how to detect vector shapes in pdf pages

[darkworks]

Attach (recommended) or Link to PDF file here:
test4.pdf

Configuration:

• Web browser and its version: chrome latest
• Operating system and its version: win 7
• PDF.js version: latest from mozilla cdn
• Is a browser extension: no

i can detect text , images in pdf pages but do not know how to detect vector shapes , i checked wiki pages but no found any useful related to it

so any idea how to detect vectors in pdf that page have vectors or not , like the pdf i attached , page 1 have text , page 2 have vector shapes , and page 3 have image. so am stuck with page 2 vector shape detection that page have it or not .

examples

with page 1 to detect text am doing something like

      page.getTextContent({ normalizeWhitespace: true }).then(function (textContent) {
        textContent.items.forEach(function (textItem) {
	//	console.log(textItem.str);	  
        });
      });

with page 3 to detect images doing something like

				var ops = await page.getOperatorList();						
				for (let j=0; j < ops.fnArray.length; j++) {
			        if (ops.fnArray[j] == pdfjsLib.OPS.paintJpegXObject || ops.fnArray[j] == pdfjsLib.OPS.paintImageXObject) {
				   var op = ops.argsArray[j][0];
			            var img = page.objs.get(op);
}
}

page 2 .... problem

thanks

[timvandermeij]

Vector shapes are not one entity in PDF documents, but are instead composed of various operations, such as line drawing and filling (see https://github.com/mozilla/pdf.js/blob/master/src/shared/util.js#L174 for the list of existing operations), hence there is no easy way to extract vector shapes unfortunately.

[darkworks]

ok thanks , then how to just detect that page have vectors or not i tried something like

 if (ops.fnArray[j] == pdfjsLib.OPS.rectangle || ops.fnArray[j] == pdfjsLib.OPS.closePath) {
var op = ops.argsArray[j][0];

which not worked , then i just looped to see what inside
so i did something like

				var ops = await page.getOperatorList();						
				for (let j=0; j < ops.fnArray.length; j++) {
					console.log(ops.fnArray[j]);

which output

78
79

and then i checked : https://github.com/mozilla/pdf.js/blob/master/src/shared/util.js#L174

  beginAnnotations: 78,
  endAnnotations: 79,

as you can see my test pdf page 2 have one rectangle and circle so i was aspecting some rectange and circle codes

any idea what am doing wrong

[timvandermeij]

In that case the rectangle and circle are drawn onto the annotation layer, which are already SVG elements. Refer to https://github.com/mozilla/pdf.js/blob/master/src/display/annotation_layer.js#L898-L938 for how they're drawn.