Barcode reading issue with huge zoomed

[UbikZ]

Hi there,

I have notice some issue within the repository code (and the blog thus the blog documentation https://www.pdftron.com/blog/annotation/stamp-and-read-barcode-on-pdfs-with-pdftron-sdk). I have tried to find some solution but nothing relevant since I would like to handle this properly.

The issues is not really associated to the barcode reading, but the extract zone from pageCanvas.

1. First, I think the canvas multiplier is missing on x and y coordinates :

        const x = (annotation.X * zoom - leftOffset)  * canvasMultiplier;
        const y = (annotation.Y * zoom - topOffset)  * canvasMultiplier;
        const width = annotation.Width * zoom * canvasMultiplier;
        const height = annotation.Height * zoom * canvasMultiplier;

2. Secondly, the main issue is about the "pageCanvas" part. Regarding the document, a zoom threshold turns the pageCanvas from a "complete" render to a "partial" render. It's pretty obvious, it's to avoid getting a too much complexe canvas and to only keep what we see. But in our zone "extraction", we do not consider this behavior thus what we extract is wrong.

I have checked the API and I have found nothing to solve this issue.

I had the idea at the moment to use the DocViewer.getScrollViewElement() to get scrollLeft and scrollTop in order to substract this numbers from the x and y... But I have not found and event which could tell me that the canva is only one part of the initial document.

To reproduce it's pretty simple, just need to create a stamp which minimize in the bottom right of the document. We zoom at 500% for exemple and we try to read the barcode.
I hava added this code at the end in order to see draw and see the extracted zone.

            const pageCanvasCtx = pageCanvas.getContext('2d');
            pageCanvasCtx.strokeStyle = 'red';
            pageCanvasCtx.strokeRect(x, y, width, height);
            console.log(copyCanvas.toDataURL('image/png'));

Thank's for the help.

Gabriel