AI-generated concept cover: fictional cartoon children and an original AI helper discovering holographic comic panels in a storybook

AI-generated concept artwork · fictional original characters · Prompt and provenance

Comic Companion Course
从一本书到漫画伴读

Turn a story into a five-page comic reading workshop.

简体中文 · English

Free non-commercial v1.1 · 12 lessons · CC BY-NC 4.0: course text and code

Online course · Cases · Start · Workflow · Status · Contact

A static HTML course for medical researchers and AI learners. Lessons are in Chinese, with reusable prompts and runnable examples. All reader profiles and medical scenarios are fictional. Project creator, adaptation organiser and maintainer: **Haixiang Ding, MD**, emergency and trauma surgeon, Clinical Medicine PhD student at Soochow University. **AI disclosure:** the existing comic companions were organised with Hark assistance. Codex assisted with course text, teaching prompts, HTML/CSS/JS, SVG code and Python/JSON examples. The new concept cover was generated with the built-in OpenAI image generation tool; it is neither a book illustration nor an actual product screenshot. New prompts and code are teaching examples, not original production records. Original cast sheets, model/package details, image references and prompt logs remain **pending**. The current direct Codex content review is an AI-assisted review, not independent human, legal or medical expert sign-off. Earlier failed Hark reviews remain incomplete. ## Use cases | Goal | Start here | | --- | --- | | Set interest and reading constraints | Lesson 01: fictional reader, sentence length and vocabulary | | Adapt a small event | Lessons 03–06: panels, character states, image/text separation and layout | | Generate a modifiable layout example | Lesson 07 and examples directory | | Transfer the workflow to medical education | Lesson 10: fictional scenarios, sources and professional review | ## Two actual cases
Actual content page 1 of the AI-assisted Winnie-the-Pooh adaptation
Winnie-the-Pooh: 50 content pages / ten chapters		Actual content page 1 of the AI-assisted Peter Rabbit adaptation
Peter Rabbit: 20 content pages / four teaching units
These screenshots come from the creator-supplied companion PDFs, organised by Haixiang Ding, MD with Hark assistance. They are adapted pages, not original book scans or new display illustrations. Original text sources are Gutenberg #67098 and #14838; text and image rights are recorded separately. [PDF downloads and previews](../index.html#books). Public v1.1 PDFs add a separate disclosure strip below every original page: organiser, AI involvement, source and use boundaries. There remain 50/20 content pages; page height is slightly increased to accommodate the strip. Original files and v1.0 copies are unchanged, retained locally and excluded from this release. Both compressed previews are below 10MB; they reduce image detail, not known content defects. > Historical adaptation examples retain documented issues, including balloon colour consistency and subjective comprehension questions. Lessons 06/08 explain evidence, possible causes and proposed corrections; the book content itself has not been silently repaired or declared fully approved. ## Quick start Open the [online course](https://tuanziding.github.io/comic-companion-course/) or download the full repository and run: ```sh python3 -m http.server 8875 --bind 127.0.0.1 ``` Keep the server running and open `http://127.0.0.1:8875/` in Chrome. No build system, API account or key is needed. Local assets have no CDN, analytics or external font dependencies. A complete download works offline through the local server; external references and GitHub contact links need internet access. Run the five-page teaching example: ```sh python3 -m venv .venv .venv/bin/python -m pip install -r examples/示例依赖v1.0.txt .venv/bin/python examples/排版示例v1.1.py ``` Only reportlab is a direct dependency; its dependencies install automatically. Windows users should use the virtual environment's Scripts/python. The default input is examples/章节数据v1.1.json and output examples/输出/五页排版示例v1.1.pdf. Optional arguments: `--data PATH --output PATH`. Built-in Helvetica demonstrates English layout; Chinese notes remain in JSON/HTML. Illustration areas are explicitly marked placeholders. [Script](../examples/%E6%8E%92%E7%89%88%E7%A4%BA%E4%BE%8Bv1.1.py) · [Data](../examples/%E7%AB%A0%E8%8A%82%E6%95%B0%E6%8D%AEv1.1.json) · [Generated example](../examples/%E8%BE%93%E5%87%BA/%E4%BA%94%E9%A1%B5%E6%8E%92%E7%89%88%E7%A4%BA%E4%BE%8Bv1.1.pdf) ## Workflow Lawful source → fictional reader → event boundary and six panels → character states → text-free illustrations → shared utterance IDs and five-page layout → rendered review and corrections. Every lesson includes goals, explanation, an actual PDF case, a copyable teaching prompt, checklist and assignment. The twelve lessons cover introduction, readers, selection/rights, adaptation, cast continuity, image generation, layout, ReportLab, quality checks, family use, medical transfer and final submission. ## Capabilities and limits | Item | Current scope | | --- | --- | | Twelve lessons, themes, copy, gallery and progress | Static implementation; current direct content review and browser checks are documented | | Five-page layout code | Runs in a clean environment; reportlab only as direct dependency; artwork placeholders | | Audio | Read-along script design and suggested three-second gaps only; no generated audio or measured timeline | | Curriculum alignment and reading outcomes | Source lists and observation records pending; no outcomes claimed | | Rights | Four categories, separate source/regional boundaries; US public domain is not worldwide public domain | | Hark review | Historical failures remain incomplete, not acceptance evidence |
Maintainer checks scripts/本地检查v1.1.py needs html5lib, pypdf and Pillow. scripts/浏览器检查v1.1.js needs Playwright and local Chrome. These are QA tools, not website runtime dependencies. scripts/公开副本加注v1.0.py needs reportlab and pypdf. It reads local v1.0 copies and creates v1.1 disclosure copies without overwriting sources.
## Contribute and contact Use the verified [creator's GitHub profile](https://github.com/TUANZIDING) or [repository Issues](https://github.com/TUANZIDING/comic-companion-course/issues) for reproducible problems, suggestions and collaboration enquiries. Never post identifying child or patient data. **Future commercial collaboration may be discussed with the creator; commercial use requires a separate authorisation review.** This edition is free and non-commercial, with no payment flow or commercial licence promise. Hark has a separate written-authorisation condition for commercial outputs. Contacting the creator does not replace permissions from other rights holders. [Materials and rights](%E7%B4%A0%E6%9D%90%E4%B8%8E%E7%89%88%E6%9D%83%E8%AF%B4%E6%98%8Ev1.3.md) · [Content review and publication](%E5%86%85%E5%AE%B9%E5%AE%A1%E9%98%85%E4%B8%8E%E5%8F%91%E5%B8%83%E8%AE%B0%E5%BD%95v1.1.md) · [Checks and pending items](%E4%BA%A4%E4%BB%98%E6%A3%80%E6%9F%A5v1.1.md) · [Inventory](%E6%96%87%E4%BB%B6%E6%B8%85%E5%8D%95v1.2.md)

Haixiang Ding, MD · Course assisted by Codex / historical adaptations assisted by Hark
CC BY-NC 4.0 covers only newly written course text and code. Third-party content, original artwork, AI images, book PDFs and screenshots are excluded.

Only the 1926 Milne text / Shepard illustrations are discussed for Winnie-the-Pooh. The project is unrelated to Disney and must not use Disney designs or images as references. Mainland China and other non-US regional status remain pending review. Medical transfer uses fictional cases only, is not medical advice, and requires professional review and de-identification before real material is used.