PdfOcrExtraction@1

Node PdfOcrExtraction@1 is used to extract text and data from PDF files using Optical Character Recognition (OCR). This node uses IronOCR to process PDF documents, including scanned documents, and can extract text, tables, and barcodes.

Adapter Prerequisites

• Mesh Adapter

Node Configuration

For fields path, targetPath, targetValueWriteMode, and targetValueKind, see Overview.

transformations:
  - type: PdfOcrExtraction@1
    path: $.PdfContent # The path to the base64-encoded PDF content
    targetPath: $.ExtractedText # The path where the extracted text will be stored
    targetValueWriteMode: Overwrite # The target value write mode
    targetValueKind: Simple # The target value kind
    pageNumbers: # Specific page numbers to process (optional, if not set all pages will be processed)
      - 1
      - 2
      - 5
    language: "en" # OCR language code (e.g., 'en', 'de', 'fr', default: "en")
    extractTables: false # Whether to extract tables from the PDF (default: false)
    tablesOutputPath: $.Tables # Output path for extracted tables (default: $.Tables)
    extractBarcodes: false # Whether to extract barcodes from the PDF (default: false)
    barcodesOutputPath: $.Barcodes # Output path for extracted barcodes (default: $.Barcodes)
    includeConfidence: false # Whether to include OCR confidence score in output (default: false)
    confidenceOutputPath: $.Confidence # Output path for OCR confidence score (default: $.Confidence)
    continueOnError: false # Whether to continue processing if OCR extraction fails (default: false)

Configuration Parameters

Parameter	Type	Required	Default	Description
pageNumbers	int[]	No	All pages	Specific page numbers to process
language	string	No	en	OCR language code
extractTables	bool	No	false	Whether to extract tables from the PDF
tablesOutputPath	string	No	$.Tables	Output path for extracted tables
extractBarcodes	bool	No	false	Whether to extract barcodes from the PDF
barcodesOutputPath	string	No	$.Barcodes	Output path for extracted barcodes
includeConfidence	bool	No	false	Whether to include OCR confidence score
confidenceOutputPath	string	No	$.Confidence	Output path for OCR confidence score
continueOnError	bool	No	false	Whether to continue if extraction fails

Supported Languages

The node supports the following language codes for OCR:

• en or english - English
• de or german - German
• fr or french - French
• es or spanish - Spanish
• it or italian - Italian
• pt or portuguese - Portuguese
• nl or dutch - Dutch
• ru or russian - Russian
• zh or chinese - Chinese (Simplified)
• ja or japanese - Japanese
• ko or korean - Korean
• ar or arabic - Arabic

Input Requirements

• The input must be a base64-encoded PDF file content
• Maximum file size: 1 MB (1,000,000 bytes)
• The PDF content must be available at the specified path

Output Structure

The node can produce multiple outputs based on configuration:

1. Main text output (targetPath): The extracted text from the PDF
2. Tables (tablesOutputPath): Structured table data if extractTables is enabled
3. Barcodes (barcodesOutputPath): Barcode data if extractBarcodes is enabled
4. Confidence score (confidenceOutputPath): OCR confidence percentage if includeConfidence is enabled

Error Handling

• If continueOnError is set to false (default), the pipeline will stop if OCR extraction fails
• If continueOnError is set to true, errors will be logged but the pipeline will continue
• Common errors include: missing PDF content, file size exceeded, invalid base64 encoding, or OCR processing failures

Example: Basic Text Extraction

triggers:
  - type: FromHttpRequest@1
    path: /extract/pdf
    method: POST
transformations:
  - type: PdfOcrExtraction@1
    path: $.pdfData
    targetPath: $.extractedContent
    language: "en"

Example: Extract Text with Tables and Confidence

transformations:
  - type: PdfOcrExtraction@1
    path: $.documentPdf
    targetPath: $.text
    language: "en"
    extractTables: true
    tablesOutputPath: $.tables
    includeConfidence: true
    confidenceOutputPath: $.ocrConfidence

Example: Process Specific Pages with Barcode Extraction

transformations:
  - type: PdfOcrExtraction@1
    path: $.pdfContent
    targetPath: $.extractedText
    pageNumbers:
      - 1
      - 3
      - 5
    language: "de"
    extractBarcodes: true
    barcodesOutputPath: $.barcodes
    continueOnError: true

Performance Considerations

note

OCR processing can be computationally intensive, especially for:

• Large PDF files
• Documents with many pages
• High-resolution scanned images
• Complex layouts with tables and graphics

Consider using pageNumbers to limit processing to specific pages when full document extraction is not required.