<?php
/**
 * Checker for 404 Scout. Pings unique URLs in parallel and handles loopbacks.
 */

class AILM_Scout_Checker
{

    public static function process_batch($batch_size = 20)
    {
        global $wpdb;
        $table_name = esc_sql(AILM_Scout_DB::get_table_name());
        $home_url = untrailingslashit(home_url());

        // 1. Get unique URLs that are still Pending (0)
        // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared, WordPress.DB.DirectDatabaseQuery.DirectQuery, WordPress.DB.DirectDatabaseQuery.NoCaching
        $unique_urls = $wpdb->get_col($wpdb->prepare(
            "SELECT DISTINCT url FROM {$table_name} WHERE status_code = 0 LIMIT %d", // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared
            $batch_size
        ));

        if (empty($unique_urls)) {
            return 0;
        }

        $requests = array();
        $results = array();

        // 2. Distribute checks (Internal vs External)
        foreach ($unique_urls as $url) {
            $clean_url = trim(html_entity_decode($url));

            // IF LINK IS INTERNAL (Including AILM Auto) -> Skip HTTP ping
            if (strpos($clean_url, $home_url) === 0) {
                $post_id = url_to_postid($clean_url);
                if ($post_id > 0 && get_post_status($post_id) === 'publish') {
                    $results[$url] = 200; // Post exists, set 200 instantly
                    continue;
                }
                // If url_to_postid didn't find it (e.g., category or tag link), add to HTTP ping queue
            }

            // Prepare request for external links and complex internal ones
            $requests[$url] = array(
                'url' => $clean_url,
                'type' => 'HEAD',
                'options' => array('timeout' => 5, 'redirection' => 5, 'verify' => false),
                'headers' => array(
                    // Disguise as a real browser to avoid 403 from Google and Cloudflare
                    'User-Agent' => 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
                    'Accept' => 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8'
                )
            );
        }

        // 3. Send HTTP requests in parallel
        if (!empty($requests)) {
            $responses = Requests::request_multiple($requests);

            foreach ($responses as $url => $response) {
                $status_code = 500;

                if (!is_wp_error($response) && !($response instanceof Exception) && !($response instanceof Requests_Exception)) {
                    $status_code = (int) $response->status_code;

                    // Fallback 1: If server rejected HEAD request (403, 405), try regular GET
                    if (in_array($status_code, [400, 403, 405, 406, 503])) {
                        $fallback = wp_remote_get($url, array(
                            'timeout' => 5,
                            'redirection' => 5,
                            'sslverify' => false,
                            'user-agent' => 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) Chrome/120.0.0.0'
                        ));
                        if (!is_wp_error($fallback)) {
                            $status_code = (int) wp_remote_retrieve_response_code($fallback);
                        }
                    }

                    // Fallback 2: If http:// returned an error, try substituting https://
                    if ($status_code >= 400 && strpos($url, 'http://') === 0) {
                        $https_url = str_replace('http://', 'https://', $url);
                        $https_fallback = wp_remote_head($https_url, array('timeout' => 5, 'user-agent' => 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) Chrome/120.0.0.0'));
                        if (!is_wp_error($https_fallback)) {
                            $https_code = (int) wp_remote_retrieve_response_code($https_fallback);
                            if ($https_code == 200 || $https_code == 301 || $https_code == 302) {
                                $status_code = 200; // Forgive if HTTPS version works
                            }
                        }
                    }
                }
                $results[$url] = $status_code;
            }
        }

        // 4. Bulk update the database
        $now = current_time('mysql');
        foreach ($results as $url => $status_code) {
            // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared, WordPress.DB.DirectDatabaseQuery.DirectQuery, WordPress.DB.DirectDatabaseQuery.NoCaching
            $wpdb->query($wpdb->prepare(
                "UPDATE {$table_name} SET status_code = %d, last_checked = %s WHERE url = %s AND status_code = 0", // phpcs:ignore WordPress.DB.PreparedSQL.InterpolatedNotPrepared
                $status_code,
                $now,
                $url
            ));
        }

        return count($unique_urls);
    }
}