create-dwy 0.16.0 → 0.17.0

package/templates/ai-tools/skills//350/207/252/345/252/222/344/275/223/uc-media-0030-facts/scripts/generate_html.py

@@ -0,0 +1,1287 @@
+#!/usr/bin/env python3
+"""Generate a Markdown report + interactive HTML viewer.
+
+Outputs:
+  - README.md    (clean Markdown, single source of truth)
+  - index.html   (self-contained viewer: CSS + JS + embedded MD content)
+  - sources.json (structured source data)
+"""
+
+import json
+import re
+import sys
+import argparse
+from datetime import datetime
+from pathlib import Path
+from html import escape
+
+
+def slugify(text):
+    """Generate a URL-safe slug from heading text."""
+    text = re.sub(r'\*\*([^*]+)\*\*', r'\1', text)
+    text = re.sub(r'\*([^*]+)\*', r'\1', text)
+    text = re.sub(r'`([^`]+)`', r'\1', text)
+    text = re.sub(r'[^\w\u4e00-\u9fff]+', '-', text.strip()).lower().strip('-')
+    text = re.sub(r'-+', '-', text)[:60]
+    return text
+
+
+def normalize_source(src):
+    """Normalize source entries to {title, url, type, ...}.
+
+    Supports alternate shapes like {project, urls:[...]} used by some angle
+    research dumps. Missing fields must not break the HTML viewer.
+    """
+    if not isinstance(src, dict):
+        return {'title': str(src), 'url': '#', 'type': 'community'}
+
+    urls = src.get('urls') or []
+    if not isinstance(urls, list):
+        urls = [urls] if urls else []
+
+    url = src.get('url') or (urls[0] if urls else '#') or '#'
+    title = src.get('title') or src.get('project') or url or '未知'
+    out = {
+        'title': title,
+        'url': url,
+        'type': src.get('type') or ('primary' if src.get('project') else 'community'),
+    }
+    if src.get('note'):
+        out['note'] = src['note']
+    elif len(urls) > 1:
+        out['note'] = '相关链接：' + ' · '.join(str(u) for u in urls[1:])
+    return out
+
+
+def format_project_details(projects):
+    """Turn a list of project license research objects into readable Markdown."""
+    lines = []
+    for p in projects:
+        if not isinstance(p, dict):
+            lines.append(f'- {p}')
+            continue
+        name = p.get('project', '未知项目')
+        lines += [f'### {name}', '']
+        if p.get('repo_url'):
+            lines.append(f'- **仓库**：[{p["repo_url"]}]({p["repo_url"]})')
+        if p.get('license_name'):
+            lines.append(f'- **协议**：{p["license_name"]}')
+        meta = p.get('github_license_metadata') or {}
+        if isinstance(meta, dict) and (meta.get('spdx_id') or meta.get('name')):
+            spdx = meta.get('spdx_id', '')
+            mname = meta.get('name', '')
+            note = meta.get('note', '')
+            meta_line = f'- **GitHub 元数据**：{mname}'
+            if spdx:
+                meta_line += f' (`{spdx}`)'
+            if note:
+                meta_line += f' — {note}'
+            lines.append(meta_line)
+        if 'is_osi_open_source' in p:
+            lines.append(f'- **是否 OSI 开源**：{"是" if p["is_osi_open_source"] else "否"}')
+        if p.get('license_file'):
+            lines.append(f'- **LICENSE 文件**：[{p["license_file"]}]({p["license_file"]})')
+        faq = p.get('official_faq_is_remotion_open_source')
+        if isinstance(faq, dict):
+            q = faq.get('question', '')
+            a = faq.get('answer_verbatim_summary', '')
+            lines.append(f'- **官方 FAQ**：{q} → {a}')
+            if faq.get('source'):
+                lines.append(f'  - 来源：[{faq["source"]}]({faq["source"]})')
+        if p.get('verification_method'):
+            lines.append(f'- **核实方式**：{p["verification_method"]}')
+        points = p.get('license_points') or []
+        if points:
+            lines += ['', '**要点**：']
+            for pt in points:
+                lines.append(f'- {pt}')
+        quotes = p.get('quotes') or []
+        if quotes:
+            lines += ['', '**原文摘录**：']
+            for q in quotes:
+                if not isinstance(q, dict):
+                    lines.append(f'> {q}')
+                    lines.append('')
+                    continue
+                text = (q.get('text') or '').replace('\n', ' ').strip()
+                qurl = q.get('url') or ''
+                lines.append(f'> {text}')
+                if qurl:
+                    lines.append('>')
+                    lines.append(f'> — [{qurl}]({qurl})')
+                lines.append('')
+        lines.append('')
+    return '\n'.join(lines).rstrip() + '\n'
+
+
+def details_to_markdown(details):
+    """Coerce angle details (str / list / dict / JSON string) into Markdown."""
+    if details is None:
+        return '（内容待补充）'
+
+    if isinstance(details, str):
+        s = details.strip()
+        if s.startswith('[') or s.startswith('{'):
+            try:
+                return details_to_markdown(json.loads(s))
+            except json.JSONDecodeError:
+                return details
+        return details
+
+    if isinstance(details, list):
+        if details and isinstance(details[0], dict) and 'project' in details[0]:
+            return format_project_details(details)
+        parts = []
+        for item in details:
+            if isinstance(item, str):
+                parts.append(item)
+            else:
+                parts.append(json.dumps(item, ensure_ascii=False, indent=2))
+        return '\n\n'.join(parts)
+
+    if isinstance(details, dict):
+        if 'project' in details:
+            return format_project_details([details])
+        return '```json\n' + json.dumps(details, ensure_ascii=False, indent=2) + '\n```'
+
+    return str(details)
+
+
+def generate_markdown(topic, angles_data):
+    """Generate clean Markdown report (single source of truth)."""
+    today = datetime.now().strftime('%Y-%m-%d')
+    angles_list = '、'.join(a['angle'] for a in angles_data)
+
+    lines = [
+        f'# {topic}',
+        '',
+        f'> 整理时间：{today}',
+        f'> 覆盖角度：{angles_list}',
+        '',
+        '---',
+        '',
+        '## 目录',
+        '1. [概述](#概述)',
+    ]
+    for i, a in enumerate(angles_data, 2):
+        anchor = slugify(a['angle'])
+        lines.append(f'{i}. [{a["angle"]}](#{anchor})')
+    lines.append(f'{len(angles_data) + 2}. [参考资料](#参考资料)')
+    lines += ['', '---', '', '## 概述', '']
+
+    for a in angles_data:
+        lines.append(f'**{a["angle"]}**：{a.get("summary", "")}')
+    lines.append('')
+
+    for i, a in enumerate(angles_data, 1):
+        lines += ['', '---', '', f'## {a["angle"]}', '']
+        details = details_to_markdown(a.get('details', a.get('summary', '（内容待补充）')))
+        # Remove leading H1 if present (it's the angle title)
+        details = re.sub(r'^#\s+.+\s*', '', details, count=1, flags=re.MULTILINE)
+        lines.append(details)
+        lines.append('')
+        if a.get('key_points'):
+            lines.append('### 关键要点')
+            lines.append('')
+            for kp in a['key_points']:
+                lines.append(f'- {kp}')
+            lines.append('')
+
+    # Sources (normalized so HTML citations never see missing title/url)
+    all_sources = []
+    for a in angles_data:
+        for src in a.get('sources', []):
+            all_sources.append(normalize_source(src))
+
+    lines += ['', '---', '', '## 参考资料', '', '### 来源列表', '']
+    for i, src in enumerate(all_sources, 1):
+        t = src.get('type', 'community')
+        title = src.get('title', src.get('url', '未知'))
+        url = src.get('url', '#')
+        lines.append(f'[{i}] [{title}]({url}) `[{t}]`')
+    lines.append('')
+    lines.append('---')
+    lines.append('')
+    lines.append('*本笔记由 AI 调研生成，所有事实均标注来源。建议独立验证关键信息。*')
+    lines.append('')
+
+    return '\n'.join(lines), all_sources
+
+
+HTML_TEMPLATE = """\
+<!DOCTYPE html>
+<html lang="zh-CN">
+<head>
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<title>{title}</title>
+<style>
+:root {{
+  --bg: #fafafa;
+  --card-bg: #ffffff;
+  --text: #1a1a1a;
+  --text-secondary: #666;
+  --border: #e5e5e5;
+  --accent: #2563eb;
+  --accent-light: #dbeafe;
+  --cite-bg: #f0f9ff;
+  --cite-border: #93c5fd;
+  --popup-bg: #1e293b;
+  --popup-text: #f8fafc;
+  --popup-link: #93c5fd;
+  --code-bg: #f1f5f9;
+  --table-stripe: #f8fafc;
+  --sidebar-width: 240px;
+}}
+
+@media (prefers-color-scheme: dark) {{
+  :root {{
+    --bg: #0f172a;
+    --card-bg: #1e293b;
+    --text: #e2e8f0;
+    --text-secondary: #94a3b8;
+    --border: #334155;
+    --accent: #60a5fa;
+    --accent-light: #1e3a5f;
+    --cite-bg: #1e3a5f;
+    --cite-border: #3b82f6;
+    --popup-bg: #0f172a;
+    --popup-text: #f1f5f9;
+    --popup-link: #60a5fa;
+    --code-bg: #1e293b;
+    --table-stripe: #1a2332;
+  }}
+}}
+
+* {{ margin: 0; padding: 0; box-sizing: border-box; }}
+
+body {{
+  font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", "Noto Sans SC", "PingFang SC", "Microsoft YaHei", sans-serif;
+  background: var(--bg);
+  color: var(--text);
+  line-height: 1.8;
+}}
+
+/* Layout: sidebar + main */
+.layout {{
+  display: flex;
+  max-width: 1120px;
+  margin: 0 auto;
+}}
+
+.sidebar {{
+  width: var(--sidebar-width);
+  flex-shrink: 0;
+  position: sticky;
+  top: 0;
+  height: 100vh;
+  overflow-y: auto;
+  padding: 2rem 1rem 2rem 1.5rem;
+  border-right: 1px solid var(--border);
+}}
+
+.main-content {{
+  flex: 1;
+  max-width: 800px;
+  padding: 2rem 3rem;
+}}
+
+@media (max-width: 900px) {{
+  .sidebar {{ display: none; }}
+  .main-content {{ padding: 1.5rem; }}
+}}
+
+/* Sidebar TOC */
+.sidebar-title {{
+  font-size: 0.875rem;
+  font-weight: 600;
+  text-transform: uppercase;
+  letter-spacing: 0.05em;
+  color: var(--text-secondary);
+  margin-bottom: 1rem;
+}}
+
+.toc-nav {{
+  list-style: none;
+  padding: 0;
+}}
+
+.toc-nav li {{
+  margin: 0;
+}}
+
+.toc-nav a {{
+  display: block;
+  padding: 0.4rem 0.5rem;
+  font-size: 0.9375rem;
+  color: var(--text-secondary);
+  text-decoration: none;
+  border-radius: 4px;
+  border-left: 2px solid transparent;
+  transition: all 0.15s ease;
+  line-height: 1.5;
+}}
+
+.toc-nav a:hover {{
+  color: var(--text);
+  background: var(--accent-light);
+}}
+
+.toc-nav a.active {{
+  color: var(--accent);
+  border-left-color: var(--accent);
+  font-weight: 500;
+}}
+
+/* TOC indentation levels */
+.toc-level-1 a {{
+  font-weight: 600;
+  font-size: 1rem;
+  padding-left: 0.5rem;
+}}
+
+.toc-level-2 a {{
+  padding-left: 1.25rem;
+  font-size: 0.9375rem;
+}}
+
+.toc-level-3 a {{
+  padding-left: 2rem;
+  font-size: 0.875rem;
+  opacity: 0.85;
+}}
+
+.toc-level-4 a {{
+  padding-left: 2.75rem;
+  font-size: 0.8125rem;
+  opacity: 0.7;
+}}
+
+.toc-nav {{
+  list-style: none;
+  padding: 0;
+}}
+
+.toc-nav li {{
+  margin: 0;
+}}
+
+.toc-nav a {{
+  display: block;
+  padding: 0.3rem 0.5rem;
+  font-size: 0.8125rem;
+  color: var(--text-secondary);
+  text-decoration: none;
+  border-radius: 4px;
+  border-left: 2px solid transparent;
+  transition: all 0.15s ease;
+  line-height: 1.4;
+}}
+
+.toc-nav a:hover {{
+  color: var(--text);
+  background: var(--accent-light);
+}}
+
+.toc-nav a.active {{
+  color: var(--accent);
+  border-left-color: var(--accent);
+  font-weight: 500;
+}}
+
+.toc-level-1 a {{
+  font-weight: 600;
+  font-size: 0.875rem;
+  padding-left: 0.5rem;
+}}
+
+.toc-level-2 a {{
+  padding-left: 1.25rem;
+  font-size: 0.8125rem;
+}}
+
+.toc-level-3 a {{
+  padding-left: 2rem;
+  font-size: 0.75rem;
+  opacity: 0.85;
+}}
+
+.toc-level-4 a {{
+  padding-left: 2.75rem;
+  font-size: 0.6875rem;
+  opacity: 0.7;
+}}
+
+/* Typography */
+h1 {{
+  font-size: 2rem;
+  font-weight: 700;
+  margin-bottom: 0.5rem;
+  line-height: 1.3;
+}}
+
+.meta {{
+  color: var(--text-secondary);
+  font-size: 0.875rem;
+  margin-bottom: 2rem;
+  padding-bottom: 1.5rem;
+  border-bottom: 1px solid var(--border);
+}}
+
+h2 {{
+  font-size: 1.5rem;
+  font-weight: 600;
+  margin: 2.5rem 0 1rem;
+  padding-bottom: 0.5rem;
+  border-bottom: 2px solid var(--accent-light);
+}}
+
+h3 {{
+  font-size: 1.125rem;
+  font-weight: 600;
+  margin: 1.5rem 0 0.75rem;
+}}
+
+h4 {{
+  font-size: 1rem;
+  font-weight: 600;
+  margin: 1.25rem 0 0.5rem;
+  color: var(--text-secondary);
+}}
+
+h5 {{
+  font-size: 0.9375rem;
+  font-weight: 600;
+  margin: 1rem 0 0.5rem;
+  color: var(--text-secondary);
+  opacity: 0.9;
+}}
+
+p {{
+  margin-bottom: 1rem;
+}}
+
+a {{
+  color: var(--accent);
+  text-decoration: none;
+}}
+a:hover {{ text-decoration: underline; }}
+
+blockquote {{
+  border-left: 3px solid var(--accent);
+  padding-left: 1rem;
+  color: var(--text-secondary);
+  margin: 1rem 0;
+}}
+
+table {{
+  width: 100%;
+  border-collapse: collapse;
+  margin: 1rem 0;
+  font-size: 0.9375rem;
+}}
+
+th, td {{
+  padding: 0.75rem 1rem;
+  border: 1px solid var(--border);
+  text-align: left;
+}}
+
+th {{
+  background: var(--accent-light);
+  font-weight: 600;
+}}
+
+tr:nth-child(even) td {{
+  background: var(--table-stripe);
+}}
+
+code {{
+  background: var(--code-bg);
+  padding: 0.15rem 0.4rem;
+  border-radius: 4px;
+  font-size: 0.875em;
+  font-family: "SF Mono", "Fira Code", "Consolas", monospace;
+}}
+
+pre {{
+  background: var(--code-bg);
+  padding: 1rem 1.25rem;
+  border-radius: 8px;
+  overflow-x: auto;
+  margin: 1rem 0;
+  border: 1px solid var(--border);
+}}
+
+pre code {{
+  background: none;
+  padding: 0;
+}}
+
+ul, ol {{
+  padding-left: 1.5rem;
+  margin-bottom: 1rem;
+}}
+
+li {{
+  margin: 0.35rem 0;
+}}
+
+/* Citation link */
+a.cit {{
+  display: inline;
+  color: var(--accent);
+  background: var(--cite-bg);
+  border-bottom: 1px dashed var(--cite-border);
+  cursor: pointer;
+  padding: 0 1px;
+  text-decoration: none;
+  border-radius: 2px;
+}}
+
+a.cit:hover {{
+  background: var(--cite-border);
+  color: var(--card-bg);
+  text-decoration: none;
+}}
+
+/* Image zoom */
+.content img {{
+  cursor: zoom-in;
+  transition: opacity 0.2s ease;
+}}
+.content img:hover {{
+  opacity: 0.9;
+}}
+
+/* Image lightbox */
+#img-lightbox {{
+  display: none;
+  position: fixed;
+  z-index: 2000;
+  top: 0;
+  left: 0;
+  width: 100%;
+  height: 100%;
+  background: rgba(0, 0, 0, 0.85);
+  justify-content: center;
+  align-items: center;
+  cursor: zoom-out;
+}}
+#img-lightbox.visible {{
+  display: flex;
+}}
+#img-lightbox img {{
+  max-width: 90vw;
+  max-height: 90vh;
+  object-fit: contain;
+  border-radius: 4px;
+  box-shadow: 0 20px 60px rgba(0, 0, 0, 0.5);
+}}
+
+/* Source popup */
+#source-popup {{
+  display: none;
+  position: fixed;
+  z-index: 1000;
+  background: var(--popup-bg);
+  color: var(--popup-text);
+  padding: 0.75rem 1rem;
+  border-radius: 8px;
+  font-size: 0.8125rem;
+  line-height: 1.5;
+  max-width: 320px;
+  box-shadow: 0 10px 25px rgba(0,0,0,0.3);
+  pointer-events: none;
+  opacity: 0;
+  transition: opacity 0.15s ease;
+}}
+
+#source-popup.visible {{
+  display: block;
+  opacity: 1;
+}}
+
+#source-popup.pinned {{
+  pointer-events: auto;
+}}
+
+#source-popup a {{
+  color: var(--popup-link);
+  text-decoration: underline;
+}}
+
+#source-popup .popup-title {{
+  font-weight: 600;
+  margin-bottom: 0.25rem;
+}}
+
+#source-popup .popup-url {{
+  opacity: 0.7;
+  font-size: 0.75rem;
+  word-break: break-all;
+}}
+
+#source-popup .popup-hint {{
+  margin-top: 0.35rem;
+  opacity: 0.5;
+  font-size: 0.6875rem;
+}}
+
+/* Sources section */
+.sources-list {{
+  padding-left: 1.25rem;
+}}
+
+.sources-list li {{
+  margin: 0.5rem 0;
+  font-size: 0.9375rem;
+}}
+
+.source-type {{
+  display: inline-block;
+  font-size: 0.6875rem;
+  padding: 0.1rem 0.5rem;
+  border-radius: 10px;
+  margin-left: 0.5rem;
+  vertical-align: middle;
+  opacity: 0.7;
+}}
+
+.source-type.official {{ background: #dcfce7; color: #166534; }}
+.source-type.wikipedia {{ background: #dbeafe; color: #1e40af; }}
+.source-type.academic {{ background: #fef3c7; color: #92400e; }}
+.source-type.community {{ background: #f3e8ff; color: #6b21a8; }}
+.source-type.media {{ background: #ffe4e6; color: #9f1239; }}
+
+.footer {{
+  margin-top: 3rem;
+  padding-top: 1.5rem;
+  border-top: 1px solid var(--border);
+  color: var(--text-secondary);
+  font-size: 0.8125rem;
+  text-align: center;
+}}
+
+::selection {{
+  background: var(--accent-light);
+}}
+</style>
+</head>
+<body>
+
+<div class="layout">
+
+<!-- Sidebar TOC -->
+<aside class="sidebar">
+<div class="sidebar-title">目录</div>
+<ol class="toc-nav">
+{toc_html}
+</ol>
+</aside>
+
+<!-- Main content -->
+<main class="main-content" id="content">
+<!-- Content rendered from Markdown -->
+</main>
+
+</div>
+
+<!-- Source popup -->
+<div id="source-popup">
+  <div class="popup-title" id="popup-title"></div>
+  <div class="popup-url" id="popup-url"></div>
+  <div class="popup-hint">点击链接访问原文</div>
+</div>
+
+<!-- Image lightbox -->
+<div id="img-lightbox">
+  <img src="" alt="Zoomed image" id="lightbox-img">
+</div>
+
+<script>
+// === Embedded Markdown content ===
+const MARKDOWN_CONTENT = {md_content_json};
+
+// === Source data ===
+const sources = {sources_json};
+
+function sourceTitle(s) {{
+  if (!s) return '来源';
+  return String(s.title || s.project || s.url || (s.urls && s.urls[0]) || '来源');
+}}
+function sourceUrl(s) {{
+  if (!s) return '#';
+  return s.url || (s.urls && s.urls[0]) || '#';
+}}
+
+// === Markdown parser ===
+function parseMarkdown(md) {{
+  const lines = md.split('\\n');
+  let html = '';
+  let inTable = false;
+  let inCode = false;
+  let inUl = false;
+  let inOl = false;
+
+  function escapeHtml(text) {{
+    return text.replace(/&/g, '&amp;').replace(/</g, '&lt;').replace(/>/g, '&gt;');
+  }}
+
+  function slugify(text) {{
+    text = text.replace(/\\*\\*([^*]+)\\*\\*/g, '$1');
+    text = text.replace(/\\*([^*]+)\\*/g, '$1');
+    text = text.replace(/`([^`]+)`/g, '$1');
+    return text.replace(/[^\\w\\u4e00-\\u9fff]+/g, '-').toLowerCase().replace(/^-|-$/g, '').replace(/-+/g, '-').substring(0, 60);
+  }}
+
+  function renderInline(text) {{
+    // Code
+    text = text.replace(/`([^`]+)`/g, '<code>$1</code>');
+    // Images: ![alt](url) → <img>
+    text = text.replace(/!\\[([^\\]]*)\\]\\(([^)]+)\\)/g, function(m, alt, url) {{
+      const isLocal = !url.startsWith('http');
+      const src = url;
+      return '<img src="' + src + '" alt="' + alt + '" style="max-width:100%;border-radius:8px;margin:1rem 0;border:1px solid var(--border)">';
+    }});
+    // Links
+    text = text.replace(/\\[([^\\]]+)\\]\\(([^)]+)\\)/g, '<a href="$2" target="_blank" rel="noopener">$1</a>');
+    // Bold
+    text = text.replace(/\\*\\*([^*]+)\\*\\*/g, '<strong>$1</strong>');
+    // Italic
+    text = text.replace(/\\*([^*]+)\\*/g, '<em>$1</em>');
+    return text;
+  }}
+
+  function renderCitations(text) {{
+    return text.replace(/\\[(\\d+)\\]/g, function(m, idx) {{
+      const i = parseInt(idx) - 1;
+      if (i >= 0 && i < sources.length) {{
+        const s = sources[i];
+        const title = sourceTitle(s).substring(0, 40).replace(/"/g, '&quot;');
+        const url = sourceUrl(s);
+        return '<a class="cit" href="' + url + '" target="_blank" rel="noopener" data-source-idx="' + i + '" title="' + title + '">[' + idx + ']</a>';
+      }}
+      return m;
+    }});
+  }}
+
+  function closeLists() {{
+    if (inUl) {{ html += '</ul>'; inUl = false; }}
+    if (inOl) {{ html += '</ol>'; inOl = false; }}
+    if (inTable) {{ html += '</table>'; inTable = false; }}
+  }}
+
+  for (let li = 0; li < lines.length; li++) {{
+    const line = lines[li];
+    const stripped = line.trim();
+
+    // Code blocks
+    if (stripped.startsWith('```')) {{
+      closeLists();
+      if (inCode) {{
+        html += '</code></pre>';
+        inCode = false;
+      }} else {{
+        const lang = stripped.substring(3).trim();
+        html += '<pre><code' + (lang ? ' class="language-' + lang + '"' : '') + '>';
+        inCode = true;
+      }}
+      continue;
+    }}
+    if (inCode) {{
+      html += escapeHtml(line) + '\\n';
+      continue;
+    }}
+
+    // Empty line
+    if (!stripped) {{
+      closeLists();
+      continue;
+    }}
+
+    // Horizontal rule
+    if (stripped === '---') {{
+      closeLists();
+      html += '<hr style="border:none;border-top:1px solid var(--border);margin:2rem 0">';
+      continue;
+    }}
+
+    // Headings
+    const h5Match = stripped.match(/^#####\\s+(.+)$/);
+    const h4Match = stripped.match(/^####\\s+(.+)$/);
+    const h3Match = stripped.match(/^###\\s+(.+)$/);
+    const h2Match = stripped.match(/^##\\s+(.+)$/);
+    const h1Match = stripped.match(/^#\\s+(.+)$/);
+
+    if (h1Match) {{
+      closeLists();
+      const id = slugify(h1Match[1]);
+      html += '<h1 id="' + id + '">' + renderCitations(renderInline(h1Match[1])) + '</h1>';
+      continue;
+    }}
+    if (h2Match) {{
+      closeLists();
+      const id = slugify(h2Match[1]);
+      html += '<h2 id="' + id + '">' + renderCitations(renderInline(h2Match[1])) + '</h2>';
+      continue;
+    }}
+    if (h3Match) {{
+      closeLists();
+      const id = slugify(h3Match[1]);
+      html += '<h3 id="' + id + '">' + renderCitations(renderInline(h3Match[1])) + '</h3>';
+      continue;
+    }}
+    if (h4Match) {{
+      closeLists();
+      const id = slugify(h4Match[1]);
+      html += '<h4 id="' + id + '">' + renderCitations(renderInline(h4Match[1])) + '</h4>';
+      continue;
+    }}
+    if (h5Match) {{
+      closeLists();
+      const id = slugify(h5Match[1]);
+      html += '<h5 id="' + id + '">' + renderCitations(renderInline(h5Match[1])) + '</h5>';
+      continue;
+    }}
+
+    // Table
+    if (stripped.includes('|') && stripped.startsWith('|')) {{
+      const cells = stripped.replace(/^\\||\\|$/g, '').split('|').map(c => c.trim());
+      if (cells.every(c => /^[-:]+$/.test(c))) continue; // skip separator
+      if (!inTable) {{
+        html += '<table><thead><tr>';
+        cells.forEach(c => {{ html += '<th>' + renderCitations(renderInline(c)) + '</th>'; }});
+        html += '</tr></thead><tbody>';
+        inTable = true;
+      }} else {{
+        html += '<tr>';
+        cells.forEach(c => {{ html += '<td>' + renderCitations(renderInline(c)) + '</td>'; }});
+        html += '</tr>';
+      }}
+      continue;
+    }} else {{
+      if (inTable) {{ html += '</table>'; inTable = false; }}
+    }}
+
+    // Unordered list
+    if (/^[-*]\\s+/.test(stripped)) {{
+      if (!inUl) {{
+        if (inOl) {{ html += '</ol>'; inOl = false; }}
+        html += '<ul>';
+        inUl = true;
+      }}
+      html += '<li>' + renderCitations(renderInline(stripped.replace(/^[-*]\\s+/, ''))) + '</li>';
+      continue;
+    }}
+
+    // Ordered list
+    if (/^\\d+\\.\\s+/.test(stripped)) {{
+      if (!inOl) {{
+        if (inUl) {{ html += '</ul>'; inUl = false; }}
+        html += '<ol>';
+        inOl = true;
+      }}
+      html += '<li>' + renderCitations(renderInline(stripped.replace(/^\\d+\\.\\s+/, ''))) + '</li>';
+      continue;
+    }}
+
+    // Blockquote
+    if (stripped.startsWith('> ')) {{
+      closeLists();
+      html += '<blockquote>' + renderCitations(renderInline(stripped.substring(2))) + '</blockquote>';
+      continue;
+    }}
+
+    // Paragraph
+    closeLists();
+    html += '<p>' + renderCitations(renderInline(stripped)) + '</p>';
+  }}
+  closeLists();
+  return html;
+}}
+
+// === Render content ===
+document.getElementById('content').innerHTML = parseMarkdown(MARKDOWN_CONTENT);
+
+// === Image lightbox ===
+(function() {{
+  const lightbox = document.getElementById('img-lightbox');
+  const lightboxImg = document.getElementById('lightbox-img');
+  document.addEventListener('click', function(evt) {{
+    const img = evt.target.closest('.content img');
+    if (img) {{
+      lightboxImg.src = img.src;
+      lightboxImg.alt = img.alt;
+      lightbox.classList.add('visible');
+    }} else if (evt.target.closest('#img-lightbox')) {{
+      lightbox.classList.remove('visible');
+      lightboxImg.src = '';
+    }}
+  }});
+  document.addEventListener('keydown', function(evt) {{
+    if (evt.key === 'Escape') lightbox.classList.remove('visible');
+  }});
+}})();
+
+// === Popup interaction ===
+(function() {{
+  const popup = document.getElementById('source-popup');
+  const popupTitle = document.getElementById('popup-title');
+  const popupUrl = document.getElementById('popup-url');
+  let pinnedSource = null;
+
+  function showPopup(el) {{
+    const idx = parseInt(el.dataset.sourceIdx);
+    const s = sources[idx];
+    if (!s) return;
+    const title = sourceTitle(s);
+    const url = sourceUrl(s);
+    popupTitle.textContent = title;
+    popupUrl.innerHTML = '<a href="' + url + '" target="_blank" rel="noopener">' + url + '</a>';
+    const rect = el.getBoundingClientRect();
+    let x = rect.left;
+    let y = rect.bottom + 8;
+    if (x + 320 > window.innerWidth) x = window.innerWidth - 330;
+    if (y + 100 > window.innerHeight) y = rect.top - 100;
+    popup.style.left = x + 'px';
+    popup.style.top = y + 'px';
+    popup.classList.add('visible');
+    pinnedSource = idx;
+  }}
+
+  function hidePopup() {{
+    popup.classList.remove('visible');
+    popup.classList.remove('pinned');
+    pinnedSource = null;
+  }}
+
+  // Click on citation: toggle pin
+  document.addEventListener('click', function(evt) {{
+    const cit = evt.target.closest('.cit');
+    if (cit) {{
+      evt.preventDefault();
+      evt.stopPropagation();
+      const idx = parseInt(cit.dataset.sourceIdx);
+      if (pinnedSource === idx) {{
+        hidePopup();
+      }} else {{
+        showPopup(cit);
+        popup.classList.add('pinned');
+      }}
+      return;
+    }}
+    // Click inside pinned popup → allow link navigation
+    if (evt.target.closest('#source-popup')) return;
+    // Click elsewhere → close
+    hidePopup();
+  }});
+
+  // Hover: preview (only when not pinned)
+  document.addEventListener('mouseover', function(evt) {{
+    const cit = evt.target.closest('.cit');
+    if (cit && !pinnedSource) showPopup(cit);
+  }});
+  document.addEventListener('mouseout', function(evt) {{
+    const cit = evt.target.closest('.cit');
+    if (cit && !pinnedSource) hidePopup();
+  }});
+
+  // Selection: detect citation
+  document.addEventListener('mouseup', function(evt) {{
+    const sel = window.getSelection();
+    if (sel.isCollapsed) return;
+    const range = sel.getRangeAt(0);
+    let node = range.startContainer;
+    if (node.nodeType === 3) node = node.parentNode;
+    const cit = node.closest('.cit');
+    if (cit && !pinnedSource) showPopup(cit);
+  }});
+
+  // Escape closes
+  document.addEventListener('keydown', function(evt) {{
+    if (evt.key === 'Escape') hidePopup();
+  }});
+}})();
+
+// === Scroll-spy ===
+(function() {{
+  const tocLinks = document.querySelectorAll('.toc-nav a');
+  const headingElements = new Map();
+  tocLinks.forEach(function(link) {{
+    const targetId = link.getAttribute('href').slice(1);
+    const el = document.getElementById(targetId);
+    if (el) headingElements.set(targetId, el);
+  }});
+
+  function updateActiveHeading() {{
+    let current = '';
+    const scrollPos = window.scrollY + 100;
+    headingElements.forEach(function(el, id) {{
+      if (el.offsetTop <= scrollPos) current = id;
+    }});
+    tocLinks.forEach(function(link) {{
+      link.classList.toggle('active', link.getAttribute('href') === '#' + current);
+    }});
+  }}
+
+  window.addEventListener('scroll', updateActiveHeading, {{ passive: true }});
+  updateActiveHeading();
+}})();
+</script>
+
+</body>
+</html>
+"""
+
+
+def extract_headings_for_toc(md_text):
+    """Extract headings from markdown for hierarchical TOC."""
+    headings = []
+    in_code = False
+    for line in md_text.split('\n'):
+        stripped = line.strip()
+        if stripped.startswith('```'):
+            in_code = not in_code
+            continue
+        if in_code:
+            continue
+        if stripped.startswith('##### '):
+            headings.append((5, stripped[6:].strip()))
+        elif stripped.startswith('#### '):
+            headings.append((4, stripped[5:].strip()))
+        elif stripped.startswith('### '):
+            headings.append((3, stripped[4:].strip()))
+        elif stripped.startswith('## '):
+            headings.append((2, stripped[3:].strip()))
+        elif stripped.startswith('# '):
+            headings.append((1, stripped[2:].strip()))
+    return headings
+
+
+def generate_html(topic, angles_data, md_content=None):
+    """Generate HTML template from angle data.
+    If md_content is provided, uses it; otherwise generates from angles_data.
+    Returns (html_content, all_sources).
+    """
+    if md_content is None:
+        md_content, all_sources = generate_markdown(topic, angles_data)
+    else:
+        all_sources = []
+        for a in angles_data:
+            for src in a.get('sources', []):
+                all_sources.append(normalize_source(src))
+
+    # Extract headings for TOC
+    headings = extract_headings_for_toc(md_content)
+
+    # Build hierarchical TOC HTML
+    toc_html = '<li class="toc-level-1"><a href="#概述">概述</a></li>\n'
+    for level, title in headings:
+        hid = slugify(title)
+        toc_level = min(level + 1, 4)
+        toc_html += f'<li class="toc-level-{toc_level}"><a href="#{hid}">{escape(title)}</a></li>\n'
+
+    # Embed markdown as JSON string
+    md_json = json.dumps(md_content, ensure_ascii=False)
+    sources_json_str = json.dumps(all_sources, ensure_ascii=False)
+
+    html = HTML_TEMPLATE.format(
+        title=escape(topic),
+        toc_html=toc_html,
+        md_content_json=md_json,
+        sources_json=sources_json_str,
+    )
+
+    return html, all_sources
+
+
+def main():
+    parser = argparse.ArgumentParser(description='Generate Markdown + interactive HTML report')
+    parser.add_argument('--topic', required=True, help='Research topic')
+    parser.add_argument('--input', required=True, help='Path to JSON file with angle data')
+    parser.add_argument('--output-dir', required=True, help='Output directory')
+    args = parser.parse_args()
+
+    with open(args.input, 'r', encoding='utf-8') as f:
+        angles_data = json.load(f)
+
+    if isinstance(angles_data, dict):
+        angles_data = angles_data.get('angles', angles_data.get('results', []))
+
+    topic = args.topic
+    out_dir = Path(args.output_dir)
+    out_dir.mkdir(parents=True, exist_ok=True)
+
+    # Generate initial Markdown
+    md_content, all_sources = generate_markdown(topic, angles_data)
+
+    # Download images and insert into markdown
+    print('Downloading images...')
+    image_map = collect_and_download_images(angles_data, out_dir)
+    local_count = sum(1 for v in image_map.values() if not v.startswith('http'))
+    remote_count = sum(1 for v in image_map.values() if v.startswith('http'))
+    print(f'Images: {local_count} local, {remote_count} remote')
+
+    md_content = insert_images_into_markdown(md_content, angles_data, image_map, out_dir)
+
+    # Write README.md
+    md_path = out_dir / 'README.md'
+    md_path.write_text(md_content, encoding='utf-8')
+    print(f'Markdown saved to: {md_path}')
+
+    # Generate HTML with updated markdown (including image references)
+    html_content, all_sources = generate_html(topic, angles_data, md_content=md_content)
+
+    # Write index.html
+    html_path = out_dir / 'index.html'
+    html_path.write_text(html_content, encoding='utf-8')
+    print(f'HTML viewer saved to: {html_path}')
+
+    # Write sources.json
+    sources_data = {
+        'generated_at': datetime.now().isoformat(),
+        'sources': all_sources,
+    }
+    sources_path = out_dir / 'sources.json'
+    sources_path.write_text(json.dumps(sources_data, ensure_ascii=False, indent=2), encoding='utf-8')
+    print(f'Sources saved to: {sources_path}')
+
+
+def download_image(url, dest_path, timeout=10):
+    """Download an image from URL to local path. Returns (success, local_path_or_url)."""
+    try:
+        import urllib.request
+        import urllib.error
+        # Set a user agent to avoid 403 errors
+        req = urllib.request.Request(url, headers={'User-Agent': 'Mozilla/5.0'})
+        with urllib.request.urlopen(req, timeout=timeout) as response:
+            data = response.read()
+            content_type = response.headers.get('Content-Type', '')
+            # Determine extension from content-type or URL
+            ext = '.png'
+            if 'svg' in content_type or url.endswith('.svg'):
+                ext = '.svg'
+            elif 'jpeg' in content_type or 'jpg' in content_type or url.endswith(('.jpg', '.jpeg')):
+                ext = '.jpg'
+            elif 'gif' in content_type or url.endswith('.gif'):
+                ext = '.gif'
+            elif 'webp' in content_type or url.endswith('.webp'):
+                ext = '.webp'
+            elif url.endswith(('.png', '.svg', '.jpg', '.jpeg', '.gif', '.webp')):
+                ext = '.' + url.split('.')[-1].split('?')[0]
+
+            dest_path = dest_path.with_suffix(ext)
+            dest_path.parent.mkdir(parents=True, exist_ok=True)
+            dest_path.write_bytes(data)
+            return True, str(dest_path)
+    except Exception as e:
+        return False, url
+
+
+def collect_and_download_images(angles_data, out_dir):
+    """Collect all image URLs from angles and download them.
+    Returns (image_map, updated_md_insertions) where image_map is {url: local_path}."""
+    images_dir = out_dir / 'images'
+    images_dir.mkdir(exist_ok=True)
+
+    image_map = {}  # url -> local_path or original url
+    all_images = []  # list of (index, image_info)
+
+    for a_idx, a in enumerate(angles_data):
+        for img in a.get('images', []):
+            url = img.get('url', '')
+            if not url or url in image_map:
+                continue
+            all_images.append((a_idx, img))
+
+    # Download images
+    for idx, (a_idx, img) in enumerate(all_images):
+        url = img['url']
+        desc = img.get('description', f'Image {idx+1}')
+
+        # SVG files: use original URL (works in browsers and markdown viewers)
+        if url.endswith('.svg') or 'wikipedia.org/wiki/File:' in url:
+            image_map[url] = url
+            continue
+
+        # Generate a filename
+        safe_name = re.sub(r'[^\w]+', '-', desc.lower()).strip('-')[:50]
+        if not safe_name:
+            safe_name = f'image-{idx+1}'
+        dest_path = images_dir / safe_name
+
+        # Check if already downloaded (from a previous run)
+        existing = list(images_dir.glob(f'{safe_name}.*'))
+        if existing:
+            image_map[url] = str(existing[0].relative_to(out_dir)).replace('\\', '/')
+            continue
+
+        success, result = download_image(url, dest_path)
+        if success:
+            image_map[url] = result.replace('\\', '/')
+            print(f'  Downloaded: {desc[:40]} -> {result}')
+        else:
+            image_map[url] = url  # Fallback to original URL
+            print(f'  Failed to download (using original URL)')
+
+    return image_map
+
+
+def insert_images_into_markdown(md_content, angles_data, image_map, out_dir):
+    """Insert image references into markdown content at appropriate positions.
+    Returns updated markdown content."""
+    # Build a list of images per angle
+    angle_image_refs = {}  # angle_index -> list of markdown image refs
+    for a_idx, a in enumerate(angles_data):
+        for img in a.get('images', []):
+            url = img.get('url', '')
+            if not url:
+                continue
+            local_path = image_map.get(url, url)
+            desc = img.get('description', '')
+            source = img.get('source', url)
+            img_md = f'![{desc}]({local_path})\n\n*来源：[{source})]({source})*'
+            angle_image_refs.setdefault(a_idx, []).append(img_md)
+
+    # Insert images after the first paragraph of each angle section
+    # We'll find each angle's section and insert images after its first paragraph
+    lines = md_content.split('\n')
+    result = []
+    current_angle = -1
+    angle_header_pattern = re.compile(r'^##\s+(.+)')
+    first_para_done = set()
+
+    for i, line in enumerate(lines):
+        result.append(line)
+
+        # Detect angle headers
+        m = angle_header_pattern.match(line)
+        if m:
+            header_text = m.group(1)
+            for a_idx, a in enumerate(angles_data):
+                if a['angle'] == header_text:
+                    current_angle = a_idx
+                    break
+
+        # After first paragraph in an angle section, insert images
+        if current_angle >= 0 and current_angle not in angle_image_refs:
+            continue
+        if current_angle >= 0 and current_angle not in first_para_done:
+            # Check if this is a paragraph (non-empty, non-header, non-table, non-list)
+            if line and not line.startswith('#') and not line.startswith('|') and not line.startswith(('-', '*', '>', '```')) and not re.match(r'^\d+\.', line):
+                # Check if next line is empty (end of paragraph)
+                next_line = lines[i + 1] if i + 1 < len(lines) else ''
+                if next_line.strip() == '':
+                    # Insert images here
+                    result.append('')
+                    for img_ref in angle_image_refs.get(current_angle, []):
+                        result.append(img_ref)
+                        result.append('')
+                    first_para_done.add(current_angle)
+
+    return '\n'.join(result)
+
+
+if __name__ == '__main__':
+    main()