pdf2json

The parser supports interactive form elements like text inputs, radio buttons, checkboxes, link buttons, and drop-down lists. These elements are grouped within the corresponding Page object:

• Checkboxes and Radio Buttons: Located in the Boxsets array. A checkbox is an object in Boxsets where the boxes array contains exactly one element. A radio button group is an object in Boxsets where the boxes array contains multiple elements.
• Other Elements: Text inputs, drop-down lists, and link buttons are located in the Fields array.

Field Types Reference

// Example of a checkbox in Boxsets
Boxsets: [{
 boxes: [{ x: 47, y: 40, w: 3, h: 1, style: 48, TI: 39, AM: 4, id: { Id: "F8888" }, T: { Name: "box" } }],
 id: { Id: "A446" }
}]

// Example of a text input in Fields
{
 style: 48,
 T: { Name: "alpha", TypeInfo: { } },
 id: { Id: "p1_t40", EN: 0 },
 TU: "alternative text",
 TI: 0,
 x: 6.19,
 y: 5.15,
 w: 30.94,
 h: 0.85,
 V: "field value"
}

The pdf2json module is a Node.js port of pdf.js. Because pdf.js was originally designed for browser environments with HTML5 support, pdf2json implements several shims and modifications to allow it to run in Node.js:

• File Loading: Replaces XMLHttpRequest with Node's fs (File System) module to load PDF files.
• XML Parsing: Replaces the browser's DOMParser with the xmldom module.
• Threading: Uses a "fake worker" approach where parsing occurs in the same thread as the main process, rather than a background Web Worker.
• Canvas API: Uses a custom PDFCanvas implementation. Instead of drawing graphics to an HTML5 canvas, PDFCanvas writes 2D context API calls to a JavaScript object in JSON format.
• DOM: All DOM manipulation code from the original pdf.js is commented out to prevent errors in the non-browser environment.
• Fonts: Instead of downloading fonts, it parses font information into a CSS font format used in the output JSON's texts array.

When a widget field type is Tx and the additional-actions dictionary AA is set, the parser detects specific formatting types. These types are populated in the field's T object. Supported types include:

• Standard Types: number, ssn, date (including custom yyyy), zip, phone, and percent.
• Arbitrary Masks: If the format is mask, the mask string is provided in the MV field (e.g., 9999 for a 4-digit PIN).

Special Single-Character Masks

If an arbitrary mask has only one character, it represents:

• a: Alphabet only
• n: Numeric only (no locale formatting)
• d: Numeric only (with locale formatting, one decimal point allowed)
• -: Negative number only
• +: Positive number only

// Example of a 'number' field
{
 style: 48,
 T: { Name: "number", TypeInfo: { } },
 id: { Id: "FAGI", EN: 0 },
 TI: 0,
 x: 68.35,
 y: 22.43,
 w: 21.77,
 h: 1.08
}

You can use pdf2json as a CLI tool to transcode local PDF files or entire directories into JSON format. This is useful for pre-processing static PDFs to improve scalability by serving JSON directly to clients instead of parsing on a web server.

Using via Node.js

If you haven't installed the package globally, run it using node:

node pdf2json.js -f [input directory or pdf file]

Using as a global command

To run pdf2json directly without specifying the path to the script, install it globally:

npm install pdf2json -g

Then use:

pdf2json -f [input directory or pdf file]

Key Behaviors

• Files: If -f points to a .pdf file, it creates a .json file with the same name in the same directory.
• Directories: If -f points to a directory, it scans and processes all .pdf files within it (skipping dotfiles).
• Output Directory: Use -o [output directory] to specify where results should be saved. The directory is created automatically if it doesn't exist.

node pdf2json.js -f [input directory or pdf file] -o [output directory]