<?php
defined( 'ABSPATH' ) || exit;

/**
 * Discovers all publicly accessible URLs on the site and manages
 * per-URL scan results.
 *
 * Storage strategy (for performance at scale):
 *   • Pages / posts / CPTs  → post meta (`_accessmate_scan`) on the post itself.
 *     WordPress handles cleanup automatically when a post is deleted, and
 *     post-meta queries are indexed.
 *   • Non-post URLs (homepage, archives, categories) → individual wp_options
 *     rows with autoload = false, keyed by OPT_PREFIX . md5($url).
 *     A lightweight index option (OPT_INDEX) tracks which hashes exist so we
 *     never need a wildcard LIKE query.
 *
 * This replaces the previous single `accessmate_page_scans` option which was
 * autoloaded and grew unboundedly, hurting memory on every page request.
 */
class AccessMate_Site_Scanner {

    private const META_KEY   = '_accessmate_scan';
    private const OPT_PREFIX = 'accessmate_scan_';
    private const OPT_INDEX  = 'accessmate_scan_index'; // array of md5 hashes for non-post URLs
    private const LEGACY_KEY = 'accessmate_page_scans'; // migrated on first call

    // Maximum number of individual posts to return per CPT (prevents huge lists).
    private const MAX_POSTS_PER_TYPE = 50;

    // ── URL discovery ─────────────────────────────────────────

    /**
     * Returns an ordered list of scannable URLs.
     * Each entry: [ 'url' => string, 'type' => string, 'label' => string ]
     */
    public static function get_scannable_urls(): array {
        $urls = [];
        $seen = [];

        $add = static function ( string $url, string $type, string $label ) use ( &$urls, &$seen ): void {
            $url = trailingslashit( esc_url_raw( $url ) );
            if ( empty( $url ) || isset( $seen[ $url ] ) ) return;
            $seen[ $url ] = true;
            $urls[]        = compact( 'url', 'type', 'label' );
        };

        // Homepage
        $add( home_url( '/' ), 'home', 'Homepage' );

        // Blog index (if different from homepage)
        $blog_page_id = (int) get_option( 'page_for_posts' );
        if ( $blog_page_id ) {
            $add( get_permalink( $blog_page_id ), 'blog', 'Blog Index' );
        }

        // Pages
        $pages = get_posts( [
            'post_type'      => 'page',
            'posts_per_page' => -1,
            'post_status'    => 'publish',
            'orderby'        => 'menu_order',
            'order'          => 'ASC',
        ] );
        foreach ( $pages as $page ) {
            $add( (string) get_permalink( $page ), 'page', get_the_title( $page ) );
        }

        // Posts
        $posts = get_posts( [
            'post_type'      => 'post',
            'posts_per_page' => self::MAX_POSTS_PER_TYPE,
            'post_status'    => 'publish',
            'orderby'        => 'date',
            'order'          => 'DESC',
        ] );
        foreach ( $posts as $post ) {
            $add( (string) get_permalink( $post ), 'post', get_the_title( $post ) );
        }

        // Category & tag archives
        $categories = get_categories( [ 'hide_empty' => true, 'number' => 20 ] );
        foreach ( $categories as $cat ) {
            $url = get_category_link( $cat->term_id );
            if ( $url ) $add( $url, 'category', 'Category: ' . $cat->name );
        }

        // Custom post types
        $cpts = get_post_types( [ '_builtin' => false, 'public' => true ], 'objects' );
        foreach ( $cpts as $cpt ) {
            if ( $cpt->has_archive ) {
                $archive_url = get_post_type_archive_link( $cpt->name );
                if ( $archive_url ) {
                    $add( $archive_url, 'archive', $cpt->label . ' Archive' );
                }
            }
            $cpt_posts = get_posts( [
                'post_type'      => $cpt->name,
                'posts_per_page' => self::MAX_POSTS_PER_TYPE,
                'post_status'    => 'publish',
                'orderby'        => 'date',
                'order'          => 'DESC',
            ] );
            foreach ( $cpt_posts as $cpt_post ) {
                $add( (string) get_permalink( $cpt_post ), $cpt->name, get_the_title( $cpt_post ) );
            }
        }

        return $urls;
    }

    // ── Storage ───────────────────────────────────────────────

    /**
     * Stores a browser-computed axe-core scan result for $url.
     * Routes to post meta or a dedicated non-autoloaded option depending on
     * whether $url maps to a known post/page/CPT.
     */
    public static function store_axe_scan( string $url, array $result ): void {
        $url = trailingslashit( $url );

        $stored = [
            'url'        => $url,
            'score'      => $result['score']      ?? 0,
            'grade'      => $result['grade']      ?? 'F',
            'failed'     => $result['failed']     ?? 0,
            'issues'     => $result['issues']     ?? [],
            'scanned_at' => $result['scanned_at'] ?? gmdate( 'c' ),
        ];

        $post_id = url_to_postid( $url );

        if ( $post_id > 0 ) {
            // Post / page / CPT — use post meta (indexed, auto-cleaned with post).
            update_post_meta( $post_id, self::META_KEY, $stored );
        } else {
            // Non-post URL (homepage, archive, category …).
            $hash = md5( $url );
            update_option( self::OPT_PREFIX . $hash, $stored, false ); // autoload = false

            // Register the hash in the index so get_page_scans() can find it.
            $index = get_option( self::OPT_INDEX, [] );
            if ( ! in_array( $hash, $index, true ) ) {
                $index[] = $hash;
                update_option( self::OPT_INDEX, $index, false );
            }
        }
    }

    /**
     * Returns the full map of per-URL scan summaries, keyed by URL.
     * Merges post-meta scans and non-post option scans efficiently.
     *
     * @return array<string, array>
     */
    public static function get_page_scans(): array {
        self::maybe_migrate_legacy();

        $scans = [];

        // ── 1. Post-meta scans (single direct DB query, no full table scan) ──
        global $wpdb;
        $rows = $wpdb->get_results(           // phpcs:ignore WordPress.DB.DirectDatabaseQuery
            $wpdb->prepare(
                "SELECT meta_value FROM {$wpdb->postmeta} WHERE meta_key = %s",
                self::META_KEY
            )
        );
        foreach ( $rows as $row ) {
            $data = maybe_unserialize( $row->meta_value );
            if ( is_array( $data ) && isset( $data['url'] ) ) {
                $scans[ $data['url'] ] = $data;
            }
        }

        // ── 2. Non-post scans (index → individual non-autoloaded options) ────
        $index = get_option( self::OPT_INDEX, [] );
        foreach ( $index as $hash ) {
            $data = get_option( self::OPT_PREFIX . $hash );
            if ( is_array( $data ) && isset( $data['url'] ) ) {
                $scans[ $data['url'] ] = $data;
            }
        }

        return $scans;
    }

    /**
     * PHP-scanner path (legacy, non-axe): runs AccessMate_Scanner on $url and
     * stores a summary. Kept for compatibility but axe-core path is preferred.
     */
    public static function scan_page( string $url ): array {
        $result = AccessMate_Scanner::run( $url );

        self::store_axe_scan( $url, [
            'score'      => $result['score']  ?? 0,
            'grade'      => $result['grade']  ?? 'F',
            'failed'     => $result['failed'] ?? 0,
            'issues'     => $result['issues'] ?? [],
            'scanned_at' => gmdate( 'c' ),
        ] );

        return $result;
    }

    // ── Migration ─────────────────────────────────────────────

    /**
     * One-time migration: reads the old single-blob `accessmate_page_scans` option
     * (if present) and redistributes each entry into the new storage layout.
     * Deletes the legacy option once done.
     */
    private static function maybe_migrate_legacy(): void {
        $legacy = get_option( self::LEGACY_KEY );
        if ( ! is_array( $legacy ) || empty( $legacy ) ) {
            return; // Already migrated or never existed.
        }

        foreach ( $legacy as $url => $data ) {
            if ( ! is_array( $data ) || empty( $url ) ) continue;
            self::store_axe_scan( (string) $url, $data );
        }

        delete_option( self::LEGACY_KEY );
    }
}