javascript - Render .pdf to single Canvas using pdf.js and ImageData

Question

I am trying to read an entire .pdf Document using PDF.js and then render all the pages on a single canvas.

My idea: render each page onto a canvas and get the ImageData (context.getImageData()), clear the canvas do the next page. I store all the ImageDatas in an array and once all pages are in there I want to put all the ImageDatas from the array onto a single canvas.

var pdf = null;
PDFJS.disableWorker = true;
var pages = new Array();
    //Prepare some things
    var canvas = document.getElementById('cv');
    var context = canvas.getContext('2d');
    var scale = 1.5;
    PDFJS.getDocument(url).then(function getPdfHelloWorld(_pdf) {
        pdf = _pdf;
        //Render all the pages on a single canvas
        for(var i = 1; i <= pdf.numPages; i ++){
            pdf.getPage(i).then(function getPage(page){
                var viewport = page.getViewport(scale);
                canvas.width = viewport.width;
                canvas.height = viewport.height;
                page.render({canvasContext: context, viewport: viewport});
                pages[i-1] = context.getImageData(0, 0, canvas.width, canvas.height);
                context.clearRect(0, 0, canvas.width, canvas.height);
                p.Out("pre-rendered page " + i);
            });
        }

    //Now we have all 'dem Pages in "pages" and need to render 'em out
    canvas.height = 0;
    var start = 0;
    for(var i = 0; i < pages.length; i++){
        if(canvas.width < pages[i].width) canvas.width = pages[i].width;
        canvas.height = canvas.height + pages[i].height;
        context.putImageData(pages[i], 0, start);
        start += pages[i].height;
    }
    });

So from the way I understnad thing this should work, right? When I run this I end up with the canvas that is big enought to contain all the pages of the pdf but doesn't show the pdf...

Thank you for helping.

Answer 1

The PDF operations are asynchronous at all stages. This means you also need to catch the promise at the last render as well. If you not catch it you will only get a blank canvas as the rendering isn't finished before the loop continues to the next page.

Tip: I would also recommend that you use something else than getImageData as this will store uncompressed bitmap, for example the data-uri instead which is compressed data.

Here is a slightly different approach eliminating the for-loop and uses the promises better for this purpose:

LIVE FIDDLE

var canvas = document.createElement('canvas'), // single off-screen canvas
    ctx = canvas.getContext('2d'),             // to render to
    pages = [],
    currentPage = 1,
    url = 'path/to/document.pdf';              // specify a valid url

PDFJS.getDocument(url).then(iterate);   // load PDF document

/* To avoid too many levels, which easily happen when using chained promises,
   the function is separated and just referenced in the first promise callback
*/

function iterate(pdf) {

    // init parsing of first page
    if (currentPage <= pdf.numPages) getPage();

    // main entry point/function for loop
    function getPage() {

        // when promise is returned do as usual
        pdf.getPage(currentPage).then(function(page) {

            var scale = 1.5;
            var viewport = page.getViewport(scale);

            canvas.height = viewport.height;
            canvas.width = viewport.width;

            var renderContext = {
                canvasContext: ctx,
                viewport: viewport
            };

            // now, tap into the returned promise from render:
            page.render(renderContext).then(function() {

                // store compressed image data in array
                pages.push(canvas.toDataURL());

                if (currentPage < pdf.numPages) {
                    currentPage++;
                    getPage();        // get next page
                }
                else {
                    done();           // call done() when all pages are parsed
                }
            });
        });
    }

}

When you then need to retrieve a page you simply create an image element and set the data-uri as source:

function drawPage(index, callback) {
    var img = new Image;
    img.onload = function() {
        /* this will draw the image loaded onto canvas at position 0,0
           at the optional width and height of the canvas.
           'this' is current image loaded 
        */
        ctx.drawImage(this, 0, 0, ctx.canvas.width, ctx.canvas.height);
        callback();          // invoke callback when we're done
    }
    img.src = pages[index];  // start loading the data-uri as source
}

Due to the image loading it will be asynchronous in nature as well which is why we need the callback. If you don't want the asynchronous nature then you could also do this step (creating and setting the image element) in the render promise above storing image elements instead of data-uris.

Hope this helps!