EN:Layout Export

English | 中文版

Layout Export

luatex-cn can export layout information to a JSON file after typesetting, including precise coordinates for every character, page info, column info, interlinear notes (jiazhu), and side notes. This feature is useful for regression testing, quality assurance, or integration with external systems.

The output format is aligned with guji-platform's PageLayout data structure, enabling direct downstream parsing without additional conversion.

Enabling Export

Add to your document preamble:

\enableLayoutExport

Or using the Chinese alias:

\开启排版导出

Export is disabled by default. When enabled, a <jobname>-layout.json file is generated automatically after compilation.

Custom Filename

\enableLayoutExport[filename=my-output.json]

Output File

After compilation, <jobname>-layout.json is generated in the same directory as the .tex file.

JSON Format

Top-Level Structure

{
  "version": "1.0",
  "generator": "luatex-cn",
  "document": { ... },
  "pages": [ ... ]
}

document Object

Global document information:

Field	Type	Description
total_pages	number	Total number of pages
page_width_pt	number	Page width (pt)
page_height_pt	number	Page height (pt)
grid_width_pt	number	Grid column width (pt)
grid_height_pt	number	Grid row height (pt)
line_limit	number	Maximum characters per column
columns_count	number	Total columns per page
split_page.enabled	boolean	Whether split page (tube page) is enabled

pages Array

Each page contains:

Field	Type	Description
page_index	number	Page number (0-indexed)
columns_count	number	Total columns on this page
margins	object	Page margins
columns	array	Array of columns with content
sidenotes	array	Array of side notes

margins Object

{
  "top_pt": 190.5,
  "bottom_pt": 91.6,
  "left_pt": 133.2,
  "right_pt": 133.2
}

All values in pt.

columns Array

Each column contains:

Field	Type	Description
col_index	number	Column index (0-indexed, right to left)
position	object	Column position {left_x, right_x} (pt)
characters	array	Array of characters in this column

characters Array

Each character contains:

Field	Type	Description
char	string	The character itself
unicode	number	Unicode code point
row_index	number	Row index (0-indexed, top to bottom)
position	object	Character position {x, y_top, y_bottom} (pt)
type	string	Character type: "normal" or "jiazhu"
confidence	number	Confidence score (always 1.0 for TeX)
jiazhu	object	Jiazhu info {sub_col: 1 or 2}, only present when type="jiazhu"
style	object	Style info {font_size_pt, font_color}, only present when different from default

position Object

{
  "x": 811.19,
  "y_top": 12.40,
  "y_bottom": 40.40
}

• x: Column center X coordinate (pt)
• y_top: Character top boundary Y coordinate (pt)
• y_bottom: Character bottom boundary Y coordinate (pt)

jiazhu Object

{
  "sub_col": 1
}

• sub_col: 1=right sub-column, 2=left sub-column

style Object

{
  "font_size_pt": 8.0,
  "font_color": "0.8 0 0"
}

sidenotes Array

Each side note contains:

Field	Type	Description
sidenote_id	number	Side note ID
anchor_col	number	Anchor column index
anchor_y_pt	number	Anchor Y coordinate (pt)
font_size_pt	number	Font size (pt)
spans_columns	boolean	Whether the note spans multiple columns
characters	array	Side note characters

Side note characters contain char, unicode, page, col, y_pt, and cell_height_pt fields.

Coordinate System

• Origin: Top-left corner of the page
• X axis: Positive to the right
• Y axis: Positive downward
• Column order: col_index=0 is the rightmost column, increasing leftward
• Row order: row_index=0 is the first character in a column (topmost)
• Unit: All coordinates and dimensions are in pt (1pt = 1/72.27 inch)

Example Output

{
  "version": "1.0",
  "generator": "luatex-cn",
  "document": {
    "total_pages": 1,
    "page_width_pt": 1136,
    "page_height_pt": 894.6,
    "grid_width_pt": 51.2096,
    "grid_height_pt": 28.0048,
    "line_limit": 21,
    "columns_count": 17,
    "split_page": { "enabled": false }
  },
  "pages": [
    {
      "page_index": 0,
      "columns_count": 17,
      "margins": {
        "top_pt": 190.5,
        "bottom_pt": 91.6,
        "left_pt": 133.2,
        "right_pt": 133.2
      },
      "columns": [
        {
          "col_index": 0,
          "position": { "left_x": 811.19, "right_x": 811.19 },
          "characters": [
            {
              "char": "史",
              "unicode": 21490,
              "row_index": 0,
              "position": { "x": 811.19, "y_top": 12.40, "y_bottom": 40.40 },
              "type": "normal",
              "confidence": 1.0
            }
          ]
        }
      ],
      "sidenotes": []
    }
  ]
}

Integration with guji-platform

The output format is aligned with guji-platform's PageLayout data structure:

Export Field	PageLayout Mapping	Notes
row_index	Character.row_index	Both 0-indexed
position.x/y_top/y_bottom	CharPosition	Direct mapping
type	Character.type	"normal" / "jiazhu"
confidence	Character.confidence	Always 1.0 for TeX
jiazhu.sub_col	JiazhuInfo.sub_col	Direct mapping
style	CharStyle	Direct mapping

Downstream from_tex_layout() can read the data directly without additional conversion.

Use Cases

• Regression Testing: Compare JSON output to detect unexpected layout changes
• Quality Assurance: Verify character grid alignment and jiazhu column balancing
• Data Extraction: Export typeset results as structured data for full-text search or digital humanities research
• External Integration: Provide precise character coordinates for OCR verification, layout analysis, etc.

---

Next: Features | Debug Mode | Command Reference