<?php
/**
 * Plugin Name:       AI Crawler Guard
 * Description:       Detect and control AI crawler access with local logging and history.
 * Version:           1.0.0
 * Author:            ServerlessGeeks.com
 * License:           GPLv2 or later
 * License URI:       https://www.gnu.org/licenses/gpl-2.0.html
 * Requires at least: 6.0
 * Requires PHP:      7.4
 * Text Domain:       ai-crawler-guard
 */

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

if ( ! defined( 'AICRG_VERSION' ) ) {
    define( 'AICRG_VERSION', '1.0.0' );
}

if ( ! defined( 'AICRG_PLUGIN_FILE' ) ) {
    define( 'AICRG_PLUGIN_FILE', __FILE__ );
}

if ( ! defined( 'AICRG_PLUGIN_DIR' ) ) {
    define( 'AICRG_PLUGIN_DIR', plugin_dir_path( __FILE__ ) );
}

global $wpdb;

if ( ! defined( 'AICRG_TABLE_CRAWLERS' ) ) {
    define( 'AICRG_TABLE_CRAWLERS', $wpdb->prefix . 'aicrg_crawlers' );
}

if ( ! defined( 'AICRG_TABLE_LOGS' ) ) {
    define( 'AICRG_TABLE_LOGS', $wpdb->prefix . 'aicrg_logs' );
}

require_once AICRG_PLUGIN_DIR . 'includes/functions-helpers.php';
require_once AICRG_PLUGIN_DIR . 'includes/class-rules-provider.php';
require_once AICRG_PLUGIN_DIR . 'includes/class-detector.php';
require_once AICRG_PLUGIN_DIR . 'includes/class-crawler-registry.php';
require_once AICRG_PLUGIN_DIR . 'includes/class-policy.php';
require_once AICRG_PLUGIN_DIR . 'includes/class-logger.php';
require_once AICRG_PLUGIN_DIR . 'includes/class-admin.php';

register_activation_hook( __FILE__, 'aicrg_activate' );
register_deactivation_hook( __FILE__, 'aicrg_deactivate' );

// phpcs:ignore WordPress.NamingConventions.PrefixAllGlobals.NonPrefixedFunctionFound -- Function is prefixed with 'aicrg_'
function aicrg_activate(): void {
    global $wpdb;

    require_once ABSPATH . 'wp-admin/includes/upgrade.php';

    $charset_collate = $wpdb->get_charset_collate();

    $crawl_table = $wpdb->prefix . 'aicrg_crawlers';
    $logs_table  = $wpdb->prefix . 'aicrg_logs';

    $sql_crawlers = "CREATE TABLE {$crawl_table} (
        id BIGINT UNSIGNED NOT NULL AUTO_INCREMENT,
        bot_slug VARCHAR(100) NOT NULL,
        display_name VARCHAR(255) NOT NULL,
        status VARCHAR(20) NOT NULL DEFAULT 'unknown',
        first_seen DATETIME NOT NULL,
        last_seen DATETIME NOT NULL,
        total_hits INT NOT NULL DEFAULT 0,
        allowed_hits INT NOT NULL DEFAULT 0,
        blocked_hits INT NOT NULL DEFAULT 0,
        PRIMARY KEY  (id),
        UNIQUE KEY bot_slug (bot_slug)
    ) {$charset_collate};";

    $sql_logs = "CREATE TABLE {$logs_table} (
        id BIGINT UNSIGNED NOT NULL AUTO_INCREMENT,
        time DATETIME NOT NULL,
        bot_slug VARCHAR(100) NOT NULL,
        bot_name VARCHAR(255) NOT NULL,
        user_agent TEXT,
        url TEXT,
        action VARCHAR(20) NOT NULL,
        PRIMARY KEY  (id),
        KEY time (time)
    ) {$charset_collate};";

    dbDelta( $sql_crawlers );
    dbDelta( $sql_logs );

    if ( false === get_option( 'aicrg_settings', false ) ) {
        add_option( 'aicrg_settings', aicrg_get_default_settings() );
    }
}

// phpcs:ignore WordPress.NamingConventions.PrefixAllGlobals.NonPrefixedFunctionFound -- Function is prefixed with 'aicrg_'
function aicrg_deactivate(): void {
    // Intentionally left blank for future cleanup.
}

// phpcs:ignore WordPress.NamingConventions.PrefixAllGlobals.NonPrefixedFunctionFound -- Function is prefixed with 'aicrg_'
function aicrg_services(): array {
    static $services = null;

    if ( null === $services ) {
        global $wpdb;

        $rules_provider = new AICRG_Rules_Provider();
        $detector       = new AICRG_Detector( $rules_provider );
        $registry       = new AICRG_Crawler_Registry( $wpdb, AICRG_TABLE_CRAWLERS );
        $logger         = new AICRG_Logger( $wpdb, AICRG_TABLE_LOGS );
        $policy         = new AICRG_Policy( $registry, $rules_provider );

        if ( is_admin() ) {
            new AICRG_Admin( $registry, $logger, AICRG_TABLE_LOGS, $rules_provider );
        }

        $services = compact( 'rules_provider', 'detector', 'registry', 'logger', 'policy' );
    }

    return $services;
}

add_action( 'plugins_loaded', 'aicrg_services' );
add_action( 'template_redirect', 'aicrg_maybe_block_crawler', 0 );

// phpcs:ignore WordPress.NamingConventions.PrefixAllGlobals.NonPrefixedFunctionFound -- Function is prefixed with 'aicrg_'
function aicrg_maybe_block_crawler(): void {
    static $processed = false;

    if ( $processed ) {
        return;
    }

    if (
        is_admin()
        || wp_doing_cron()
        || ( defined( 'WP_CLI' ) && WP_CLI )
        || ( defined( 'REST_REQUEST' ) && REST_REQUEST )
    ) {
        return;
    }

    $processed = true;

    $services = aicrg_services();

    // phpcs:ignore WordPress.Security.ValidatedSanitizedInput.MissingUnslash -- User-Agent is sanitized below
    $user_agent = sanitize_text_field( wp_unslash( $_SERVER['HTTP_USER_AGENT'] ?? '' ) );
    $detection  = $services['detector']->detect( $user_agent );

    // Only process if a bot was detected - regular users are not tracked
    if ( empty( $detection['is_bot'] ) ) {
        return;
    }

    $settings        = aicrg_get_settings();
    $history_enabled = ! empty( $settings['history_enabled'] );

    $bot_slug   = $detection['bot_slug'];
    $bot_name   = $detection['bot_name'];
    $rule_info  = $services['rules_provider']->get_category_info( $bot_slug );
    $default_status = $rule_info['default_status'] ?? 'unknown';

    if ( $history_enabled ) {
        $services['registry']->get_or_create( $bot_slug, $bot_name, $default_status );
    }

    $decision = $services['policy']->decide( $bot_slug );
    $action   = 'block' === $decision ? 'blocked' : 'allowed';

    $protocol = ( isset( $_SERVER['HTTPS'] ) && 'on' === $_SERVER['HTTPS'] ) ? 'https://' : 'http://';
    // phpcs:ignore WordPress.Security.ValidatedSanitizedInput.MissingUnslash -- HTTP_HOST is sanitized below
    $host     = sanitize_text_field( wp_unslash( $_SERVER['HTTP_HOST'] ?? '' ) );
    // phpcs:ignore WordPress.Security.ValidatedSanitizedInput.MissingUnslash -- REQUEST_URI is sanitized below
    $uri      = sanitize_text_field( wp_unslash( $_SERVER['REQUEST_URI'] ?? '' ) );
    $url      = $protocol . $host . $uri;

            $services['logger']->log( $bot_slug, $bot_name, $user_agent, $url, $action );

    if ( $history_enabled ) {
        $services['registry']->update_stats( $bot_slug, $action );
    }

    if ( 'block' === $decision ) {
        status_header( 403 );
        nocache_headers();
        echo esc_html__( 'Access denied for automated crawlers.', 'ai-crawler-guard' );
        exit;
    }
}