wren
547 lines
16 KiB
YAML
547 lines
16 KiB
YAML
metadata:
|
||
type_id: govdoc_general
|
||
name: 内部公文通用规则
|
||
version: "0.1.0"
|
||
source: 公文文稿常见错误汇编（第一期）·2025-11
|
||
description: 基于旧内部公文正式规则语义整理的当前平台规则集。
|
||
|
||
extract:
|
||
# 8 个内置实体（title / doc_number / recipient / date /
|
||
# signature / attachments / wenzhong / issuer）由代码自动产出。
|
||
entities: []
|
||
|
||
rules:
|
||
- group: 标题（错误汇编 一）
|
||
rules:
|
||
- rule_id: GW-T-001
|
||
name: 标题文种合规性
|
||
severity: error
|
||
category: 标题
|
||
target: title
|
||
on_missing: fail
|
||
stages:
|
||
- check: ai
|
||
prompt: |
|
||
审查公文标题是否符合规范。
|
||
标题：{{title.text}}
|
||
|
||
15 种合法文种：决议、决定、命令（令）、公报、公告、通告、意见、
|
||
通知、通报、报告、请示、批复、议案、函、纪要
|
||
|
||
检查要点：
|
||
1. 是否使用了合法文种
|
||
2. 方案/规划/办法/细则等是否以"通知"形式下发（应为"关于印发〈xxx〉的通知"）
|
||
3. 标题中是否有"印发"等动词
|
||
messages:
|
||
pass: 标题文种合规
|
||
fail: 标题文种不合规
|
||
|
||
- rule_id: GW-T-002
|
||
name: 标题不可有"请求"+"请示"重复
|
||
severity: error
|
||
category: 标题
|
||
target: title
|
||
on_missing: skip
|
||
stages:
|
||
- check: regex_forbid
|
||
pattern: '关于请求.*的请示'
|
||
messages:
|
||
pass: ok
|
||
fail: '"请示"已包含"请求"之意，应删去"请求"'
|
||
|
||
- rule_id: GW-T-003
|
||
name: 标题不可有"上报"+"报告"重复
|
||
severity: error
|
||
category: 标题
|
||
target: title
|
||
on_missing: skip
|
||
stages:
|
||
- check: regex_forbid
|
||
pattern: '关于上报.*的报告'
|
||
messages:
|
||
pass: ok
|
||
fail: '"报告"已包含"上报"之意，应删去"上报"'
|
||
|
||
- rule_id: GW-T-004
|
||
name: 标题介词连用
|
||
severity: warning
|
||
category: 标题
|
||
target: title
|
||
on_missing: skip
|
||
stages:
|
||
- check: regex_forbid
|
||
pattern: '关于对.*的(批复|通知|通报)'
|
||
messages:
|
||
pass: ok
|
||
fail: '"关于"+"对" 介词连用不规范'
|
||
|
||
- rule_id: GW-T-005
|
||
name: 标题文种白名单
|
||
severity: error
|
||
category: 文种
|
||
target: wenzhong
|
||
on_missing: skip
|
||
stages:
|
||
- check: wenzhong_whitelist
|
||
messages:
|
||
pass: 文种合规
|
||
fail: 非法定文种（出现"工作情况""汇报""方案""办法"等当文种）
|
||
|
||
- rule_id: GW-T-006
|
||
name: 标题回行词意完整
|
||
severity: warning
|
||
category: 标题
|
||
target: title
|
||
on_missing: skip
|
||
stages:
|
||
- check: ai
|
||
prompt: |
|
||
只在标题里**明确出现破词**时才报错。
|
||
破词示例：「广东省烟草专卖局关于xx的通知」如果在"专"和"卖"之间有换行 → fail
|
||
其它情况（单行标题、合理换行点、词意完整）→ **必须 pass**
|
||
|
||
判断准则：
|
||
- 标题已经是单行字符串，没有明显断点 → pass
|
||
- 不要凭直觉揣测，只判断是否能在原文中**逐字定位**破词位置
|
||
- 找不到具体破词位置就 pass
|
||
|
||
标题原文：
|
||
{{title.text}}
|
||
messages:
|
||
pass: 标题回行合规
|
||
fail: 标题回行破词
|
||
|
||
- group: 发文字号（错误汇编 三、六.3）
|
||
rules:
|
||
- rule_id: GW-N-001
|
||
name: 发文字号必须用六角括号
|
||
severity: error
|
||
category: 发文
|
||
target: doc_number
|
||
on_missing: fail
|
||
stages:
|
||
- check: forbid_chars
|
||
chars: ["[", "]"]
|
||
messages:
|
||
pass: ok
|
||
fail: 发文字号年份应用六角括号「〔〕」，不得使用方括号
|
||
|
||
- rule_id: GW-N-002
|
||
name: 发文字号不可加"第"字
|
||
severity: error
|
||
category: 发文
|
||
target: doc_number
|
||
on_missing: fail
|
||
stages:
|
||
- check: regex_forbid
|
||
pattern: '〔\d{4}〕第\d+号'
|
||
messages:
|
||
pass: ok
|
||
fail: 发文字号顺序号前不应加"第"字
|
||
|
||
- rule_id: GW-N-003
|
||
name: 发文字号顺序号不编虚位
|
||
severity: error
|
||
category: 发文
|
||
target: doc_number
|
||
on_missing: fail
|
||
stages:
|
||
- check: regex_forbid
|
||
pattern: '〔\d{4}〕0\d+号'
|
||
messages:
|
||
pass: ok
|
||
fail: 发文字号顺序号不编虚位（如"02号"应为"2号"）
|
||
|
||
- group: 格式（错误汇编 二）
|
||
rules:
|
||
- rule_id: GW-F-001
|
||
name: 主标题用方正小标宋简体二号
|
||
severity: error
|
||
category: 格式
|
||
target: title
|
||
on_missing: fail
|
||
stages:
|
||
- check: font
|
||
expect:
|
||
eastasia: 方正小标宋简体
|
||
size_pt: 22
|
||
messages:
|
||
pass: ok
|
||
fail: 主标题应使用方正小标宋简体二号
|
||
|
||
- rule_id: GW-F-002
|
||
name: 一级标题用黑体三号
|
||
severity: error
|
||
category: 格式
|
||
applies_to:
|
||
role: heading_1
|
||
on_missing: skip
|
||
stages:
|
||
- check: font
|
||
expect:
|
||
eastasia: 黑体
|
||
size_pt: 16
|
||
messages:
|
||
pass: ok
|
||
fail: 一级标题应使用黑体三号
|
||
|
||
- rule_id: GW-F-003
|
||
name: 二级标题用楷体三号
|
||
severity: error
|
||
category: 格式
|
||
applies_to:
|
||
role: heading_2
|
||
on_missing: skip
|
||
stages:
|
||
- check: font
|
||
expect:
|
||
eastasia: 楷体
|
||
size_pt: 16
|
||
messages:
|
||
pass: ok
|
||
fail: 二级标题应使用楷体三号
|
||
|
||
- rule_id: GW-F-004
|
||
name: 正文用仿宋三号
|
||
severity: warning
|
||
category: 格式
|
||
applies_to:
|
||
role: body
|
||
on_missing: skip
|
||
stages:
|
||
- check: font
|
||
expect:
|
||
eastasia: 仿宋
|
||
size_pt: 16
|
||
messages:
|
||
pass: ok
|
||
fail: 正文应使用仿宋（GB2312）三号
|
||
|
||
- rule_id: GW-F-005
|
||
name: 附件后不加冒号
|
||
severity: error
|
||
category: 格式
|
||
applies_to:
|
||
role: attachment_marker
|
||
on_missing: skip
|
||
stages:
|
||
- check: regex_forbid
|
||
pattern: '^附件\d+：'
|
||
messages:
|
||
pass: ok
|
||
fail: '"附件1"等字样后不应加冒号'
|
||
|
||
- rule_id: GW-F-006
|
||
name: 不使用"（此页无正文）"
|
||
severity: warning
|
||
category: 格式
|
||
applies_to:
|
||
role: any
|
||
on_missing: skip
|
||
stages:
|
||
- check: forbid_phrase
|
||
phrases:
|
||
- （此页无正文）
|
||
- (此页无正文)
|
||
messages:
|
||
pass: ok
|
||
fail: 应通过编辑排版避免出现"（此页无正文）"
|
||
|
||
- rule_id: GW-F-007
|
||
name: 附件项末尾不加标点
|
||
severity: warning
|
||
category: 格式
|
||
applies_to:
|
||
role: any
|
||
on_missing: skip
|
||
stages:
|
||
- check: cross_role
|
||
rules:
|
||
- type: attachment_item_no_trailing_punct
|
||
messages:
|
||
pass: ok
|
||
fail: 附件名称（内容）后不应使用标点符号
|
||
|
||
- rule_id: GW-F-008
|
||
name: 三级标题用仿宋三号
|
||
severity: warning
|
||
category: 格式
|
||
applies_to:
|
||
role: heading_3
|
||
on_missing: skip
|
||
stages:
|
||
- check: font
|
||
expect:
|
||
eastasia: 仿宋
|
||
size_pt: 16
|
||
messages:
|
||
pass: ok
|
||
fail: 三级标题应使用仿宋（GB2312）三号
|
||
|
||
- rule_id: GW-F-009
|
||
name: 四级标题用仿宋三号
|
||
severity: warning
|
||
category: 格式
|
||
applies_to:
|
||
role: heading_4
|
||
on_missing: skip
|
||
stages:
|
||
- check: font
|
||
expect:
|
||
eastasia: 仿宋
|
||
size_pt: 16
|
||
messages:
|
||
pass: ok
|
||
fail: 四级标题应使用仿宋（GB2312）三号
|
||
|
||
- rule_id: GW-F-010
|
||
name: 附件标记用黑体三号不加粗
|
||
severity: error
|
||
category: 格式
|
||
applies_to:
|
||
role: attachment_marker
|
||
on_missing: skip
|
||
stages:
|
||
- check: attachment_marker_style
|
||
expect:
|
||
eastasia: 黑体
|
||
size_pt: 16
|
||
bold: false
|
||
messages:
|
||
pass: ok
|
||
fail: '"附件："或"附件1"等标记应使用黑体三号，且不加粗'
|
||
|
||
- group: 层级序号（错误汇编 四）
|
||
rules:
|
||
- rule_id: GW-H-001
|
||
name: 层级序号格式
|
||
severity: error
|
||
category: 层级
|
||
applies_to:
|
||
role: any
|
||
on_missing: skip
|
||
stages:
|
||
- check: hierarchy
|
||
forbid_patterns:
|
||
- '^[一二三四五六七八九十]+、.*[、。]$'
|
||
- '^\d+、'
|
||
- '^（[一二三四五六七八九十]+）、'
|
||
messages:
|
||
pass: ok
|
||
fail: 层级序号格式错误
|
||
|
||
- rule_id: GW-H-002
|
||
name: 二级标题换行不带句号
|
||
severity: warning
|
||
category: 层级
|
||
applies_to:
|
||
role: heading_2
|
||
on_missing: skip
|
||
stages:
|
||
- check: cross_role
|
||
rules:
|
||
- type: h2_no_period_then_break
|
||
messages:
|
||
pass: ok
|
||
fail: 二级标题在换行分段时不应使用句号
|
||
|
||
- group: 标点符号（错误汇编 六）
|
||
rules:
|
||
- rule_id: GW-P-001
|
||
name: 多书名号/引号并列不加顿号
|
||
severity: warning
|
||
category: 标点
|
||
applies_to:
|
||
role: any
|
||
on_missing: skip
|
||
stages:
|
||
- check: punctuation
|
||
rules:
|
||
- type: no_dunhao_between_quotes
|
||
messages:
|
||
pass: ok
|
||
fail: 多个书名号/引号并列时不应用顿号分隔
|
||
|
||
- rule_id: GW-P-002
|
||
name: 句内括号末尾不加标点
|
||
severity: warning
|
||
category: 标点
|
||
applies_to:
|
||
role: any
|
||
on_missing: skip
|
||
stages:
|
||
- check: punctuation
|
||
rules:
|
||
- type: no_punct_inside_inline_paren
|
||
messages:
|
||
pass: ok
|
||
fail: 句内括号行文末尾通常不应含标点
|
||
|
||
- rule_id: GW-P-003
|
||
name: 引号嵌套不规范
|
||
severity: warning
|
||
category: 标点
|
||
applies_to:
|
||
role: any
|
||
on_missing: skip
|
||
stages:
|
||
- check: punctuation
|
||
rules:
|
||
- type: no_outer_quote_when_inner_quote
|
||
messages:
|
||
pass: ok
|
||
fail: 双引号内已含单引号强调时，外层不应再加双引号（如"卓'粤'创一流"应为 卓"粤"创一流）
|
||
|
||
- group: 文字表述与提法（错误汇编 七、八、九）
|
||
rules:
|
||
- rule_id: GW-W-001
|
||
name: 易混淆词使用
|
||
severity: warning
|
||
category: 文字
|
||
applies_to:
|
||
role: any
|
||
on_missing: skip
|
||
stages:
|
||
- check: confused_pair
|
||
pairs:
|
||
- wrong: 截至到
|
||
correct: 截止到
|
||
reason: '"截至" 已含"到"之意'
|
||
- wrong: 下称
|
||
correct: 以下简称
|
||
reason: 标注简称应用"以下简称"
|
||
- wrong_pattern: '截止\d{4}年'
|
||
suggest: 截至YYYY年
|
||
reason: 用于到某时点应为"截至"
|
||
messages:
|
||
pass: ok
|
||
fail: 易混淆词使用不当
|
||
|
||
- rule_id: GW-W-002
|
||
name: 简称使用规范
|
||
severity: warning
|
||
category: 简称
|
||
applies_to:
|
||
role: body
|
||
on_missing: skip
|
||
stages:
|
||
- check: ai
|
||
prompt: |
|
||
只在文中出现以下两种省级职务简称错误时才报错，否则一律 pass：
|
||
- "X省省委书记" 错误（应为 "X省委书记"，省字不重复）
|
||
- "X省长" 错误（应为 "X省省长"，省字不可省略）
|
||
|
||
若文中没有"省委书记"或"省长"等省级职务字样，**必须 pass**。
|
||
若不能在文中找到准确的错误原文，**必须 pass**。
|
||
不要做语气、措辞、其它简称的检查。
|
||
|
||
全文片段：
|
||
{{paragraphs[0]}}
|
||
messages:
|
||
pass: 简称规范
|
||
fail: 简称使用不规范
|
||
|
||
- rule_id: GW-W-003
|
||
name: 成文日期用阿拉伯数字
|
||
severity: error
|
||
category: 提法
|
||
target: date
|
||
on_missing: fail
|
||
stages:
|
||
- check: regex_forbid
|
||
pattern: '[一二三四五六七八九十○〇零]+年'
|
||
messages:
|
||
pass: ok
|
||
fail: 成文日期应使用阿拉伯数字（如"2023年10月9日"）
|
||
|
||
- rule_id: GW-W-004
|
||
name: 成文日期不编虚位
|
||
severity: warning
|
||
category: 提法
|
||
target: date
|
||
on_missing: fail
|
||
stages:
|
||
- check: regex_forbid
|
||
pattern: '\d{4}年0\d月|\d{4}年\d{1,2}月0\d日'
|
||
messages:
|
||
pass: ok
|
||
fail: 成文日期月、日不编虚位
|
||
|
||
- group: 发文机关（错误汇编 十）
|
||
rules:
|
||
- rule_id: GW-S-001
|
||
name: 发文机关署名不能用简称
|
||
severity: error
|
||
category: 机关
|
||
target: signature
|
||
on_missing: fail
|
||
stages:
|
||
- check: ai
|
||
prompt: |
|
||
判断署名是否含**明确的简称错误**。
|
||
典型错误：
|
||
- "广东省烟草专卖局（公司）" — 用括号缩短两个机关 → 错
|
||
- "省局" / "粤烟" 等单独缩写 → 错
|
||
典型正确：
|
||
- "广东省烟草专卖局" 单独出现 → pass（即使可能存在配套总公司，但单独存在不算简称）
|
||
- "广东省烟草专卖局 中国烟草总公司广东省公司" → pass
|
||
|
||
判断准则：
|
||
- 若署名是一个完整、官方、可独立成立的机关名 → **必须 pass**
|
||
- 若署名带"（公司）"、"省局"、明显缩写、行业内部代号 → fail
|
||
|
||
署名原文：
|
||
{{signature.text}}
|
||
messages:
|
||
pass: 署名规范
|
||
fail: 发文机关署名使用了简称
|
||
|
||
- rule_id: GW-S-002
|
||
name: 发文机关确定严谨性
|
||
severity: warning
|
||
category: 机关
|
||
target: signature
|
||
on_missing: fail
|
||
stages:
|
||
- check: ai
|
||
prompt: |
|
||
只判断**这一个明确条件**：
|
||
- 标题或正文里明确涉及"党组""党的xx工作""组织部""纪委"等党务事项，
|
||
但署名是行政机关（局/公司/委员会等），未署"党组"或党务机构 → fail
|
||
- 其它情况（行政事务、缺乏证据、性质模糊）→ **必须 pass**
|
||
|
||
判断时需要看到**明确的党务关键词**（党组/党委/党的xx会议/党风/反腐倡廉等），
|
||
没有这些关键词就 pass。
|
||
|
||
署名原文：{{signature.text}}
|
||
标题：{{title.text}}
|
||
messages:
|
||
pass: 发文机关一致
|
||
fail: 发文机关与文稿性质不一致
|
||
|
||
- group: 标题字体（target 通道示例）
|
||
rules:
|
||
- rule_id: GW-T-008
|
||
name: 标题字体（语义实体通道）
|
||
severity: warning
|
||
category: 标题
|
||
target: title
|
||
on_missing: warn
|
||
stages:
|
||
- check: ai
|
||
prompt: |
|
||
判断公文标题的字体与字号是否合规。
|
||
要求：字体 = 方正小标宋简体；字号 = 22pt（或 22.0）。
|
||
|
||
实际：
|
||
- 标题：{{title.text}}
|
||
- 字体：{{title.style.font_eastasia}}
|
||
- 字号：{{title.style.font_size_pt}}pt
|
||
|
||
若实际字体为空或与要求一致 → pass
|
||
若字体明显不符（例如 仿宋/楷体/黑体）→ fail
|
||
若仅字号轻微差异 → warn
|
||
messages:
|
||
pass: 标题字体字号合规
|
||
fail: 标题字体或字号不符合 GB/T 9704
|