<?php
/**
 * Persian text normalisation.
 *
 * @package Rahkar\WP
 */

declare( strict_types = 1 );

namespace Rahkar\WP\Modules\IranLocations;

defined( 'ABSPATH' ) || exit;

/**
 * Folds the variations that make two identical Persian names fail to match.
 *
 * Iranian address data arrives with Arabic ك and ي mixed into Persian ک and ی,
 * with zero-width non-joiners in some copies of a name and spaces in others, and
 * with Arabic-Indic digits alongside Persian ones. Compared raw, «کرمانشاه» from
 * one source and «كرمانشاه» from another are different strings — so a customer's
 * city silently fails validation for no visible reason.
 *
 * The same folding runs in the generator, so compiled data and runtime input are
 * always compared on equal terms.
 */
final class Normaliser {

	/**
	 * Character substitutions applied to every place name.
	 */
	private const FOLD = [
		// Arabic letters that look like their Persian counterparts.
		'ك'        => 'ک',
		'ي'        => 'ی',
		'ى'        => 'ی',
		'ة'        => 'ه',
		'ﻻ'        => 'لا',
		// Invisible characters that survive copy-paste.
		"\u{200C}" => ' ',  // zero-width non-joiner.
		"\u{200D}" => '',   // zero-width joiner.
		"\u{200E}" => '',   // left-to-right mark.
		"\u{200F}" => '',   // right-to-left mark.
		"\u{FEFF}" => '',   // byte-order mark.
		"\u{00A0}" => ' ',  // non-breaking space.
		// Diacritics, which are decorative in place names.
		"\u{064B}" => '',
		"\u{064C}" => '',
		"\u{064D}" => '',
		"\u{064E}" => '',
		"\u{064F}" => '',
		"\u{0650}" => '',
		"\u{0651}" => '',
		"\u{0652}" => '',
	];

	/**
	 * Arabic-Indic and Persian digits mapped to ASCII.
	 */
	private const DIGITS = [
		// Arabic-Indic.
		'٠' => '0',
		'١' => '1',
		'٢' => '2',
		'٣' => '3',
		'٤' => '4',
		'٥' => '5',
		'٦' => '6',
		'٧' => '7',
		'٨' => '8',
		'٩' => '9',
		// Persian.
		'۰' => '0',
		'۱' => '1',
		'۲' => '2',
		'۳' => '3',
		'۴' => '4',
		'۵' => '5',
		'۶' => '6',
		'۷' => '7',
		'۸' => '8',
		'۹' => '9',
	];

	/**
	 * Normalise a province or city name for storage and comparison.
	 *
	 * @param string $value Raw name.
	 */
	public static function place( string $value ): string {
		$value = strtr( $value, self::FOLD );
		$value = (string) preg_replace( '/\s+/u', ' ', $value );

		return trim( $value );
	}

	/**
	 * Normalise digits to ASCII, leaving everything else alone.
	 *
	 * @param string $value Raw text.
	 */
	public static function digits( string $value ): string {
		return strtr( $value, self::DIGITS );
	}
}