TPCUtil.php

<?php

/**
  * Miscellaneous helper functions that don't fit anywhere else.
  *
  * @file
  * @author Nmlgc
  */

class TPCUtil {
	/**
	  * Normalizes a hook name.
	  *
	  * @param string &$hook Hook name
	  * @return string Normalized hook name.
	  */
	public static function normalizeHook( $hook ) {
		// Normalize hook name... this should totally do for now
		$hook = strtolower( $hook );
		$hook = preg_replace( '/ /', '_', $hook );
		return $hook;
	}

	/**
	  * Checks whether an array is associative or numeric.
	  *
	  * @param array &$array The array to check.
	  * @return int > 0: associative, = 0: numeric, < 0: both
	  */
	public static function isAssoc( &$array ) {
		if ( !is_array( $array ) ) {
			return 0;
		}
		return count( array_filter( array_keys( $array ), 'is_string' ) );
	}

	/**
	  * Sanitizes a wikitext string for in-game display.
	  *
	  * @param string &$param The string to sanitize.
	  * @param bool $escape_percents Keep literal percent signs for a printf format string.
	  * @return string Sanitized string.
	  */
	public static function sanitize( &$param, $escape_percents = true ) {
		/**
		  * Remove translation markup. And yes, we use regex to do it.
		  * Justification:
		  * * The source format is plaintext, not XML. In fact, this is the only
		  *   instance of XML tags being used inside translatable strings.
		  * * Thus, using full-fledged XML parsers is counter-productive. PHP's
		  *   DOM, for example, ignores everything up to the first HTML tag.
		  *   Should we build some tags around the text by default just to work
		  *   around this now?
		  * * Some of the translation units extend to more template parameters
		  *   and we might not even have an opening or closing tag due to this.
		  * * Even worse, DOM might get confused with our layout markup... and
		  *   by working around *that*, we have ultimately defeated any reason
		  *   to use DOM in the first place.
		  */
		$param = preg_replace( '~<translate>\n?~', '', $param );
		$param = preg_replace( '~</translate>~', '', $param );
		$param = preg_replace( '~<!--T\:(.*?)-->\n?~', '', $param );

		// HTML comments
		$param = preg_replace( '/<!--(.*?)-->/', '', $param );

		// Remove {{lang}} wrappers
		$REGEX_LANG_PAT = '/\{\{\s*lang*\s*\|.*?\|\s*(.*?)\s*\}\}/i';
		$REGEX_LANG_REP = '\1';
		$param = preg_replace( $REGEX_LANG_PAT, $REGEX_LANG_REP, $param );
		// &nbsp;
		$param = preg_replace( '/&nbsp;/', "\u{00a0}", $param );
		// space
		$param = preg_replace( '/&#32;/', " ", $param );
		// MediaWiki markup
		$param = preg_replace( "/'''''(.*?)'''''/", '<bi$\1>', $param );
		$param = preg_replace( "/'''(.*?)'''/", '<b$\1>', $param );
		$param = preg_replace( "/''(.*?)''/", '<i$\1>', $param );
		$param = preg_replace( "/<u>(.*?)<\/u>/", '<u$\1>', $param );
		$param = preg_replace( "/<ins>(.*?)<\/ins>/", '<u$\1>', $param );
		$param = preg_replace( "/<br ?\\/?>/", "\n", $param );
		if ( $escape_percents ) {
			$param = preg_replace( '/%/', '%%', $param );
		}
		// TL notes. We don't remove trailing whitespace here because you
		// can *technically* have more than one per template…
		$REGEX_TLNOTE = '/(\n)*\{\{\s*tlnote\s*\|\s*(.*?)(\|.*)*}}/is';
		if ( preg_match_all( $REGEX_TLNOTE, $param, $tlnotes ) ) {
			$param = preg_replace( $REGEX_TLNOTE, '', $param );
			$param .= "\u{0014}" . implode( $tlnotes[2] );
		}
		// Do more MediaWiki stuff...
		return $param;
	}


	/**
	  * Returns a table cell with the full GeoIP location information for a given URL.
	  *
	  * @param string &$url
	  * @return string
	  */
	public static function getGeoIPCell( &$url ) {
		if ( function_exists( 'geoip_record_by_name' ) ) {
			$parsedURL = parse_url( $url, PHP_URL_HOST );
			$record = geoip_record_by_name( $parsedURL );

			$ret = "| ";
			if ( $record['city'] ) {
				$ret .= "{$record['city']}, ";
			}
			if ( $record['region'] ) {
				$ret .= "{$record['region']}, ";
			}
			$ret .= "{$record['country_name']}, ";
			$ret .= "{$record['continent_code']}";
			$ret .= "\n";
		} else {
			$ret = '';
		}
		return $ret;
	}
}