Nested Substrings

nest.php

<?php

/**
 * Return a substring, honoring nested strings
 *
 * Find from "start" to "end" in a given "haystack" string, 
 * but if a second "start" is encountered before the first "end",
 * skip over the nested repetition and return all the way up to the proper "end" for the outer "start"
 *
 * @author Brooks Boyd <[email protected]>
 * @param string $str The "haystack" string to search through for a substring
 * @param string $start A regular expression to be passed to preg_match() to find the beginning of the string to extract
 * @param string $end A regular expression to be passed to preg_match() to find the ending of the string to extract
 * @param int $offset Optional parameter to indicate how far into $str to start searching
 * @return bool|string Returns FALSE if error was encountered, or returns the substring (including $start and $end pieces)
 */
function find_nested($str, $start, $end, $offset = 0) {
  if (empty($str) || empty($start) || empty($end)) return false;
  
  if (preg_match($start, $str, $matches, PREG_OFFSET_CAPTURE, $offset)) {
    // Match was found
    $i = $matches[0][1]; // Grab the matched location
    $beginning = $i; // Where the target starts
    $level = 1; // One level deep at this point
    $offset = $i+strlen($matches[0][0]);
    
    while(true) {
      if (!preg_match($end, $str, $matches, PREG_OFFSET_CAPTURE, $offset)) {
        trigger_error('No ending string found', E_USER_WARNING);
        return false;
      }
      $next_end = $matches[0][1];
      $end_str = $matches[0][0];
      preg_match($start, $str, $matches, PREG_OFFSET_CAPTURE, $offset);
      $next_start = (count($matches) > 0)? $matches[0][1] : false;
      $start_str = (count($matches) > 0)? $matches[0][0] : false;

      if ($next_start !== false && $next_start < $next_end) {
        // There's a nested item here
        $level++;
        $offset = $next_start+strlen($start_str);
      } else {
        $level--;
        $offset = $next_end+strlen($end_str);
      }
      if ($level == 0) {
        // That's it, we're done!
        return substr($str, $beginning, $offset-$beginning);
      }
      if ($level > 100) {
        trigger_error('Too many levels deep!', E_USER_WARNING);
        return false;
      }
    }
  } else {
    trigger_error('No starting string found', E_USER_WARNING);
    return false;
  }
}