Skip to content

fgilio/pdf-extractor

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

49 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

pdf-extractor

Pdf-extractor is a wrapper around pdf.js to generate images, svgs, html files, text files and json files from a pdf on node.js.

  • Image: A DOM Canvas is used to render and export the graphical layer of the pdf. Canvas exports *.png as a default but can be extended to export to other file types like *.jpg.
  • SVG: Pdf objects are converted to svg using the SVGGraphics parser of pdf.js.
  • HTML: Pdf text is converted to HTML. This can be used as a (transparent) layer over the image to enable text selection.
  • Text: Pdf text is extracted to a text file for different usages (e.g. indexing the text).

PDF.js on Node.js

This library is in it's most basic form a node.js wrapper for pdf.js. It has default renderers to generate a default output, but is easily extended to incorporate custom logic or to generate different output. It uses a node.js DOM and the node domstub from pdf.js do make pdf parsing available on node.js without a browser.

Box View / Crocodoc

This project is inspired by the Box View / Crocodoc way of converting documents (with this tool pdfs) to web-assets: images and html. The generated files match the files of Box View. This makes this library an option to transition from the Box View API to an open-source solution.

Examples

This library can be used as-is to generate assets from a pdf. The only requirements are a pdf as input and a writable directory as output. The extractor can also be used for rendering in different ways. The renderers can be extended or new ones can be injected into the extractor to render a pdf in new ways.

Extractor

How to use the default extractor to render png, html and text files for pdf pages:

const PdfExtractor = require('pdf-extractor').PdfExtractor;

let outputDir = '/path/to/output',

pdfExtractor = new PdfExtractor(outputDir, {
    viewportScale: (width, height) => {
        //dynamic zoom based on rendering a page to a fixed page size 
        if (width > height) {
            //landscape: 1100px wide
            return 1100 / width;
        }
        //portrait: 800px wide
        return 800 / width;
    },
    pageRange: [1,5],
});

pdfExtractor.parse('/path/to/dummy.pdf').then(function () {
	console.log('# End of Document');
}).catch(function (err) {
	console.error('Error: ' + err);
});

This results in these generated files:

info.json
page-1.png
page-2.png
page-3.png
page-4.png
page-5.png
stylesheet.css
text-1.html
text-1.txt
text-2.html
text-2.txt
text-3.html
text-3.txt
text-4.html
text-4.txt
text-5.html
text-5.txt

Using a custom renderer

It is relatively easy to extend or create new renderers/writers. Below an example to use the Canvas capability for *.jpg file generation.

const PdfExtractor = require('pdf-extractor').PdfExtractor;
const CanvasRenderer = require('pdf-extractor').CanvasRenderer;
const SvgRenderer = require('pdf-extractor').SvgRenderer;
const FileWriter = require('pdf-extractor').FileWriter;

class JPGWriter extends FileWriter
{
	getFilePathForPage(page) {
		return super.getPagePath(page.pageNumber, 'png');
	}

	writeCanvasPage(page, viewport, canvas) {
		return this.writeStreamToFile(canvas.jpgStream(), this.getFilePathForPage(page))
	}
}

class JPGCanvasRenderer extends CanvasRenderer
{
	getWriters(writerOptions) {
		let writers = super.getWriters(writerOptions);
		writers.push(new JPGWriter(this.outputDir, writerOptions));
		return writers;
	}
}

let outputDir = '/path/to/output',

pdfExtractor = new PdfExtractor(outputDir, {
	renderers: [
		new JPGCanvasRenderer(outputDir, rendererOptions),
		new SvgRenderer(outputDir, rendererOptions)
	]
});

pdfExtractor.parse('/path/to/dummy.pdf').then(function () {
	console.log('# End of Document');
}).catch(function (err) {
	console.error('Error: ' + err);
});

This adds jpg images to the generated files.

About

Node.js module for rendering pdf pages to images, svgs, html files, text files and json metadata

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • JavaScript 94.7%
  • HTML 5.3%