mono/packages/content/ref/pdf-to-images

@polymech/pdf

Installation

1. Clone the repository (optional):

   git clone <repository-url> # Replace with your repository URL
   cd <directory-name> # e.g., cd pdf
   

2. Install dependencies:

   npm install
   

3. Build the project:

   npm run build
   

CLI Usage

When running from within the cloned project directory after building:

npm start -- convert [options]
# or directly execute the built script
node dist/index.js convert [options]

(Note: If you publish this package and install it elsewhere, you might execute it differently, potentially using npx @polymech/pdf convert ... if a bin entry is added to package.json)

Available command: convert - Convert PDF to images

Options for convert command:

• -i, --input <string>: Input PDF file (required)
• -o, --output <string>: Output directory prefix for images (required)
• --dpi <number>: DPI for output images (default: 300)
• --format <string>: Output image format (choices: 'png', 'jpg', default: 'png')
• -s, --startPage <number>: First page to convert (1-based)
• -e, --endPage <number>: Last page to convert (1-based, inclusive)

Example:

node dist/index.js convert -i mydocument.pdf -o output/image

This will generate images like output/image_1.png, output/image_2.png, etc.

Another example (using JPG format and 150 DPI):

node dist/index.js convert -i report.pdf -o images/report_page --format jpg --dpi 150

This generates images/report_page_1.jpg, images/report_page_2.jpg, etc.

Example specifying a page range (pages 3 to 5):

node dist/index.js convert -i long_doc.pdf -o pages/doc_pg --startPage 3 --endPage 5

This generates pages/doc_pg_3.png, pages/doc_pg_4.png, pages/doc_pg_5.png.

API Usage

import { convertPdfToImages, ImageFormat, PdfToImageOptions } from './dist/lib/pdf'; // Adjust path based on your project structure
import { readFile } from 'node:fs/promises';

async function example() {
  try {
    const pdfBuffer = await readFile('mydocument.pdf');

    const options: PdfToImageOptions = {
      outputPathPrefix: 'output/image',
      dpi: 300,
      format: 'png'
    };

    const outputFilePaths = await convertPdfToImages(pdfBuffer, options);
    console.log('Generated images:', outputFilePaths);
  } catch (error) {
    console.error('Error:', error);
  }
}

example();

Example using JPG format:

import { convertPdfToImages, PdfToImageOptions } from './dist/lib/pdf'; // Adjust path
import { readFile } from 'node:fs/promises';
import { Logger } from 'tslog'; // Assuming you want logging

async function exampleJpg() {
  const logger = new Logger();
  try {
    const pdfBuffer = await readFile('report.pdf');
    const options: PdfToImageOptions = {
      outputPathPrefix: 'images/report_page',
      dpi: 150,
      format: 'jpg',
    };
    const outputFilePaths = await convertPdfToImages(pdfBuffer, options);
    logger.info('Generated JPG images:', outputFilePaths);
  } catch (error) {
    logger.error('Error generating JPGs:', error);
  }
}

exampleJpg();

Example with specific page range:

import { convertPdfToImages, PdfToImageOptions } from './dist/lib/pdf'; // Adjust path
import { readFile } from 'node:fs/promises';

async function examplePageRange() {
  try {
    const pdfBuffer = await readFile('long_doc.pdf');
    const options: PdfToImageOptions = {
      outputPathPrefix: 'pages/doc_pg',
      dpi: 200,
      format: 'png',
      startPage: 3,
      endPage: 5
    };
    const outputFilePaths = await convertPdfToImages(pdfBuffer, options);
    console.log('Generated specific pages:', outputFilePaths);
  } catch (error) {
    console.error('Error generating page range:', error);
  }
}

examplePageRange();

Exports

• convertPdfToImages(pdfData: Buffer, options: PdfToImageOptions): Promise<string[]>: Converts a PDF buffer to images.
• ImageFormat: Type alias for 'png' | 'jpg'.
• PdfToImageOptions: Interface for conversion options (outputPathPrefix, dpi, format, optional startPage, optional endPage, optional logger).

References

• https://github.com/opendatalab/PDF-Extract-Kit/tree/main/project/pdf2markdown (Using Yolo)