<?php
	// echo "BOOST_CACHE_WARMER.php v.0.71\n";

	// CACHE WARMER W/ DRUPAL BOOST AND XML SITEMAP
	// this script traverses the $base_sitemap, then individual XML files
	// traverses URL's comapres them to boost cache located at $base_path
	// if no cache file exists makes a request to create it
	// limited to $max_updated # of files to boost
	// meant to be run by cron

	//check for matching file file; w/o http:// & w/ _.html
	$base_path = "/path/to/d7_install/cache/normal/"; // path to Boost cache root
	$base_sitemap = "http://example.com/sitemap.xml"; // starting sitemap
	$max_updated = 50;	// max nodes create cache for per run


	echo("\n-- Starting BOOST_CACHE_WARMER @ ". date('l jS \of F Y h:i:s A') ); // log

	// load up base sitemap
	$ch = curl_init();
	curl_setopt($ch, CURLOPT_URL, $base_sitemap);
	curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
	$data = curl_exec ($ch);
	curl_close ($ch);
	$xml_file_list = new SimpleXMLElement($data);

	$index_reviewed = 0;
	$index_updated = 0;
	// traverse list of XML files from sitemap
	foreach ($xml_file_list->sitemap as $xml_file_url_list) {
		if ($index_updated >= $max_updated) {break;}; // abort if done max
		$xml_file_url = $xml_file_url_list->loc;
		echo("Starting on file: ". $xml_file_url ); // log

		// traverse XML file getting URL's
		$ch = curl_init();
		curl_setopt($ch, CURLOPT_URL, $xml_file_url);
		curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
		$data = curl_exec ($ch);
		curl_close ($ch);
		$xml = new SimpleXMLElement($data);

		// for each URL
		foreach ($xml->url as $url_list) {
			$index_reviewed++;	//index URL's reviewed
			if ($index_updated >= $max_updated) {break;}; // abort if done max
			// echo " DOES NOT EXIST, create it.\n";

			$url = $url_list->loc;
			//check for matching file file; w/o http:// & w/ _.html
			$temp_file =  $base_path . substr($url, 7) ."_.html";
			//echo($temp_file ."\n"); // log

			// check for file
			if (!(file_exists($temp_file))) {
					// echo " DOES NOT EXIST. create via Curl request\n";
					$ch = curl_init();
					curl_setopt($ch, CURLOPT_URL, $url);
					curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
					$data = curl_exec ($ch);
					curl_close ($ch);
					$index_updated++;	//index URL's updated

			}else{
					// echo " DOES EXIST.\n";
			}

		}
	// Done w/ file
	echo("Ending on file: ". $xml_file_url);
	}
		echo("Reviewed: ". $index_reviewed ." | Updated: ". $index_updated );
		echo("\n-- Ending BOOST_CACHE_WARMER @ ". date('l jS \of F Y h:i:s A') ."\n\n");
?>