These days I needed a script to backup only a part of a customers website using a CRON. Most of the control panels I know allow only a complete website backup and this is not what I needed. While plaaning the script, I thought about a solution for webmaster without full SSH access to their hosting account. A typical situation could be:

• A shared hosting account that allows only backups for the whole site incl. database, emails and other settings
• No administration rights via SSH
• A FTP host for the storage of the the backup files
• Support for cURL and a default PHP5 configuration (sorry no more code for PHP4).

First of all we need to create a variable to define the local source directory on your website where you like to start the backup from:

define('BASEPATH', $_SERVER['DOCUMENT_ROOT'].'/');

The next code is to create a new class including the constructor function and some variables used in that class:

class Curl_ftp_backup {
 
	public $basePath, $errors, $ftp_server, $ftp_user_pw , $msg, $email;
 
	public function __construct($ftp_server, $ftp_userpw, $email_adr) {
		$this->basePath = BASEPATH;
		$this->errors = '';
		$this->ftp_server = $ftp_server;
		$this->ftp_user_pw = $ftp_userpw;
		$this->email = $email_adr;
		$this->msg = '';
	}
}

Our class will store file and directories on the remote FTP server, the next function will create a directory if it not already exists:

	private function curl_create_ftp_directory($name, $curr_dir = '') {
		$ch = curl_init();
		curl_setopt($ch, CURLOPT_URL, $this->ftp_server);
		curl_setopt($ch, CURLOPT_USERPWD, $this->ftp_user_pw);
		curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
		$cmd = array();
		if ($curr_dir != '') $cmd[] = 'CWD '.$curr_dir;
		$cmd[] = 'MKD '.$name;
		curl_setopt($ch, CURLOPT_POSTQUOTE, $cmd);
		curl_exec ($ch);
		return;
	}

The following function will open a file and uploads the data to specified directory on the target server. For each file that is not stored the error number is returned, this information is used later to build up an error string.

	private function curl_put_file($path) {
		$ch = curl_init();
		$fp = fopen($this->basePath.$path, "r");
		curl_setopt($ch, CURLOPT_URL, $this->ftp_server.$path);
		curl_setopt($ch, CURLOPT_USERPWD, $this->ftp_user_pw);
		curl_setopt($ch, CURLOPT_UPLOAD, 1);
		curl_setopt($ch, CURLOPT_INFILE, $fp);
		curl_setopt($ch, CURLOPT_INFILESIZE, filesize($this->basePath.$path));
		curl_exec ($ch);
		$error = curl_errno($ch);
		curl_close ($ch);
		if ($error != 0) $this->errors .= $path.PHP_EOL;
		return;
	}

Now we need a function that will loop through the specified base directory and will switch between directories and files; the previous functions will create the directory or upload the file.

	public function backupfiles($directory) {
		if ($handle = opendir($this->basePath.$directory)) {
			$this->curl_create_ftp_directory($directory);
			while (false !== ($file = readdir($handle))) {
				if (is_file($this->basePath.$directory.$file) && $file != '.htaccess') {
					$this->curl_put_file($directory.$file);
				} elseif ($file != '.' && $file != '..' && is_dir($this->basePath.$directory.$file)) {
					$this->backupfiles($directory.$file.'/', $directory);
				}
			}
			closedir($handle);
			if ($this->errors != '') {
				$this->msg = 'Missing files: '.PHP_EOL.$this->errors;
			} else {
				$this->msg = 'FTP upload successful.';
			}
		} else {
			$this->msg = 'Can\'t open local directory.';
		}
	}
 
	public function send_emailmsg() {
		if ($this->msg != '') {
			$body = PHP_EOL.'Backup result:'.PHP_EOL.PHP_EOL.$this->msg;
			mail($this->email, 'Your Backup on '.date('Y-m-d H:i:s'), $body);
		}
	}

At last there is a tiny function that will send you some short email message about the backup result. You can call the class in your PHP script:

$cb = new Curl_ftp_backup('ftp://ftp.yourserver.com/source_folder/', 'user:password', 'your@email.com');
$cb->backupfiles('includes/');
$cb->send_emailmsg();

Thats all, now you need to create a CRON job for the file that will execute the backup. If your hosting provider doesn’t provide CRON jobs, just google for “free cron jobs”.

Disclaimer:
We tested the script “successful” on two modern Linux web servers with ~100 files between 10kb and 300MB. If you have a backup with a lot of files you should think about a different solution. With every file or directory a new CURL function is called, this could result in a high server load. We are not responsible for any damage and/or data loss as the result off using this script.

Click here to download the PHP class script.Similar Posts:

• Tutorial: FTP Upload via cURL
• Curl: Location redirect while open_basedir is set
• Tutorial: Create a zip file from folders on the fly

For the most of the PHP developer which are using preg_match or preg_replace frequently is the function preg_match_all a smaller advantage, but for all others it’s maybe hard to understand. The biggest difference between preg_match_all and the regular preg_match is that all matched values are stored inside a multi-dimensional array to store an unlimited number of matches. With the following example I will try to make clear how its possible to store the image paths inside a web page:

$data = file_get_contents("http://www.finalwebsites.com");
$pattern = "/src=[\\"']?([^\\"']?.*(png|jpg|gif))[\\"']?/i";
preg_match_all($pattern, $data, $images);

We take a closer look to the pattern:

"/src=[\\"']?([^\\"']?.*(png|jpg|gif))[\\"']?/i"

The first part and the last part are searching for everything that starts with src and ends with a optional quote or double quote. This could be a long string because the outer rule is very global. Next we check the rule starts within the first bracket:

"/src=[\\"']?([^\\"']?.*(png|jpg|gif))[\\"']?/i"

Now we are looking inside this long string from the outer rule for strings starting with an optional quote or double quote followed by any characters. The last part inside the inner brackets is the magic:

"/src=[\\"']?([^\\"']?.*(png|jpg|gif))[\\"']?/i"

We are looking next for a string that is followed by a file extension and match we get all the paths from the html file.

We need all the rules to isolate the string parts (image paths) from the rest of the html. The result looks like this (access the array $images with these indexes, or just use print_r($images)):

$images[0][0] -> src="/images/english.gif"
$images[1][0] -> /images/english.gif
$images[2][0] -> gif

The index 1 is the information we need, try this example with other part of html code for a better understanding. Check our partner finalwebsites.com for more scripts/snippets.

Similar Posts:

• Modifiers in regular expression patterns
• Upload images for usage in TinyMCE
• Upload in modal window and pass values with jQuery