ASCII Preview Generate Skill
Purpose
This skill uses AI to create a meaningful ASCII text-based representation of the PDF page layout. The AI:
- Analyzes both the PDF page image and extracted text data
- Creates a structured ASCII preview showing page organization
- Represents heading hierarchy, paragraphs, lists, and tables visually
- Provides clear structural context for downstream HTML generation
The ASCII preview serves as a structured representation of page layout, enabling the downstream AI-HTML generator to understand both visual layout and content structure.
What to Do
-
Load input files
- Read
02_page_XX.png(visual reference) - Read
01_rich_extraction.json(text spans with metadata) - Verify all required fields present
- Read
-
Invoke AI to analyze and represent
- Send PNG image to Claude (visual reference)
- Include extracted text data (JSON)
- Request ASCII layout representation
- AI determines structure and relationships
-
AI generates ASCII preview
- AI analyzes heading hierarchy from extraction data
- AI identifies paragraphs and their relationships
- AI represents lists with proper formatting
- AI shows table layouts if present
- AI creates visual representation using ASCII characters
-
Validate ASCII output
- Verify output is readable text
- Check that all content elements represented
- Confirm structure is clear and logical
-
Save to persistent file
- Save to:
output/chapter_XX/page_artifacts/page_YY/03_page_XX_ascii.txt - Include metadata (page number, chapter, timestamp)
- Save to:
Input Parameters
rich_extraction_file: <str> - Path to 01_rich_extraction.json
page_png: <str> - Path to 02_page_XX.png (optional, for visual validation)
output_file: <str> - Path to save ASCII preview
book_page: <int> - Book page number (for display)
chapter: <int> - Chapter number (for display)
Output Format
ASCII Preview Structure
The ASCII preview uses this structure:
================================================================================
PAGE XX ASCII LAYOUT PREVIEW
================================================================================
Book Page: YY | Chapter: Z | PDF Index: XX | Dimensions: 612x792px
--------------------------------------------------------------------------------
[PAGE CONTENT WITH ASCII ART AND ANNOTATIONS]
STRUCTURE SUMMARY:
[Element counts and types]
================================================================================
Example Output
================================================================================
PAGE 16 ASCII LAYOUT PREVIEW
================================================================================
Book Page: 17 | Chapter: 2 | PDF Index: 16 | Dimensions: 612x792px
--------------------------------------------------------------------------------
┌─────────────────────────────────────────────────────────────────────────┐
│ │
│ [H1 Heading - 27pt Bold] │
│ ████████████████████████████████ │
│ █ Rights in Real Estate █ │
│ ████████████████████████████████ │
│ │
│ ───────────────────────────────────────────────────────────────── │
│ [Divider] │
│ │
│ [H2 Section Heading - 11pt Bold All-Caps] │
│ REAL PROPERTY RIGHTS │
│ │
│ [Paragraph 1 - 11pt Regular] │
│ Real property consists of physical land and the legal rights │
│ associated with ownership. These rights form what is commonly │
│ referred to as the "bundle of rights." │
│ │
│ [H4 Subsection - 11pt Bold] │
│ Physical characteristics. │
│ │
│ [Paragraph 2 - 11pt Regular] │
│ Land has unique physical characteristics that distinguish it from │
│ other types of assets. The most important characteristics include: │
│ │
│ [Bullet List - 3 items] │
│ ▶ Immobility - Land cannot be moved from one location to │
│ another. This fact has significant legal and economic │
│ implications. │
│ ▶ Indestructibility - Land is permanent and cannot be │
│ destroyed. Although structures on land can be destroyed, │
│ the land itself endures. │
│ ▶ Uniqueness - Each parcel of land is unique due to its │
│ location. "Not two parcels of land are exactly alike due to │
│ their geographic location." │
│ │
│ [H4 Subsection - 11pt Bold] │
│ Interdependence. │
│ │
│ [Paragraph 3 - 11pt Regular] │
│ All land exists in relation to other land. Therefore, the value │
│ of land is dependent on the land around it... │
│ │
└─────────────────────────────────────────────────────────────────────────┘
STRUCTURE SUMMARY:
─────────────────
Page Type: Chapter Continuation
Total Elements: 12
Headings (3):
• H1: 1 instance (27pt, bold) - "Rights in Real Estate"
• H2: 1 instance (11pt, bold, all-caps) - "REAL PROPERTY RIGHTS"
• H4: 2 instances (11pt, bold) - "Physical characteristics", "Interdependence"
Paragraphs (3):
• Body text: 3 paragraphs total
• Average paragraph length: ~45 words
• Text flow: Top to bottom, left to right
Lists (1):
• Bullet list with 3 items
• Indentation: 2em from left margin
• Separator: "▶" character
Tables: 0
Images: 0
Font Summary:
• 27pt: 1 instance (heading)
• 11pt: 7 instances (body, subsections, lists)
• Total unique fonts: 2 (Arial-Bold, Times-Regular)
Confidence Assessment:
• Heading hierarchy: 100% confident
• Element types: 95% confident
• Content flow: 90% confident
• (Lower confidence = ambiguous or unusual formatting)
================================================================================
AI Prompt Template
The prompt sent to Claude to generate ASCII preview:
You are creating an ASCII text-based layout preview of a PDF page.
VISUAL REFERENCE (PDF Page):
[PNG Image Attached]
EXTRACTED TEXT DATA:
[JSON Attached]
TASK:
Generate an ASCII text representation of this page's layout and structure.
REQUIREMENTS:
1. ASCII Format:
- Use box drawing characters for visual clarity (┌─┐│└┘)
- Show page boundaries
- Use spacing to represent actual layout
2. Element Representation:
- Headings: Show in ASCII, mark with [H1], [H2], [H3], [H4]
- Paragraphs: Show first line, indicate continuation
- Lists: Use ▶ or • characters, maintain structure
- Tables: