Class for working with data to be used in latin-1 encoded HTML documents, or for normalising user input.

charset-converter.php

<?php

class CharsetConverter
{
	private static $WORD_CHARS_MAP = array(
		"\xE2\x80\x9A" => "&sbquo;",
		"\xE2\x80\x9E" => "&bdquo;",
		"\xE2\x80\x98" => "'",
		"\xE2\x80\x99" => "'",
		"\xE2\x80\x9C" => "\"",
		"\xE2\x80\x9D" => "\"",
		"\xE2\x80\xA2" => "-",
		"\xE2\x80\x93" => "-",
		"\xE2\x80\x94" => "-",
		"\xE2\x80\xA6" => "...",
		"\xC2\xBC" => "&frac14;",
		"\xC2\xBD" => "&frac12;",
		"\xC2\xBE" => "&frac34;",
	);

	//==========================================================================

	/**
	 * Removes microsoft "smart quote" characters from the string
	 * @see https://github.com/pospi/cp1252fixer
	 *
	 * @return the string with characters stripped
	 */
	public static function stripMSCharacters($str)
	{
		return str_replace(array_keys(CharsetConverter::$WORD_CHARS_MAP), array_values(CharsetConverter::$WORD_CHARS_MAP), $str, $count);
	}

	/**
	 * Replaces all low (< 32) & high (> 0x9F) ASCII characters with HTML numbered entities
	 *
	 * @return the string with characters replaced
	 */
	public static function convertHighASCII($strin)
	{
		$strout = '';

		for ($i = 0; $i < strlen($strin); $i++) {
			$chr = $strin[$i];
			$ord = ord($chr);

			if ($ord > 0x9F) {
				$strout .= "&#{$ord};";
			} else {
				$strout .= $chr;
			}
		}

		return $strout;
	}

	/**
	 * Strips all low (< 32) & high (> 0x9F) ASCII characters from a string
	 *
	 * @return the string with characters stripped
	 */
	public static function stripHighASCII($strin)
	{
		$strout = '';

		for ($i = 0; $i < strlen($strin); $i++) {
			$chr = $strin[$i];
			$ord = ord($chr);

			if ($ord <= 0x9F) {
				$strout .= $chr;
			}
		}

		return $strout;
	}
}