📘 Automated Book Publication Workflow

An intelligent agent-driven pipeline to automate the rewriting and publication of public-domain books with human-in-the-loop editing and content versioning.

---

✨ Features

• 🌐 Web Scraping
  Scrapes book chapters from online sources (e.g., Wikisource) using Playwright.

• 🧠 AI Content Generation
  Uses a multi-agent architecture (Writer & Reviewer) powered by LLMs (like Google Gemini) to rewrite and refine chapters.

• 🧑‍💻 Human-in-the-Loop Editing
  Optional manual feedback after AI rewriting allows human writers/editors to make changes before finalizing.

• 📚 Content Versioning
  Final outputs are saved into ChromaDB with automatic version labels (v1, v2, ...), UUIDs, and metadata.

• 🔍 RL-inspired Retrieval
  Query past versions using TF-IDF + cosine similarity (stubbed for future reinforcement learning-based retrieval).

---

📂 Directory Structure

auto_book_pub/
├── main.py                    # Entry point
├── README.md
├── LICENSE
├── requirements.txt
├── .env
├── .gitignore
├── config/
│   └── settings.yaml
├── data/
│   ├── raw/                   # Raw scraped HTML + screenshots
│   ├── processed/             # AI-edited versions
│   └── versions/              # Finalized, versioned outputs
├── scraping/
│   └── scraper.py             # Playwright-based scraper
├── ai_agents/
│   ├── writer_agent.py        # AI "spinner"
│   ├── reviewer_agent.py      # Reviewer LLM
│   └── editor_agent.py        # Optional human/AI edit flow
├── human_loop/
│   └── feedback_manager.py    # Handle user input iterations
├── versioning/
│   └── chromadb_manager.py    # Store/retrieve versioned content
├── rl_search/
│   └── retriever.py           # Reinforcement Learning-based retriever
└── utils/
    └── helpers.py             # Common utilities

🚀 How It Works

1. Scrape Content

   raw_text = scrape_chapter(target_url)

2. AI Rewriting + Review

rewritten = WriterAgent().spin(raw_text)
reviewed  = ReviewerAgent().review(rewritten)

3. Human Feedback (Optional)

final_version = collect_feedback(reviewed)

4. Save to ChromaDB

save_version(final_version)

5. Retrieve Similar Versions

results = retrieve_version("version-number")

⚙️ Setup

🔧 Prerequisites

-Python 3.10+

-Google Gemini API Key (optional)

-Playwright dependencies

📦 Installation

git clone https://github.com/yourusername/auto_book_pub
cd auto_book_pub

Set up Python environment

pip install -r requirements.txt

Set up Playwright

playwright install

🔐 .env Configuration

Create a .env file:

GEMINI_API_KEY=your_google_gemini_key

🧪 Dev Mode (Optional)

To avoid re-scraping each time, enable dev mode in main.py:

dev_mode = True
if dev_mode:
    raw_text = load_cached_text()
else:
    raw_text = scrape_chapter(url)

📄 License

MIT License. See LICENSE for details.