Skip to content

QWEN.md

Latest commit

 

History

History
110 lines (89 loc) · 4.97 KB

QWEN.md

File metadata and controls

110 lines (89 loc) · 4.97 KB

Book Writer System - QWEN.md

Project Overview

The Book Writer System is a comprehensive book development application that helps authors create, organize, and refine book content through a structured workflow. The system combines traditional book writing processes with AI-powered content generation using local Ollama models.

Key Features

  • Outline Creation: Create hierarchical book outlines in YAML/JSON format with parts, chapters, and subtopics
  • Note Processing Pipeline: Embed notes using BGE-M3 model and store in ChromaDB with metadata
  • AI-Powered Content Expansion: Expand raw notes into polished text using AI models
  • Retrieval-Augmented Writing (RAG): Context-aware writing with Ollama models
  • AI Function Calling & Tool Integration: Support for AI models to call specific tools and functions
  • Content Organization: Cluster content, detect gaps, suggest reorganization, and visualize content structure
  • Book Assembly: Merge chapters according to outline and generate complete manuscripts
  • Multiple Export Options: Export to PDF and ePub formats
  • Dual Interface: Interactive CLI and web-based Gradio UI

Architecture

The system follows a modular architecture with clear separation of concerns:

  • app.py: Main application interface that ties all components together
  • outline.py: Handles book outline creation and management
  • note_processor.py: Manages note processing and storage in ChromaDB
  • content_expansion.py: Expands notes into polished content using AI models
  • rag_writing.py: Implements RAG loop for context-aware writing
  • book_assembly.py: Combines content into complete manuscripts
  • export.py: Handles PDF and ePub export functionality
  • model_manager.py: Manages different Ollama models for various tasks
  • organized_content.py: Provides tools for content organization, including the ContentOrganizer and ContentOrganization classes.
  • ui.py: Gradio web interface implementation
  • tool_registry.py: Manages AI tool integration for function calling

Building and Running

Prerequisites

  1. Install Ollama on your system
  2. Pull the required models: 
    ollama pull stable-beluga:13b
ollama pull deepseek-r1:8b

Installation

pip install -r requirements.txt

Running the Application

Interactive CLI Mode

python main.py

Web-based Gradio UI

python main.py --ui

Configuration

Model configuration can be customized in book_writer/model_config.yaml:

  • Content Expansion & RAG: stable-beluga:13b (optimized for generating rich, detailed content)
  • Outline Generation & Organization: deepseek-r1:8b (optimized for structure and organization)

Development Conventions

AI Model Usage

  • Stable Beluga 13B: Used for content generation tasks (expanding notes into polished text, RAG generation)
  • DeepSeek R1 8B: Used for organizational tasks (outline generation, content classification)

Tool Integration

The system includes a flexible tool registry that allows AI models to call specific functions:

  • search_notes: Search for notes based on a query
  • classify_content: Classify content into appropriate chapter and subtopic
  • expand_note: Expand a note into detailed content
  • validate_outline: Validate if an outline is properly structured

Data Storage

  • Notes and content are stored with embeddings in ChromaDB for semantic search
  • Outlines are stored in YAML/JSON format in the outlines/ directory
  • Content is stored in markdown format in the content/ directory
  • Project configuration is stored in book_config.yaml

Streaming Implementation

The system implements real-time streaming of Ollama output to the web UI. The expand_note_stream_with_real_time function in the UI progressively yields content as it's received from Ollama, allowing users to see responses as they are generated.

Usage Examples

Interactive CLI Commands

  • create_outline <title> [author] [description]: Create a new book outline
  • add_note <text> [source] [topic1,topic2,...]: Add a new note to the system
  • expand_note <note_id> [style]: Expand a note into polished text
  • organize_content <content_id1> <content_id2> ...: Organize content and get a summary
  • get_suggestions: Get suggestions for content gaps and reorganization
  • visualize_content <output_path>: Generate a visualization of the content structure
  • build_book: Build the current book according to the outline
  • export_book <format>: Export the book to PDF or ePub format

Web Interface Features

  • Project creation and loading
  • Note adding with real-time AI expansion
  • Writing style selection (academic, narrative, technical, conversational)
  • Content editing capabilities
  • Simplified content organization with a step-by-step process:
    • Organize content and get a summary
    • Review organization suggestions
    • Visualize content structure
  • Outline visualization and management
  • Book building and export functionality