<?php
/**
 * Turns names into comparable keys: lower case, no accents, no punctuation. Pure PHP.
 *
 * @package Tidelist
 */

namespace Tidelist\Mapping;

/**
 * Text normaliser.
 */
final class Normalizer {

	/**
	 * Latin letters with accents, as used in PL, RO, PT (and others), mapped to plain letters.
	 */
	const ACCENTS = array(
		'á' => 'a',
		'à' => 'a',
		'â' => 'a',
		'ä' => 'a',
		'ã' => 'a',
		'å' => 'a',
		'ą' => 'a',
		'ă' => 'a',
		'ç' => 'c',
		'ć' => 'c',
		'č' => 'c',
		'é' => 'e',
		'è' => 'e',
		'ê' => 'e',
		'ë' => 'e',
		'ę' => 'e',
		'í' => 'i',
		'ì' => 'i',
		'î' => 'i',
		'ï' => 'i',
		'ł' => 'l',
		'ñ' => 'n',
		'ń' => 'n',
		'ó' => 'o',
		'ò' => 'o',
		'ô' => 'o',
		'ö' => 'o',
		'õ' => 'o',
		'ø' => 'o',
		'ő' => 'o',
		'ř' => 'r',
		'ś' => 's',
		'š' => 's',
		'ș' => 's',
		'ş' => 's',
		'ț' => 't',
		'ţ' => 't',
		'ú' => 'u',
		'ù' => 'u',
		'û' => 'u',
		'ü' => 'u',
		'ű' => 'u',
		'ý' => 'y',
		'ÿ' => 'y',
		'ź' => 'z',
		'ż' => 'z',
		'ž' => 'z',
		'ß' => 'ss',
	);

	/**
	 * Comparable key of a text: "Dark Blue", "dark-blue" and "DARK  BLUE!" all give "dark blue".
	 *
	 * @param string $text Text (percent-encoded text is decoded first).
	 */
	public static function key( string $text ): string {
		$text = rawurldecode( $text );
		$text = function_exists( 'mb_strtolower' ) ? mb_strtolower( $text, 'UTF-8' ) : strtolower( $text );
		$text = strtr( $text, self::ACCENTS );
		$text = (string) preg_replace( '/[^\p{L}\p{N}]+/u', ' ', $text );
		return trim( $text );
	}

	/**
	 * The words of a text key.
	 *
	 * @param string $text Text.
	 * @return string[]
	 */
	public static function words( string $text ): array {
		$key = self::key( $text );
		return '' === $key ? array() : explode( ' ', $key );
	}
}