<?php
/**
 * AI Crawler Analytics — Track AI bot visits.
 *
 * Logs which AI bots visit your site, how often, and which pages.
 * Uses WordPress options for lightweight storage (no custom tables).
 *
 * @since 3.0.0
 */

if ( ! defined( 'ABSPATH' ) ) {
    exit;
}

class AIREWP_Crawler_Analytics {

    private const OPTION_KEY  = 'airewp_crawler_stats';
    private const LOG_KEY     = 'airewp_crawler_log';
    private const MAX_LOG     = 200; // Keep last 200 entries

    /**
     * Regex patterns to match AI bot user agents.
     */
    private const BOT_PATTERNS = [
        'GPTBot'             => '/GPTBot/i',
        'OAI-SearchBot'      => '/OAI-SearchBot/i',
        'ChatGPT-User'       => '/ChatGPT-User/i',
        'ClaudeBot'          => '/ClaudeBot/i',
        'Claude-SearchBot'   => '/Claude-SearchBot/i',
        'Claude-User'        => '/Claude-User/i',
        'Google-Extended'    => '/Google-Extended/i',
        'GoogleOther'        => '/GoogleOther/i',
        'PerplexityBot'      => '/PerplexityBot/i',
        'Perplexity-User'    => '/Perplexity-User/i',
        'Meta-ExternalAgent' => '/Meta-ExternalAgent/i',
        'Bytespider'         => '/Bytespider/i',
        'CCBot'              => '/CCBot/i',
        'Applebot-Extended'  => '/Applebot-Extended/i',
        'Amazonbot'          => '/Amazonbot/i',
        'Diffbot'            => '/Diffbot/i',
        'Cohere-ai'          => '/cohere-ai/i',
        'FacebookBot'        => '/FacebookBot/i',
    ];

    public function __construct() {
        $settings = AI_Ready_WP::get_settings();
        if ( empty( $settings['enable_crawler_analytics'] ) ) {
            return;
        }

        // Track on frontend only, skip admin/cron
        if ( ! is_admin() && ! wp_doing_cron() && ! wp_doing_ajax() ) {
            add_action( 'wp', [ $this, 'track_visit' ] );
        }
    }

    /**
     * Track an AI bot visit.
     */
    public function track_visit(): void {
        $ua = isset( $_SERVER['HTTP_USER_AGENT'] ) ? sanitize_text_field( wp_unslash( $_SERVER['HTTP_USER_AGENT'] ) ) : '';
        if ( empty( $ua ) ) {
            return;
        }

        $bot = self::detect_bot( $ua );
        if ( ! $bot ) {
            return;
        }

        // Update aggregate stats
        $stats = get_option( self::OPTION_KEY, [] );
        if ( ! isset( $stats[ $bot ] ) ) {
            $stats[ $bot ] = [
                'visits'     => 0,
                'first_seen' => current_time( 'mysql' ),
                'last_seen'  => null,
            ];
        }
        $stats[ $bot ]['visits']++;
        $stats[ $bot ]['last_seen'] = current_time( 'mysql' );
        update_option( self::OPTION_KEY, $stats, false );

        // Append to recent log (capped)
        $log   = get_option( self::LOG_KEY, [] );
        $log[] = [
            'bot'  => $bot,
            'url'  => isset( $_SERVER['REQUEST_URI'] ) ? esc_url_raw( wp_unslash( $_SERVER['REQUEST_URI'] ) ) : '/',
            'time' => current_time( 'mysql' ),
        ];

        // Keep only the last MAX_LOG entries
        if ( count( $log ) > self::MAX_LOG ) {
            $log = array_slice( $log, -self::MAX_LOG );
        }
        update_option( self::LOG_KEY, $log, false );
    }

    /**
     * Detect which AI bot is visiting based on User-Agent.
     */
    public static function detect_bot( string $user_agent ): ?string {
        foreach ( self::BOT_PATTERNS as $name => $pattern ) {
            if ( preg_match( $pattern, $user_agent ) ) {
                return $name;
            }
        }
        return null;
    }

    /**
     * Get aggregate stats.
     */
    public static function get_stats(): array {
        return get_option( self::OPTION_KEY, [] );
    }

    /**
     * Get recent visit log.
     */
    public static function get_log( int $limit = 50 ): array {
        $log = get_option( self::LOG_KEY, [] );
        return array_slice( array_reverse( $log ), 0, $limit );
    }

    /**
     * Get total visits across all bots.
     */
    public static function get_total_visits(): int {
        $stats = self::get_stats();
        $total = 0;
        foreach ( $stats as $bot_stats ) {
            $total += $bot_stats['visits'] ?? 0;
        }
        return $total;
    }

    /**
     * Reset all analytics data.
     */
    public static function reset(): void {
        delete_option( self::OPTION_KEY );
        delete_option( self::LOG_KEY );
    }
}