dunnewrite-scripts

Free, open-source scripts that prepare a book or a pile of notes for reading into DunneWrite, the story-bible app from Dunne, Corp. (DunneCorp).

DunneWrite reads plain text and Markdown. What you have is often something else: a book from Project Gutenberg wrapped in its licence, a PDF, an EPUB. These scripts turn it into clean text first, so the AI that reads it for DunneWrite reads only the story, and you pay only for the story.

Each script is one file, with no dependencies. Download the one you need and run it with Node.js 18 or newer. Nothing here connects to the internet, and no script ever changes the file you give it. It writes a new one.

Script	What it does	Status
clean-source.mjs	Removes Project Gutenberg's licence and boilerplate, keeping the book's title, author and language	available
PDF to text	Converts a PDF to clean prose (rejoins hyphenated words, drops page numbers and running heads)	planned
EPUB to text	Converts an EPUB to clean text, chapter by chapter	planned

clean-source.mjs

node clean-source.mjs pg1342.txt

  Project Gutenberg recognised.

  title   Pride and Prejudice
  author  Jane Austen
  kept    3 details of the book (title, author, language), at the top
  story   127,360 words
  removed 3,056 words of wrapper, about 4,953 tokens

  Dropped release date, most recently updated, other information and formats, credits.
  Project Gutenberg's licence lets you remove it only when every reference to Project Gutenberg goes
  with it (section 1.C), so the edition details were dropped and the book's own details were kept.

  Written: pg1342.clean.txt
  Your original file was not changed.

The cleaned file starts with the book's details and then the story, word for word:

---
title: "Pride and Prejudice"
author: "Jane Austen"
language: "English"
---

[the book, from its first page to its last]

Import pg1342.clean.txt in DunneWrite (Ingestion Review → Ingest a file). The author is filled in for you.

Why the Project Gutenberg details are removed too. Their licence lets you take the licence off a book only if every reference to Project Gutenberg goes with it (section 1.C). Otherwise the full licence has to stay with the file (section 1.E.4). So the script removes the release date, eBook number, link and volunteer credits as well, and leaves you with the plain public-domain book. If the story itself mentions Project Gutenberg, usually in a transcriber's note, the script tells you and leaves that line for you to delete. The licence is written for the United States. If you live elsewhere, check that the book is out of copyright where you are too. None of this is legal advice.

Project Gutenberg is run by volunteers. If their work made your book possible, consider supporting them.

Options:

node clean-source.mjs <book.txt> [--out <cleaned.txt>] [--source gutenberg]
node clean-source.mjs --help
node clean-source.mjs --selftest

Testing

npm test

runs every script's self-test. There is nothing to install.

Contributing and security

Bug reports, a book a script gets wrong, and pull requests (including new sources and formats) are welcome: see CONTRIBUTING.md. Please report security problems privately, as SECURITY.md describes, not in a public issue.

Licence

MIT — see LICENSE.

DunneWrite and DunneCorp are trademarks of Dunne, Corp. The licence covers this repository's contents. It does not grant the right to use those names or logos, except to say truthfully that software prepares files for DunneWrite. Project Gutenberg is a registered trademark of the Project Gutenberg Literary Archive Foundation. This project is not affiliated with or endorsed by it.