Legacy ImageMagick Discussions Archive

My goal is to extract about 30 JPG or PNG images from a PDF using PHP. I'm using ImageMagick 6.9.4 from PHP 7.1.8 in a shared hosting account using this code to extract one specific image:

$imagick = new Imagick();
$imagick->readImage(bound.pdf[0]);
$imagick->setImageFormat("jpeg");
$imagick->writeImage("image1.jpg");

It works perfect for the first 10 or 15 images, but gives an exception "ReadPDFImage/664 Failed" during the readImage command if I try to read bound.pdf[18] for example. I have tried several source PDFs having from 20 to 30 images each. It always works fine for the first 10 images, but starting at index X and thereafter it always gets an exception.

I tried the same test on another hosting account with PHP 7.2.9 and ImageMagick 6.7.2-7 with similar results. It always gives an exception in the readImage command when index is around 15 or higher depending on the source PDF.

Any suggestions?

harold wrote:My goal is to extract about 30 JPG or PNG images from a PDF using PHP.

IM doesn't do that. IM will rasterize each page, no matter how many images each page has. I suspect you are trying to extract more pages than are in the document. Extracting embedded images from a PDF is a different task, which "pdfimages" does fairly well.

Perhaps my goal description was unclear. I have a 20 page PDF file and I want to extract out 20 individual JPG files (one for each page). Each page is essentially a big image. The code above works for the first X pages, then fails with an exception. I'm trying to extract pages 0 to 19.

I don't use PHP or IMagick so can't advise, but are you freeing memory correctly?

Just these two line of code alone will cause the exception so there is never a chance to free any memory. I tried changing the memory_limit assigned to PHP, but that doesn't seem to have an affect. It seems that when you are reading a page that is far enough into the PDF file, it gets the exception.

$imagick = new Imagick();
$imagick->readImage(bound.pdf[15]);

Are you sure that pdf file has 16 or more pages? The first page is [0] and you are requesting page [15] which would be the 16 page.

The Imagemagick developers have nothing to do with Imagick. Does it work with a simple convert command line?
If it does, then you will need to report the issue to the Imagick developers. See the Imagick forum.

You could try working with the first 10 pages and then the second 10 and see what happens " Each page is essentially a big image. " could be the problem as fmw42 suggested.

If it still fails try reading 5 pages at a time which will hopefully isolate the problem.

I have tested multiple PDF files each with over 20 pages. They all fail on the readImage[X] command. Bear in mind this is running in PHP from a shared hosting account so I do not have a Command line environment to run a convert command.

To clarify, the entire script consists of only two lines. It is not running a loop and finally getting the exception on the 16th page.

$imagick = new Imagick();
$imagick->readImage(bound.pdf[15]);

My "guess" and results of testing thus far seems to indicate that the implementation of the readImage[X] command exceeds some kind of memory limitation or calculation limitation starting at page X so the exception will occur at every index of X or above. The value of X depends on the specific PDF but appears consist for a specific PDF when running this tiny PHP script from different hosting environments.

It could be a limitation within PHP or Imagick and not ImageMagick.

Do you have access to PHP exec(), so that you could run the command line and see if that fails?