<?php
/**
 * Reads the data Yoast SEO leaves behind.
 *
 * @package ElevenRank
 * @since   1.0.0
 */

namespace ElevenRank\Modules\Import;

use ElevenRank\Modules\Meta\Keys;

defined( 'ABSPATH' ) || exit;

/**
 * Post meta `_yoast_wpseo_*`, term data in the option `wpseo_taxonomy_meta`, author meta `wpseo_*`
 * and the options `wpseo_titles` and `wpseo_social`. The keys were checked on a real install of
 * Yoast SEO 28.8 (see `docs/IMPORT_MAPPING.md`). Yoast's redirects are a paid feature whose storage
 * could not be checked here, so they are not imported.
 *
 * @since 1.0.0
 */
final class Yoast_Source implements Source {

	/**
	 * Meta key prefix.
	 *
	 * @since 1.0.0
	 *
	 * @var string
	 */
	const PREFIX = '_yoast_wpseo_';

	/**
	 * Variable converter.
	 *
	 * @since 1.0.0
	 *
	 * @var Variable_Converter
	 */
	private $variables;

	/**
	 * Constructor.
	 *
	 * @since 1.0.0
	 */
	public function __construct() {
		$this->variables = new Variable_Converter(
			'/%%([A-Za-z0-9_\-]+?)%%/',
			array(
				'title'                => 'title',
				'sitename'             => 'sitename',
				'sitedesc'             => 'sitedesc',
				'sep'                  => 'sep',
				'page'                 => 'page',
				'pagenumber'           => 'pagenumber',
				'pagetotal'            => 'pagetotal',
				'excerpt'              => 'excerpt',
				'excerpt_only'         => 'excerpt',
				'category'             => 'categories',
				'primary_category'     => 'category',
				'tag'                  => 'tag',
				'term_title'           => 'term',
				'term_description'     => 'term_description',
				'category_title'       => 'term',
				'category_description' => 'term_description',
				'name'                 => 'author',
				'date'                 => 'date',
				'modified'             => 'modified',
				'currentyear'          => 'currentyear',
				'currentmonth'         => 'currentmonth',
				'searchphrase'         => 'searchphrase',
				'pt_single'            => 'post_type',
			),
			static function ( $name ) {
				if ( 0 === strpos( $name, 'cf_' ) && strlen( $name ) > 3 ) {
					return '%custom_field(' . substr( $name, 3 ) . ')%';
				}

				if ( 0 === strpos( $name, 'ct_' ) && strlen( $name ) > 3 ) {
					return '%custom_tax(' . substr( $name, 3 ) . ')%';
				}

				return null;
			}
		);
	}

	/**
	 * Source id.
	 *
	 * @since 1.0.0
	 *
	 * @return string
	 */
	public function id(): string {
		return 'yoast';
	}

	/**
	 * Plugin name.
	 *
	 * @since 1.0.0
	 *
	 * @return string
	 */
	public function label(): string {
		return 'Yoast SEO';
	}

	/**
	 * Whether Yoast left data.
	 *
	 * @since 1.0.0
	 *
	 * @return bool
	 */
	public function detected(): bool {
		return false !== get_option( 'wpseo_titles', false ) || Source_Support::count_with_meta( 'postmeta', 'post_id', self::PREFIX ) > 0;
	}

	/**
	 * Whether Yoast is active.
	 *
	 * @since 1.0.0
	 *
	 * @return bool
	 */
	public function is_active(): bool {
		return defined( 'WPSEO_VERSION' );
	}

	/**
	 * What there is to import.
	 *
	 * @since 1.0.0
	 *
	 * @return array{posts: int, terms: int, authors: int, redirects: int, settings: bool}
	 */
	public function counts(): array {
		$terms = 0;

		foreach ( $this->term_data() as $entries ) {
			$terms += count( $entries );
		}

		return array(
			'posts'     => Source_Support::count_with_meta( 'postmeta', 'post_id', self::PREFIX ),
			'terms'     => $terms,
			'authors'   => Source_Support::count_with_meta( 'usermeta', 'user_id', 'wpseo_' ),
			'redirects' => 0,
			'settings'  => false !== get_option( 'wpseo_titles', false ),
		);
	}

	/**
	 * A batch of posts.
	 *
	 * @since 1.0.0
	 *
	 * @param int $after Only posts with a higher id.
	 * @param int $limit Most posts to return.
	 * @return array<int, array{id: int, fields: array<string, mixed>, notes: string[]}>
	 */
	public function posts( int $after, int $limit ): array {
		$ids  = Source_Support::ids_with_meta( 'postmeta', 'post_id', self::PREFIX, $after, $limit );
		$meta = Source_Support::meta_for( 'postmeta', 'post_id', $ids, self::PREFIX );
		$out  = array();

		foreach ( $ids as $id ) {
			$out[ $id ]       = $this->record( $meta[ $id ] ?? array(), self::PREFIX, true );
			$out[ $id ]['id'] = $id;
		}

		return $out;
	}

	/**
	 * A batch of terms.
	 *
	 * @since 1.0.0
	 *
	 * @param int $after Only terms with a higher id.
	 * @param int $limit Most terms to return.
	 * @return array<int, array{id: int, fields: array<string, mixed>, notes: string[]}>
	 */
	public function terms( int $after, int $limit ): array {
		$all = array();

		foreach ( $this->term_data() as $entries ) {
			foreach ( $entries as $term_id => $values ) {
				if ( (int) $term_id > $after && is_array( $values ) ) {
					$all[ (int) $term_id ] = $values;
				}
			}
		}

		ksort( $all );

		$out = array();

		foreach ( array_slice( $all, 0, $limit, true ) as $term_id => $values ) {
			$prefixed = array();

			foreach ( $values as $key => $value ) {
				$name = 0 === strpos( (string) $key, 'wpseo_' ) ? substr( (string) $key, 6 ) : (string) $key;

				$prefixed[ 'desc' === $name ? 'metadesc' : $name ] = is_scalar( $value ) ? (string) $value : '';
			}

			$out[ $term_id ]       = $this->record( $prefixed, '', false, $this->term_robots( $prefixed ) );
			$out[ $term_id ]['id'] = $term_id;
		}

		return $out;
	}

	/**
	 * A batch of authors.
	 *
	 * @since 1.0.0
	 *
	 * @param int $after Only users with a higher id.
	 * @param int $limit Most users to return.
	 * @return array<int, array{id: int, fields: array<string, mixed>, notes: string[]}>
	 */
	public function authors( int $after, int $limit ): array {
		$ids  = Source_Support::ids_with_meta( 'usermeta', 'user_id', 'wpseo_', $after, $limit );
		$meta = Source_Support::meta_for( 'usermeta', 'user_id', $ids, 'wpseo_' );
		$out  = array();

		foreach ( $ids as $id ) {
			$values = $meta[ $id ] ?? array();
			$notes  = array();
			$title  = $this->template( $values['wpseo_title'] ?? '', $notes );
			$desc   = $this->template( $values['wpseo_metadesc'] ?? '', $notes );

			$out[ $id ] = array(
				'id'     => $id,
				'fields' => Source_Support::filled(
					array(
						Keys::TITLE       => $title,
						Keys::DESCRIPTION => $desc,
						Keys::ROBOTS      => 'on' === ( $values['wpseo_noindex_author'] ?? '' ) ? array( 'noindex' => true ) : array(),
					)
				),
				'notes'  => array_values( $notes ),
			);
		}

		return $out;
	}

	/**
	 * The site-wide settings.
	 *
	 * @since 1.0.0
	 *
	 * @return array{settings: array<string, array<string, mixed>>, notes: string[]}
	 */
	public function settings(): array {
		$titles = get_option( 'wpseo_titles', array() );
		$social = get_option( 'wpseo_social', array() );
		$titles = is_array( $titles ) ? $titles : array();
		$social = is_array( $social ) ? $social : array();
		$notes  = array();
		$out    = array(
			'titles'  => array(),
			'social'  => array(),
			'profile' => array(),
		);

		$separators = array(
			'sc-dash'   => '-',
			'sc-ndash'  => '–',
			'sc-mdash'  => '—',
			'sc-pipe'   => '|',
			'sc-middot' => '·',
			'sc-bull'   => '•',
			'sc-tilde'  => '~',
			'sc-raquo'  => '»',
			'sc-laquo'  => '«',
		);

		if ( isset( $titles['separator'], $separators[ $titles['separator'] ] ) ) {
			$out['titles']['separator'] = $separators[ $titles['separator'] ];
		}

		$out['titles']['home_title']         = $this->template( (string) ( $titles['title-home-wpseo'] ?? '' ), $notes );
		$out['titles']['home_description']   = $this->template( (string) ( $titles['metadesc-home-wpseo'] ?? '' ), $notes );
		$out['titles']['author_title']       = $this->template( (string) ( $titles['title-author-wpseo'] ?? '' ), $notes );
		$out['titles']['author_description'] = $this->template( (string) ( $titles['metadesc-author-wpseo'] ?? '' ), $notes );
		$out['titles']['date_title']         = $this->template( (string) ( $titles['title-archive-wpseo'] ?? '' ), $notes );
		$out['titles']['date_description']   = $this->template( (string) ( $titles['metadesc-archive-wpseo'] ?? '' ), $notes );
		$out['titles']['search_title']       = $this->template( (string) ( $titles['title-search-wpseo'] ?? '' ), $notes );
		$out['titles']['not_found_title']    = $this->template( (string) ( $titles['title-404-wpseo'] ?? '' ), $notes );

		foreach ( array(
			'author_noindex' => 'noindex-author-wpseo',
			'date_noindex'   => 'noindex-archive-wpseo',
		) as $target => $key ) {
			if ( array_key_exists( $key, $titles ) ) {
				$out['titles'][ $target ] = (bool) $titles[ $key ];
			}
		}

		if ( array_key_exists( 'noindex-author-noposts-wpseo', $titles ) ) {
			$out['titles']['author_noindex_single_author'] = (bool) $titles['noindex-author-noposts-wpseo'];
		}

		foreach ( get_post_types( array( 'public' => true ) ) as $type ) {
			$entry = $this->entry( $titles, $type, '', $notes );

			if ( $entry ) {
				$out['titles']['post_types'][ $type ] = $entry;
			}
		}

		foreach ( get_taxonomies( array( 'public' => true ) ) as $taxonomy ) {
			$entry = $this->entry( $titles, $taxonomy, 'tax-', $notes );

			if ( $entry ) {
				$out['titles']['taxonomies'][ $taxonomy ] = $entry;
			}
		}

		if ( ! empty( $social['twitter_site'] ) ) {
			$out['social']['twitter_site'] = ltrim( (string) $social['twitter_site'], '@' );
			$out['profile']['x']           = 'https://x.com/' . rawurlencode( $out['social']['twitter_site'] );
		}

		if ( isset( $social['twitter_card_type'] ) && in_array( $social['twitter_card_type'], array( 'summary', 'summary_large_image' ), true ) ) {
			$out['social']['twitter_card_type'] = $social['twitter_card_type'];
		}

		$image = Source_Support::image_id( $social['og_default_image_id'] ?? 0, (string) ( $social['og_default_image'] ?? '' ) );

		if ( $image > 0 ) {
			$out['social']['default_image_id'] = $image;
		}

		foreach ( array(
			'facebook_site' => 'facebook',
			'instagram_url' => 'instagram',
			'linkedin_url'  => 'linkedin',
			'youtube_url'   => 'youtube',
			'pinterest_url' => 'pinterest',
		) as $key => $target ) {
			if ( ! empty( $social[ $key ] ) ) {
				$out['profile'][ $target ] = (string) $social[ $key ];
			}
		}

		$person = 'person' === ( $titles['company_or_person'] ?? '' );

		$out['profile']['entity_type'] = $person ? 'person' : 'organization';

		$name = $person ? ( $titles['company_or_person_user_id'] ?? '' ) : ( $titles['company_name'] ?? '' );

		if ( $person && is_numeric( $name ) ) {
			$user = get_userdata( (int) $name );
			$name = $user ? $user->display_name : '';
		}

		$out['profile']['entity_name'] = is_string( $name ) ? $name : '';

		if ( ! $person ) {
			$logo = Source_Support::image_id( $titles['company_logo_id'] ?? 0, (string) ( $titles['company_logo'] ?? '' ) );

			if ( $logo > 0 ) {
				$out['profile']['entity_logo_id'] = $logo;
			}
		}

		return array(
			'settings' => $out,
			'notes'    => array_values( $notes ),
		);
	}

	/**
	 * Yoast keeps its redirects in a paid add-on, so there is nothing to read.
	 *
	 * @since 1.0.0
	 *
	 * @param int $after Unused.
	 * @param int $limit Unused.
	 * @return array<int, array{id: int, rules: array<int, array<string, mixed>>, notes: string[]}>
	 */
	public function redirects( int $after, int $limit ): array {
		return array();
	}

	/**
	 * Turns Yoast's values for one object into a record.
	 *
	 * @since 1.0.0
	 *
	 * @param array<string, string> $meta     Values by key.
	 * @param string                $prefix   Key prefix to strip (post meta only).
	 * @param bool                  $is_post  Whether the keys are post meta keys.
	 * @param array<string, bool>   $robots   Robots flags worked out by the caller (terms).
	 * @return array{id: int, fields: array<string, mixed>, notes: string[]}
	 */
	private function record( array $meta, string $prefix, bool $is_post, array $robots = array() ): array {
		$values = array();

		foreach ( $meta as $key => $value ) {
			$values[ '' !== $prefix && 0 === strpos( $key, $prefix ) ? substr( $key, strlen( $prefix ) ) : $key ] = $value;
		}

		$notes = array();
		$get   = static function ( $name ) use ( $values ) {
			return isset( $values[ $name ] ) ? trim( $values[ $name ] ) : '';
		};

		if ( $is_post ) {
			$robots = array();

			if ( '1' === $get( 'meta-robots-noindex' ) ) {
				$robots['noindex'] = true;
			} elseif ( '2' === $get( 'meta-robots-noindex' ) ) {
				$robots['noindex'] = false;
			}

			if ( '1' === $get( 'meta-robots-nofollow' ) ) {
				$robots['nofollow'] = true;
			}

			foreach ( explode( ',', $get( 'meta-robots-adv' ) ) as $flag ) {
				$flag = trim( $flag );

				if ( in_array( $flag, array( 'noimageindex', 'noarchive', 'nosnippet' ), true ) ) {
					$robots[ $flag ] = true;
				}
			}
		}

		$fields = array(
			Keys::TITLE               => $this->template( $get( 'title' ), $notes ),
			Keys::DESCRIPTION         => $this->template( $get( 'metadesc' ), $notes ),
			Keys::FOCUS_KEYWORDS      => '' !== $get( 'focuskw' ) ? Source_Support::keywords( $get( 'focuskw' ) ) : array(),
			Keys::CANONICAL           => esc_url_raw( $get( 'canonical' ) ),
			Keys::ROBOTS              => $robots,
			Keys::BREADCRUMB_TITLE    => sanitize_text_field( $get( 'bctitle' ) ),
			Keys::OG_TITLE            => $this->template( $get( 'opengraph-title' ), $notes ),
			Keys::OG_DESCRIPTION      => $this->template( $get( 'opengraph-description' ), $notes ),
			Keys::OG_IMAGE            => Source_Support::image_id( $get( 'opengraph-image-id' ), $get( 'opengraph-image' ) ),
			Keys::TWITTER_TITLE       => $this->template( $get( 'twitter-title' ), $notes ),
			Keys::TWITTER_DESCRIPTION => $this->template( $get( 'twitter-description' ), $notes ),
			Keys::TWITTER_IMAGE       => Source_Support::image_id( $get( 'twitter-image-id' ), $get( 'twitter-image' ) ),
		);

		if ( $is_post && '1' === $get( 'is_cornerstone' ) ) {
			$fields[ Keys::PILLAR ] = true;
		}

		return array(
			'id'     => 0,
			'fields' => Source_Support::filled( $fields ),
			'notes'  => array_values( $notes ),
		);
	}

	/**
	 * Robots flags of a term entry.
	 *
	 * @since 1.0.0
	 *
	 * @param array<string, string> $values Term values with `wpseo_` removed from the keys.
	 * @return array<string, bool>
	 */
	private function term_robots( array $values ): array {
		$noindex = $values['noindex'] ?? '';

		if ( 'noindex' === $noindex ) {
			return array( 'noindex' => true );
		}

		return 'index' === $noindex ? array( 'noindex' => false ) : array();
	}

	/**
	 * Converts a template and remembers what could not be converted.
	 *
	 * @since 1.0.0
	 *
	 * @param string   $template Yoast template.
	 * @param string[] $notes    Notes to add to.
	 * @return string
	 */
	private function template( string $template, array &$notes ): string {
		$converted = $this->variables->convert( $template );

		foreach ( $converted['unsupported'] as $name ) {
			$notes[ $name ] = $name;
		}

		return $converted['text'];
	}

	/**
	 * Title, description and noindex of a post type or taxonomy.
	 *
	 * @since 1.0.0
	 *
	 * @param array<string, mixed> $titles The `wpseo_titles` option.
	 * @param string               $name   Post type or taxonomy name.
	 * @param string               $infix  `tax-` for taxonomies, empty for post types.
	 * @param string[]             $notes  Notes to add to.
	 * @return array<string, mixed>
	 */
	private function entry( array $titles, string $name, string $infix, array &$notes ): array {
		$entry = array(
			'title'       => $this->template( (string) ( $titles[ 'title-' . $infix . $name ] ?? '' ), $notes ),
			'description' => $this->template( (string) ( $titles[ 'metadesc-' . $infix . $name ] ?? '' ), $notes ),
		);

		if ( array_key_exists( 'noindex-' . $infix . $name, $titles ) ) {
			$entry['noindex'] = (bool) $titles[ 'noindex-' . $infix . $name ];
		}

		return array_filter(
			$entry,
			static function ( $value ) {
				return '' !== $value;
			}
		);
	}

	/**
	 * Yoast's term data: taxonomy => term id => values.
	 *
	 * @since 1.0.0
	 *
	 * @return array<string, array<int|string, mixed>>
	 */
	private function term_data(): array {
		$data = get_option( 'wpseo_taxonomy_meta', array() );

		return is_array( $data ) ? $data : array();
	}
}