New | <300 字自动标注低置信度 |
+ | Format | 指纹 JSON v2：全部指标改为比率（每百句/每千字），跨文本可比 |
- ## Quick Start | 快速开始
+ ## Recommended workflow | 推荐工作流
```bash
- # Analyze text | 分析文本
- python3 style_fingerprint.py analyze --text "你的文本内容" --name "我的风格"
-
- # Analyze file | 分析文件
- python3 style_fingerprint.py analyze --file ./article.txt --name "作者A"
-
- # List saved fingerprints | 列出保存的指纹
- python3 style_fingerprint.py list
+ # 1. 提取作者指纹（建议 ≥300 字，多篇可用 merge 合并）
+ python3 style_fingerprint.py analyze --file author_a_1.txt --name 作者A
+ python3 style_fingerprint.py analyze --file author_a_2.txt --name 作者A2
+ python3 style_fingerprint.py merge --names "作者A,作者A2" --name 作者A综合
- # Show fingerprint details | 显示指纹详情
- python3 style_fingerprint.py show --name "我的风格"
+ # 2. 导出风格指南（规则 + 范例句）给写作 Agent
+ python3 style_fingerprint.py export --name 作者A综合 --output style_guide.md
- # Export style guide for writing agent | 导出写作助手指南
- python3 style_fingerprint.py export --name "我的风格" --output ./guide.txt
+ # 3. Agent 写作后，校验新稿是否还原了风格
+ python3 style_fingerprint.py compare --name 作者A综合 --file draft.txt
```
- ---
-
- ## What It Analyzes | 分析维度
-
- **English**:
- - **Basic Stats**: Sentence length, rhythm, comma density
- - **Syntax Patterns**: Rhetorical questions, passive voice, ellipsis
- - **Lexical Fingerprint**: Top words, word diversity
- - **Logic Flow**: Transition word preferences
- - **Rhetoric**: Metaphors, sensory description types
-
- **中文**:
- - **基础统计**: 句子长度、节奏感、逗号密度
- - **句法模式**: 修辞问句、被动语态、省略号使用
- - **词汇指纹**: 高频词汇、词汇多样性
- - **逻辑流**: 过渡词偏好
- - **修辞手法**: 隐喻、感官描述类型
-
- ---
-
- ## Storage | 存储
-
- **English**: Fingerprints are saved in `./fingerprints/` as JSON files.
-
- **中文**: 指纹以 JSON 格式保存在 `./fingerprints/` 目录中。
-
- ---
-
- ## For Writing Agents | 供写作助手使用
-
- **English**: Use `export` command to get a compact style guide that can be included in prompts.
-
- **中文**: 使用 `export` 命令获取紧凑的风格指南，可嵌入到提示词中使用。
-
- ---
-
- ## Features | 特性
-
- **English**:
- - ✅ **No external dependencies** - Pure Python implementation
- - ✅ **Chinese optimized** - Designed for Chinese text analysis
- - ✅ **Persistent storage** - Saves fingerprints to `./fingerprints/`
- - ✅ **Writing agent ready** - Export compact style guides
+ ## Notes | 注意
- **中文**:
- - ✅ **无外部依赖** - 纯 Python 实现
- - ✅ **中文优化** - 专为中文文本分析设计
- - ✅ **持久化存储** - 将指纹保存到 `./fingerprints/`
- - ✅ **写作助手就绪** - 导出紧凑的风格指南
+ - 指纹保存在技能目录 `fingerprints/` 下（自包含设计，拷贝技能即带走数据）
+ - <300 字的样本会标注 low_confidence，结论仅供参考
+ - compare 一致率是参考值：句式类维度 ±30% 内为 ✅
+ - 零外部依赖，Python 3.7+ 即可运行