Skip to content

toc-generation-requirements.md

Latest commit

 

History

History
102 lines (76 loc) · 3.43 KB

toc-generation-requirements.md

File metadata and controls

102 lines (76 loc) · 3.43 KB
Raw

Table of Contents Generation Requirements

This document outlines the technical requirements and integration details for the Auto Author Table of Contents (TOC) generation functionality.

Overview

The TOC generation feature uses AI to automatically create a structured table of contents based on the user's book summary and responses to clarifying questions. This creates a foundation for the book's structure that can be further refined through the editing interface.

Requirements

Summary Input Requirements

  • Minimum word count: 100 words
  • Recommended word count: 250-500 words
  • Summary must contain sufficient thematic content to generate a meaningful structure
  • Confidence score is calculated based on:
    • Word count
    • Topic clarity
    • Genre identification
    • Narrative progression identifiers

TOC Structure Requirements

  • Each TOC must generate:
    • Top-level chapters (minimum 3, maximum 20)
    • Optional subchapters (nested up to 2 levels)
    • Brief description for each chapter/subchapter
    • Estimated page count
    • Structure notes with AI insights

AI Integration Requirements

Input Processing

  • Summary text is preprocessed to:
    • Remove irrelevant content
    • Identify key themes and topics
    • Extract genre and audience information
    • Analyze narrative or argumentative structure

Generation Process

  1. Readiness Assessment

    • AI analyzes summary for completeness
    • Calculates confidence score
    • Provides specific feedback if requirements aren't met

  2. Clarifying Questions

    • AI generates 3-5 targeted questions based on gaps in the summary
    • Questions focus on genre, audience, structure, and content depth
    • Responses are incorporated into the final TOC generation

  3. TOC Generation

    • AI constructs hierarchical chapter structure
    • Determines logical chapter sequence
    • Creates descriptive chapter titles
    • Generates brief content descriptions for each chapter
    • Adds structure notes with rationale for the organization

LLM Prompt Engineering

The system uses carefully engineered prompts to guide the AI in producing high-quality TOC structures:

  1. Readiness Assessment Prompt

    • Evaluates summary completeness
    • Identifies information gaps
    • Calculates confidence score
    • Suggests improvements

  2. Clarifying Questions Prompt

    • Identifies specific gaps in the summary
    • Generates questions to fill those gaps
    • Ensures questions cover genre, audience, structure, and content depth

  3. TOC Generation Prompt

    • Combines summary and question responses
    • Specifies desired TOC format and depth
    • Guides hierarchical structure creation
    • Sets constraints for chapter/subchapter count
    • Requests brief descriptions and structure notes

Error Handling

  • Graceful degradation if AI service fails
  • Retry mechanism for transient errors
  • User-friendly error messages
  • Fallback to manual TOC creation if automated generation fails
  • Timeout handling for long-running AI operations

Performance Considerations

  • AI requests are rate-limited (2 per 5 minutes per user)
  • Caching of TOC generation results
  • Asynchronous processing for long-running operations
  • Progress indicators during generation
  • Efficient storage of TOC structures in MongoDB

Related Documentation