Ghostscript PDF file compression using PHP's exec (Laravel on Docker)

Question

What needs to be done:

User has to be able to upload a PDF, then the file is uploaded to an Amazon S3 bucket, the file should be compressed then.

Current environment:

Laravel application (mounted on Docker) (php:7.4-fpm-alpine3.11, GPL Ghostscript 9.50, Laravel Framework 5.8.37)
Amazon S3 bucket to save documents in
Script is in a shell file which is made executable and added to /usr/local/bin as shrink
Shell is not explicitly added in Docker container, should it be?

Current flow:

User uploads file
File is uploaded to S3
Laravel then downloads the file to local temp folder
Ghostscript is then ran to compress said file (this script)
Compressed file is uploaded back to S3

Problem:

The file is found and being compressed, but the output is a blank (1 page, which is white) pdf file.

Dockerfile:

# Add compression shell script to global executables
COPY .docker/config/shrink.sh /usr/local/bin/shrink
RUN chmod u+x /usr/local/bin/shrink
RUN chown nobody.nobody /usr/local/bin/shrink

nobody is the user PHP is running on.

PHP function that should compress the file:

public function optimizeFile($file_path)
    {
        // Copy file from default disk to temp disk
        Storage::disk('temp')->put($file_path, Storage::get($file_path));

        $fullTempFilePath = Storage::disk('temp')->path($file_path);

        if (Storage::mimeType($file_path) == 'application/pdf') {
            $output = shell_exec("shrink " . $fullTempFilePath . " " . $fullTempFilePath);
            if ($output != null) {
                Log::error($output);
            }
        } else {
            ImageOptimizer::optimize($fullTempFilePath);
        }

        // Write the compressed file back to default disk
        Storage::put($file_path, Storage::disk('temp')->get($file_path));

        // Delete temp file
        Storage::disk('temp')->delete($file_path);
    }

If the file isn't a PDF, ImageOptimizer does it's job and compresses the image successfully.

What have I tried:

(Local) Successfully compressed the file without Docker, using php artisan serve to launch Laravel app;
(Local) Successfully compressed the file with Docker using docker exec -it <container_id> shrink in.pdf out.pdf;
(Local) Successfully compressed the file with Docker in it's shell using docker exec -it <container_id> /bin/bash;
(Local) My coworker did the same things locally and the response is also a blank pdf

KiprasT · Accepted Answer · 2020-04-02T09:24:43.087

2

Ok, so the problem is in here:

$output = shell_exec("shrink " . $fullTempFilePath . " " . $fullTempFilePath);

Ghostscript PDF compression does not work as expected if the input and output files are the same. Solution:

$output = shell_exec("shrink " . $fullTempFilePath  . $fullTempFilePath . "-compressed ");
shell_exec("mv " . $fullTempFilePath . "-compressed " . $fullTempFilePath);

edited Apr 02 '20 at 09:24

answered Apr 02 '20 at 08:41

KiprasT

370
5
14

score 0 · Answer 2 · answered Apr 02 '20 at 10:25

OK so first point; Ghostscript (more accurately the Ghostscript pdfwrite device) does not shrink PDF files. The actual process is described in the documentation in the overview in VectorDevices.htm. I'd reccomend you read it.

Secondly you can't use the same name for the input and output files. Ghostscript will still be reading from the input file while the pdfwrite device wants to write to the output file.