book-concept-atlas

KO | EN

gitlite — search

#kubernetes #hacktoberfest #docker #api #cli #agent #claude-code #privacy #reactjs #data-science

book-concept-atlas

(★ 9)

A Codex skill that turns books, PDFs, OCR dumps, and long notes into ultra-dense, print-ready SVG concept atlases.

파일 탐색기

ZIP 다운로드

- openai.yaml
- - china-modern-history-a5.png
  - financial-anticorruption-a4.png
- extraction_protocol.md
- layout_recipes.md
- smoke_sample.json
- clean_book_text.py
- render_concept_atlas.py
.gitattributes
.gitignore
LICENSE
README.md
SKILL.md

# CDN으로 사용하기

jsDelivr

jsDelivr는 공개 GitHub 리포지토리를 별도 설정 없이 CDN으로 즉시 서빙합니다. 버전과 파일을 고르면 웹페이지에 바로 붙일 수 있는 링크와 예시 코드가 만들어집니다.

버전

파일

링크

예시

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

面向 Codex 的技能：把整本书、PDF、OCR 文本或长笔记压缩成超高信息密度、可打印的 SVG 概念图谱，并通过空白感知布局与容量感知补丁渲染尽量写满页面。

A Codex skill for turning books, PDFs, OCR dumps, and long notes into ultra-dense, print-ready SVG concept atlases with whitespace-aware layout and capacity-aware filler rendering.

This repository contains a Codex skill plus its renderer and references. The skill is optimized for single-sheet or tightly packed multi-sheet concept atlases that avoid chapter-by-chapter summaries, large whitespace, and repetitive card layouts.

## Showcase

### 中国近代史纲要 / Modern Chinese History

![中国近代史纲要展示](https://raw.githubusercontent.com/SweeneyJun/book-concept-atlas/main/assets/showcase/china-modern-history-a5.png)

### 金融反腐论 / Financial Anti-Corruption

![金融反腐论展示](https://raw.githubusercontent.com/SweeneyJun/book-concept-atlas/main/assets/showcase/financial-anticorruption-a4.png)

## What It Does

- Extracts concept-first notes from long source material
- Reassembles them into mixed block families such as timelines, matrices, chains, compares, QA, glossary ribbons, and cross-links
- Renders dense SVG output with flexible-width panels, gap backfill, and whitespace-aware layout
- Supports capacity-aware filler variants such as `micro`, `compact`, `sentence`, and `ribbon`

## Repository Layout

- `SKILL.md`: Codex skill manifest and runtime instructions
- `agents/openai.yaml`: UI metadata for skill chips/prompts
- `scripts/clean_book_text.py`: OCR cleanup helper
- `scripts/render_concept_atlas.py`: canonical SVG renderer
- `references/extraction_protocol.md`: extraction and synthesis contract
- `references/layout_recipes.md`: layout heuristics
- `references/smoke_sample.json`: renderer smoke-test fixture

## Quick Start

Render the bundled smoke sample:

```bash
uv run python scripts/render_concept_atlas.py references/smoke_sample.json --out smoke.svg --paper a5 --columns 8 --font-size 8.2 --min-fill 0.85
```

Render your own atlas JSON:

```bash
uv run python scripts/render_concept_atlas.py atlas.json --out book_atlas.svg --paper a4 --columns 10 --font-size 7.2 --min-fill 0.90
```

If the source is OCR-heavy, clean the text first:

```bash
uv run python scripts/clean_book_text.py input.txt
```

## Skill Usage

In Codex, trigger the skill with:

```text
Use $book-concept-atlas to compress this full book into a dense non-chapter SVG concept atlas.
```

The full workflow, extraction contract, and rendering rules live in `SKILL.md` and the `references/` files.

## Development Notes

- The renderer is pure Python stdlib and intended to run via `uv run python`.
- Generated outputs such as `.svg`, `.html`, `.metrics.json`, `.png`, and `.pdf` are ignored by default.
- `SKILL.md` is the runtime-facing source of truth for Codex; `README.md` is only for GitHub readers.

book-concept-atlas

파일 탐색기

# CDN으로 사용하기

Was this content helpful?

Recent Feedback

Download README