OCRs an image and adds the extracted text as Spotlight comment

ocrcomment

#!/usr/bin/env php
<?php
/**
 * ocrcomment INPUT
 * ================
 *
 * > Takes an image, performs OCR and stores the text as Spotlight comment
 * > BAM! You can search for your images using Spotlight.
 *
 * By Florian Eckerstorfer (https://florian.ec, @Florian_)
 *
 * Installation
 * ============
 * This script uses Tesseract to perform the OCR, you can use Homebrew to install it:
 *   $ brew install tesseract
 *
 * Add execution permissions and copy this script in a directory that is in your $PATH
 *   $ chmod +x ocrcomment
 *   $ mv ocrcomment ~/bin
 *
 * Optionally ocrcomment will run the image through "textcleaner" if it is installed. This can
 * improve the results for many images.
 * The script is available from: http://www.fmwconcepts.com/imagemagick/textcleaner/index.php
 *
 * Usage
 * =====
 *   $ ocrcomment img.jpg
 *
 * You can check if the Spotlight Comment has been set:
 *   $ xattr -l img.jpg
 *
 */

if (!isset($_SERVER['argv'][1])) {
    echo "Usage: ocrcomment INPUT_IMAGE\n";
    exit(1);
}
if (empty(exec('which tesseract'))) {
    echo "Please install \"tesseract\".";
    exit(1);
}

$fileName = $cleanName = $_SERVER['argv'][1];

// If "textcleaner" is installed execute it
if (exec('which textcleaner')) {
    $cleanName = preg_replace('/\.(jpg|jpeg|png|gif)/', '.clean.$1', $fileName);
    exec("textcleaner \"$fileName\" \"$cleanName\"");
}

// Perform OCR
exec("tesseract \"$cleanName\" \"$fileName\"");

// Spotlight Comments must be a plist in binary form
// → http://stackoverflow.com/a/8555633/776654
file_put_contents("$fileName.plist", '<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd"><plist version="1.0"><string>'.file_get_contents("$fileName.txt").'</string></plist>');
exec("plutil -convert binary1 \"$fileName.plist\"");
// Set Spotlight Comment
exec("xattr -wx \"com.apple.metadata:kMDItemFinderComment\" \"`xxd -ps $fileName.plist`\" \"$fileName\"");

// Cleaning up
unlink("$fileName.txt");
unlink("$fileName.plist");
if ($fileName !== $cleanName) {
    unlink($cleanName);
}