Skip to content

yueying526/paper-visual-explainer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Paper Visual Explainer 📄✨

Turn academic papers into beautiful, interactive visual reports that anyone can understand.

将学术论文转化为小白也能看懂的互动可视化解读报告。

Demo — Attention Is All You Need

Example: Full interactive report for "Attention Is All You Need" (download and open in browser)

Why This Tool?

Reading academic papers is hard. This tool transforms dense research papers into engaging, visual reports with:

  • 🎯 Plain language summaries — what problem, what method, what limitations, in one paragraph
  • 📖 Term + Analogy cards — every key term gets a plain explanation, a life analogy, and a real case (expandable)
  • 🏭 Industry applications — specific industries, specific use cases, no fluff
  • 📊 Data-driven insights — quantified impact with real numbers
  • 🎨 AI-generated illustrations — via Nano Banana API
  • 📱 Interactive HTML report — reading progress bar, nav highlighting, scroll animations, number counters
  • 📄 PDF export — print-ready, full Chinese support

Perfect for: tech bloggers, WeChat/公众号 writers, students, and anyone who wants to explain research to a broader audience.

Quick Start

# Install dependencies
pip install requests PyPDF2 pdfplumber reportlab

# One-click: download → extract → generate illustrations → build report
python3 scripts/generate_all.py 1706.03762

# Fast mode (simpler illustrations, ~40% faster)
python3 scripts/generate_all.py 1706.03762 --fast

Or run each step individually:

python3 scripts/download_paper.py "1706.03762" -o /tmp/papers/
python3 scripts/extract_pdf.py paper.pdf -o extracted --structure
python3 scripts/call_nano_banana.py "your prompt" -o img.png --fast
python3 scripts/generate_html_report.py meta.json content.json -o report.html

Writing Style: Kazike (卡兹克风格)

Reports use an engaging Chinese science communication style:

  • Extended metaphors running through the entire article (e.g., comparing a neural network to a "101-story company building")
  • Conversational tone — "对吧", "咋办呢", "牛逼坏了"
  • Shocking data comparisons — "3000:1.6 — a 3 order of magnitude reduction"
  • Progressive narrative — history → problem → solution → impact

Use as Claude Code Skill

This tool is also available as a Claude Code skill. Once installed, just say:

"帮我解读这篇论文: [arXiv link]"

It will automatically download, analyze, and generate a visual report.

Project Structure

paper-visual-explainer/
├── scripts/
│   ├── generate_html_report.py   # Interactive HTML report generator
│   ├── generate_pdf_report.py    # PDF report generator  
│   ├── download_paper.py         # arXiv paper downloader
│   ├── extract_pdf.py            # PDF content extractor
│   └── call_nano_banana.py       # AI illustration generator
├── references/
│   ├── kazike_style_guide.md     # Writing style guide
│   └── nano_banana_api.md        # Illustration API docs
└── SKILL.md                      # Claude Code skill config

Contributing

PRs welcome! Especially for:

  • New writing style templates
  • Additional report themes
  • Support for more paper sources beyond arXiv
  • Multilingual report generation (English output)

Examples

Attention Is All You Need (Vaswani et al., 2017) — Download HTML report

More examples coming soon.

License

MIT


Author: Yueying Wu · GitHub · X

About

Turn academic papers into beautiful interactive visual reports anyone can understand | 论文可视化解读工具

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages