@egintegrations/document-services

Document processing library with Google Cloud Vision OCR for text extraction from PDFs and images, plus text processing utilities.

Installation

npm install @egintegrations/document-services

Peer Dependencies

# For Google Cloud Vision OCR
npm install @google-cloud/vision

Features

• Google Cloud Vision OCR: Extract text from images and PDFs
• Text Processing: Normalize whitespace, clean OCR artifacts, extract dates/amounts
• TypeScript: Full type safety

Quick Start

OCR from Images/PDFs

import { GoogleVisionOCR } from '@egintegrations/document-services';

const ocr = new GoogleVisionOCR({
  credentials: {
    client_email: process.env.GCP_CLIENT_EMAIL,
    private_key: process.env.GCP_PRIVATE_KEY,
  },
  projectId: process.env.GCP_PROJECT_ID,
});

// Extract text from image
const result = await ocr.extractText({
  data: imageBuffer,
  mimeType: 'image/jpeg',
  filename: 'receipt.jpg',
});

if (result.success) {
  console.log(result.extractedText);
}

// Extract text from PDF
const pdfResult = await ocr.extractText({
  data: pdfBuffer,
  mimeType: 'application/pdf',
  filename: 'invoice.pdf',
});

Text Processing

import {
  cleanOCRText,
  extractAmounts,
  extractDates,
  extractLines,
} from '@egintegrations/document-services';

const rawText = 'Total: $ 100 . 00 Date: 01 / 15 / 2023';

// Clean OCR artifacts
const cleaned = cleanOCRText(rawText);
// "Total: $100.00 Date: 01/15/2023"

// Extract amounts
const amounts = extractAmounts(cleaned);
// [100.00]

// Extract dates
const dates = extractDates(cleaned);
// ['01/15/2023']

// Extract lines
const lines = extractLines(text);

API Reference

GoogleVisionOCR

interface OCRConfig {
  credentials?: {
    client_email: string;
    private_key: string;
  };
  projectId?: string;
}

class GoogleVisionOCR {
  constructor(config: OCRConfig);
  extractText(document: DocumentInput): Promise<OCRResult>;
}

Text Processing Functions

• normalizeWhitespace(text: string): string - Normalize whitespace
• cleanOCRText(text: string): string - Clean OCR artifacts
• extractLines(text: string): string[] - Extract non-empty lines
• extractAmounts(text: string): number[] - Extract currency amounts
• extractDates(text: string): string[] - Extract dates (MM/DD/YYYY)

Environment Variables

For Google Cloud Vision:

• GCP_CLIENT_EMAIL - Service account email
• GCP_PRIVATE_KEY - Service account private key
• GCP_PROJECT_ID - Google Cloud project ID

License

MIT

Credits

Extracted from BRS-Inbox-Scanner with Google Cloud Vision OCR integration.