nature-reader
Purpose
Transform a scientific paper (PDF path, DOI, or URL) into a richly annotated bilingual Markdown document: original English with inline Chinese translation, figure references grounded to actual captions, and section-level summaries.
Trigger Conditions
Activate when user mentions:
- "读论文" / "翻译论文" / "精读" / "全文翻译"
- "nature reader" / "paper reader" / "bilingual"
- "原文对照" / "图文对应" / "paper md"
- Shares a DOI, arXiv ID, PDF path, or paper URL
Input Handling
Accepted inputs:
- PDF file path — e.g.,
~/Downloads/paper.pdf→ Usereadtool to extract text content - DOI — e.g.,
10.1038/s41565-024-01234-5→ Fetch viahttps://doi.org/<DOI>orhttps://unpaywall.org/api/v2/<DOI>?email=open - arXiv ID — e.g.,
2401.12345→ Fetch viahttps://arxiv.org/abs/2401.12345 - URL — fetch directly with web_fetch
Open-access lookup:
If the paper is paywalled, try:
https://unpaywall.org/api/v2/<DOI>?email=open→ checkbest_oa_location.url_for_pdfhttps://sci-hub.se/<DOI>(mention only; do not auto-fetch)- arXiv preprint version via web_search:
arxiv "<title>" "<first author>"
Output Format
Generate a Markdown document with this structure:
# [Paper Title]
> **Journal:** Nature [Sub-journal] | **Year:** XXXX | **DOI:** [link]
> **Authors:** Author One, Author Two, ...
> **Open access:** [Yes/No] | **PDF:** [link if available]
---
## 📋 Quick Summary | 速览
| | |
|---|---|
| **核心问题** | [一句话:这篇论文解决了什么问题] |
| **核心方法** | [方法/技术核心] |
| **关键结果** | [最重要的1-2个数字/发现] |
| **意义** | [为什么重要] |
| **适合引用于** | [哪类论文的哪个部分可以引用这篇] |
---
## Abstract | 摘要
**[Original English abstract]**
> 🇨🇳 **中文翻译:**
> [Faithful Chinese translation of the abstract]
---
## Introduction | 引言
### [Subsection or paragraph grouping]
[Original English text — preserve key sentences verbatim]
> 🇨🇳 [Chinese translation of this paragraph]
**💡 Key point:** [One-sentence summary of this paragraph's main argument]
**📚 Key citations:** [[Author, Year]] — [why cited here]
[Continue paragraph by paragraph...]
---
## Results | 结果
### [Result subsection title]
[Original English — key sentences]
> 🇨🇳 [Chinese translation]
**📊 Figure X reference:** [Describe what Figure X shows and what conclusion it supports]
**🔢 Key numbers:** [Extract quantitative claims: "efficiency increased from X% to Y%"]
---
## Discussion | 讨论
[Original + Chinese + key point per paragraph]
---
## Methods | 方法
> ⚙️ [Methods summary in Chinese — full translation optional, summarize by subsection]
### [Methods subsection]
[Key parameters, instruments, conditions — bilingual]
---
## Figures | 图表解读
### Figure 1 | 图1
**Caption (original):** [Full original caption]
**中文说明:** [Chinese translation of caption]
**解读:** [What this figure proves. Which panel is most important and why.]
[Repeat for each figure]
---
## References | 参考文献
[List key references cited in the paper with brief annotations]
- [1] Author et al. (Year). *Title.* Journal. — [Why this ref matters]
---
## 🎯 How to Use This Paper | 如何使用这篇论文
**If you are writing about [topic]:**
- Cite in Introduction for: [specific claim it supports]
- Cite in Discussion for: [comparison point]
- Key sentence to reference: "[quote]"
**Limitations to note:**
- [Honest assessment of what this paper doesn't prove]
Translation Guidelines
- Translate faithfully, not literally — preserve scientific meaning
- Keep all numbers, chemical formulas, gene names, and proper nouns in original form
- For ambiguous terms, provide both: "催化活性(catalytic activity)"
- Technical terms: use standard Chinese scientific terminology
- e.g., "oxygen evolution reaction" → "析氧反应(OER)"
- "density functional theory" → "密度泛函理论(DFT)"
- Preserve hedging language: "suggest" → "表明", "indicate" → "指出", "demonstrate" → "证明"
Figure Grounding Rules
For each figure mentioned in the text:
- Find the corresponding figure caption
- Note which result/claim the figure supports
- Note the key quantitative message of each panel
- Flag any discrepancy between text claims and figure data
Output Options
Ask user:
- Full bilingual (每段都翻译) — default
- Summary only (只要速览+摘要+图表解读)
- Methods focus (重点翻译方法部分)
- Export — save as
~/Downloads/[paper-title]-reader.md
Scan to join WeChat group