use qpdf to linearize a bunch of pdf files

pdf-linearize.php

<?php
/**
 * pdf-linearize.php - linearize PDFs
 * WARNING: May run extremely slowly due to very large file sizes generated on pdf2ps
 *
 * relies on command-line tools, tested on ubuntu.
 *
 * @license GPLv3 (http://www.gnu.org/licenses/gpl-3.0.html)
 */

$folder = '.';

define('DEBUG',1);
define('VERBOSE',1);

date_default_timezone_set('UTC');
ini_set('default_charset', 'utf-8');
ini_set('mbstring.encoding_translation', 'On');
ini_set('mbstring.func_overload', 6);
ini_set('auto_detect_line_endings', TRUE);

// check CLI commands are available
$commands = getCommands();
if (empty($commands)) exit;

$ext = '.pdf';
$cmd = 'find '.$folder.' -type f -iname "*'.$ext.'" -print';
$files = cmd_execute($cmd);
$total = count($files);
$start = 99999;
foreach ($files as $k => $path) {
    $total--;
    unset($files[$k]);
    $path = trim($path);
    if (!file_exists($path)) {
        continue;
    }
    if (0 == filesize($path)) {
        unlink($path);
        continue;
    }
    if ($total > $start) {
        continue;
    }

    output("Checking [".count($files)."]: $path\n");

    // get metadata and linearize if not
    $cmd = $commands['exiftool'] . " -php -G " . escapeshellarg($path);
    debug($cmd);
    try {
        $output = cmd_execute($cmd);
    } catch (Exception $e) {
        debug("Failed getting file data with exiftool.");
        //debug(print_r($e,1));
    }
    eval('$metadata = ' . join("\n", $output));
    if (!empty($metadata))
        $metadata = $metadata[0];
    if (array_key_exists('PDF:Linearized', $metadata)) {
        $linearized = $metadata['PDF:Linearized'];
        if ('Yes' == $linearized) {
            verbose("Already linearized!");
            continue;
        }
    }
    var_dump($metadata);

    // linearize pdf
    $bak = $path . '.bak';
    rename($path, $bak);
    verbose("Linearizing...");
    $cmd = sprintf("qpdf --linearize %s %s", escapeshellarg($bak), escapeshellarg($path));
    debug($cmd);
    try {
        $output = cmd_execute($cmd);
    } catch (Exception $e) {
        rename($bak, $path);
        debug(print_r($e,1));
        continue;
    }
    debug($output);

    if (filesize($path) > filesize($bak)) {
        //cmd_execute(sprintf("%s %s %s", $commands['mv'], escapeshellarg($bak), escapeshellarg($path)));
        //verbose("\tFilesize is not smaller!\n");
    } else {
        $savedb =  filesize($bak) - filesize($path);
        output("\tShrinking, saved: ". ceil($savedb/1024) . " Kbytes\n");
        //unlink($bak);
    }
    unlink($bak);
}
exit;

//-----------------------------------------------------------------------------
// functions used above

/**
 * Execute a command and return streams as an array of
 * stdin, stdout, stderr
 *
 * @param  string $cmd command to execute
 * @return array|false array $streams | boolean false if failure
 * @see    https://secure.php.net/manual/en/function.proc-open.php
 */
function shell_execute($cmd)
{
    $process = proc_open(
        $cmd,
        [
        ['pipe', 'r'],
        ['pipe', 'w'],
        ['pipe', 'w']
        ], $pipes
    );
    if (is_resource($process)) {
        $streams = [];
        foreach ($pipes as $p => $v) {
            $streams[] = stream_get_contents($pipes[$p]);
        }
        proc_close($process);
        return [
            'stdin'  => $streams[0],
            'stdout' => $streams[1],
            'stderr' => $streams[2]
        ];
    }
    return false;
}
/**
 * Execute a command and return output of stdout or throw exception of stderr
 *
 * @param  string $cmd command to execute
 * @param  boolean $split split returned results? default on newline
 * @param  string $exp regular expression to preg_split to split on
 * @return mixed string $stdout | Exception if failure
 * @see    shell_execute($cmd)
 */
function cmd_execute($cmd, $split = true, $exp = "/\n/")
{
    $result = shell_execute($cmd);
    if (!empty($result['stderr'])) {
        throw new Exception($result['stderr']);
    }
    $data = $result['stdout'];
    if (empty($split) || empty($exp) || empty($data)) {
        return $data;
    }
    return preg_split($exp, $data);
}

// check required commands installed and get path
function getCommands()
{
  static $commands = []; // cli command paths
  if (!empty($commands)) {
    return $commands;
  }

  $requirements = [
    'qpdf' => 'http://qpdf.sourceforge.net/',
    'exiftool' => 'https://sno.phy.queensu.ca/~phil/exiftool/',
    'cp' => 'copy system command - cp',
    'mv' => 'move system command - mv',
    'find' => 'system find commmand',
  ];

  $errors = [];
  foreach ($requirements as $tool => $description) {
    $cmd = cmd_execute("which $tool");
    if (empty($cmd)) {
      $errors[] = "Error: Missing requirement: $tool - " . $description;
    } else {
      $commands[$tool] = $cmd[0];
    }
  }

  if (!empty($errors)) {
    echo join("\n", $errors) . "\n";
  }

  return $commands;
}

/**
 * Return the memory used by the script, (current/peak)
 *
 * @return string memory used
 */
function get_memory_used()
{
    return(
        ceil(memory_get_usage() / 1024 / 1024) . '/' .
        ceil(memory_get_peak_usage() / 1024 / 1024));
}

/**
 * Output string, to STDERR if available
 *
 * @param  string { string to output
 * @param  boolean $STDERR write to stderr if it is available
 */
function output($text, $STDERR = true)
{
    if (!empty($STDERR) && defined('STDERR')) {
        fwrite(STDERR, $text);
    } else {
        echo $text;
    }
}

/**
 * Dump debug data if DEBUG constant is set
 *
 * @param  optional string $string string to output
 * @param  optional mixed $data to dump
 * @return boolean true if string output, false if not
 */
function debug($string = '', $data = [])
{
    if (DEBUG) {
        output(trim('[D ' . get_memory_used() . '] ' . $string) . "\n");
        if (!empty($data)) {
            output(print_r($data, 1));
        }
        return true;
    }
    return false;
}

/**
 * Output string if VERBOSE constant is set
 *
 * @param  string $string string to output
 * @param  optional mixed $data to dump
 * @return boolean true if string output, false if not
 */
function verbose($string, $data = [])
{
    if (VERBOSE && !empty($string)) {
        output(trim('[V' . ((DEBUG) ? ' ' . get_memory_used() : '') . '] ' . $string) . "\n");
        if (!empty($data)) {
            output(print_r($data, 1));
        }
        return true;
    }
    return false;
}