@deepwhale-cn/dsh-cn-doc-formatter 0.1.0

package/skill/scripts/verify_output.py

@@ -0,0 +1,233 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+verify_output.py — 输出自检
+
+生成 docx / pdf 后核对是否真的符合本 skill 的排版规范。
+原先没有这一步，字体名写错、页边距没生效这类问题不会被发现。
+
+用法
+----
+  python verify_output.py --dir <输出目录> [--title <标题>]
+
+退出码 0 = 全部通过，1 = 有不符合项。
+"""
+import argparse
+import glob
+import os
+import sys
+
+# 规范值（与 SKILL.md 保持一致）
+SPEC = {
+    'margin_top_cm': 3.7,
+    'margin_bottom_cm': 3.5,
+    'margin_left_cm': 2.8,
+    'margin_right_cm': 2.6,
+    'title_pt': 22,          # 二号
+    'h1_pt': 16,             # 三号
+    'body_pt': 14,           # 四号
+    'title_leading_pt': (33, 36),
+    'body_leading_pt': 28,
+    'tol': 0.15,             # 容差
+}
+
+PASS, FAIL, WARN = '✅', '❌', '⚠️ '
+
+
+def check_docx(path):
+    """核对 docx 的页面设置与字号、行距、字体"""
+    issues, notes = [], []
+    try:
+        from docx import Document
+        from docx.oxml.ns import qn
+    except ImportError:
+        return ['未安装 python-docx，跳过 docx 检查'], []
+
+    d = Document(path)
+    s = d.sections[0]
+
+    # 页边距
+    for name, got, want in (
+        ('上边距', s.top_margin.cm, SPEC['margin_top_cm']),
+        ('下边距', s.bottom_margin.cm, SPEC['margin_bottom_cm']),
+        ('左边距', s.left_margin.cm, SPEC['margin_left_cm']),
+        ('右边距', s.right_margin.cm, SPEC['margin_right_cm']),
+    ):
+        if abs(got - want) > SPEC['tol']:
+            issues.append(f'{name} {got:.2f}cm，应为 {want}cm')
+        else:
+            notes.append(f'{name} {got:.2f}cm ✓')
+
+    # 段落：字号 / 行距 / 字体
+    kinds = {}
+    fonts = set()
+    for p in d.paragraphs:
+        t = p.text.strip()
+        if not t or not p.runs:
+            continue
+        r = p.runs[0]
+        sz = r.font.size.pt if r.font.size else None
+        bold = bool(r.font.bold)
+        leading = None
+        pPr = p._element.find(qn('w:pPr'))
+        if pPr is not None:
+            sp = pPr.find(qn('w:spacing'))
+            if sp is not None and sp.get(qn('w:line')):
+                leading = int(sp.get(qn('w:line'))) / 20.0    # twip → pt
+        rPr = r._element.find(qn('w:rPr'))
+        if rPr is not None:
+            rf = rPr.find(qn('w:rFonts'))
+            if rf is not None:
+                ea = rf.get(qn('w:eastAsia'))
+                if ea:
+                    fonts.add(ea)
+        kinds.setdefault((sz, bold), []).append((t[:22], leading))
+
+    # 标题：22pt 加粗
+    if not any(sz == SPEC['title_pt'] and b for (sz, b) in kinds):
+        issues.append(f'未找到 {SPEC["title_pt"]}pt 加粗的标题（二号）')
+    else:
+        notes.append(f'标题 {SPEC["title_pt"]}pt 加粗 ✓')
+
+    # 一级标题：16pt 加粗
+    if not any(sz == SPEC['h1_pt'] and b for (sz, b) in kinds):
+        issues.append(f'未找到 {SPEC["h1_pt"]}pt 加粗的一级标题（三号）')
+    else:
+        notes.append(f'一级标题 {SPEC["h1_pt"]}pt 加粗 ✓')
+
+    # 正文：14pt 不加粗，行距 28pt
+    body_ok = False
+    for (sz, b), items in kinds.items():
+        if sz == SPEC['body_pt'] and not b:
+            leads = [l for _t, l in items if l is not None]
+            if leads:
+                avg = sum(leads) / len(leads)
+                if abs(avg - SPEC['body_leading_pt']) > 1.5:
+                    issues.append(f'正文行距 {avg:.1f}pt，应为 {SPEC["body_leading_pt"]}pt')
+                else:
+                    notes.append(f'正文 {sz}pt / 行距 {avg:.0f}pt ✓')
+                    body_ok = True
+            else:
+                issues.append('正文段落未设置固定行距')
+            break
+    if not body_ok and not any(i.startswith('正文') for i in issues):
+        issues.append(f'未找到 {SPEC["body_pt"]}pt 不加粗的正文段落（四号）')
+
+    # 字体一致性
+    if len(fonts) > 1:
+        issues.append(f'全文出现多种字体：{sorted(fonts)}，规范要求统一仿宋')
+    elif fonts:
+        notes.append(f'全文统一字体：{sorted(fonts)[0]} ✓')
+    else:
+        warnings_only = True
+
+    # 字体是否本机真实存在（避免 Word 静默替换）
+    try:
+        sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
+        from docx_formatter import FOUND_FONTS
+        for f in fonts:
+            if f and f not in FOUND_FONTS:
+                issues.append(f'字体「{f}」本机不存在，Word 会替换为其他字体')
+    except Exception:
+        pass
+
+    return issues, notes
+
+
+def check_pdf(path):
+    """核对 PDF 是否嵌入并使用仿宋"""
+    issues, notes = [], []
+    try:
+        from pypdf import PdfReader
+    except ImportError:
+        return ['未安装 pypdf，跳过 pdf 检查'], []
+
+    r = PdfReader(path)
+    notes.append(f'页数 {len(r.pages)}')
+    embedded = set()
+    for pg in r.pages:
+        try:
+            res = pg.get('/Resources')
+            fonts = res.get('/Font') if res else None
+            if fonts:
+                for _k, v in fonts.items():
+                    o = v.get_object()
+                    base = str(o.get('/BaseFont', ''))
+                    if base:
+                        embedded.add(base)
+        except Exception:
+            pass
+    if embedded:
+        names = ', '.join(sorted(embedded)[:6])
+        notes.append(f'嵌入字体 {len(embedded)} 种：{names}')
+        if not any('Fang' in x or '仿宋' in x or 'Song' in x for x in embedded):
+            issues.append(f'PDF 未使用仿宋字体，实际：{sorted(embedded)}')
+    else:
+        issues.append('PDF 未检出嵌入字体')
+
+    # 尺寸应为 A4
+    try:
+        mb = r.pages[0].mediabox
+        w_cm, h_cm = float(mb.width) / 28.3465, float(mb.height) / 28.3465
+        if abs(w_cm - 21.0) > 0.3 or abs(h_cm - 29.7) > 0.3:
+            issues.append(f'页面尺寸 {w_cm:.1f}×{h_cm:.1f}cm，应为 A4 21.0×29.7cm')
+        else:
+            notes.append(f'页面 A4 {w_cm:.1f}×{h_cm:.1f}cm ✓')
+    except Exception:
+        pass
+
+    return issues, notes
+
+
+def main():
+    ap = argparse.ArgumentParser(description='排版输出自检')
+    ap.add_argument('--dir', required=True, help='输出目录')
+    a = ap.parse_args()
+
+    d = a.dir
+    docx = glob.glob(os.path.join(d, '*.docx'))
+    pdf = glob.glob(os.path.join(d, '*.pdf'))
+    txt = glob.glob(os.path.join(d, '*.txt'))
+
+    all_issues = []
+    print('════ 排版输出自检 ════')
+    print(f'  目录：{d}')
+    print(f'  docx {len(docx)} / pdf {len(pdf)} / txt {len(txt)}')
+    print()
+
+    for f in docx:
+        print(f'  ── {os.path.basename(f)}')
+        iss, notes = check_docx(f)
+        for n in notes:
+            print(f'     {PASS} {n}')
+        for i in iss:
+            print(f'     {FAIL} {i}')
+        all_issues += iss
+        print()
+
+    for f in pdf:
+        print(f'  ── {os.path.basename(f)}')
+        iss, notes = check_pdf(f)
+        for n in notes:
+            print(f'     {PASS} {n}')
+        for i in iss:
+            print(f'     {FAIL} {i}')
+        all_issues += iss
+        print()
+
+    if not docx and not pdf:
+        print('  ⚠️  目录下没有 docx / pdf，无法自检')
+        return 1
+
+    print('════ 结果 ════')
+    if all_issues:
+        print(f'  {FAIL} {len(all_issues)} 项不符合规范：')
+        for i in all_issues:
+            print(f'     · {i}')
+        return 1
+    print(f'  {PASS} 全部符合排版规范')
+    return 0
+
+
+if __name__ == '__main__':
+    sys.exit(main())