<?php
/**
 * Crawler logger — logs AI bot access to discovery files served by AIDF_Server.
 *
 * Called directly from AIDF_Server::serve_file() after a file is generated
 * successfully. This approach guarantees every request reaches PHP even on
 * sites with page caching or CDN edge caching, because discovery file
 * responses use s-maxage=0.
 *
 * @package AIDF
 * @since   1.3.0
 */

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

/**
 * Handles AI crawler detection and log persistence for discovery file requests.
 *
 * All public methods are static so the class can be called without an instance.
 *
 * @since 1.3.0
 */
class AIDF_Crawler_Logger {

	/**
	 * Detect and log an AI crawler hit for a discovery file request.
	 *
	 * Called from AIDF_Server::serve_file() after the file content has been
	 * generated successfully. Checks the user agent against the enabled bot
	 * registry and writes a single row to the crawler log table when matched.
	 *
	 * Reads the raw aidf_settings option directly instead of
	 * AIDF_Plugin::get_settings() to avoid triggering page URL lookups
	 * that require $wp_rewrite — which may not be fully initialised at
	 * serve_file() time.
	 *
	 * @since 1.3.0
	 *
	 * @param string $file_slug The file type slug (e.g. 'llms-txt', 'ai-txt').
	 * @return void
	 */
	public static function maybe_log_hit( $file_slug ) {
		// Read the raw option to avoid the $wp_rewrite timing issue.
		$saved = get_option( 'aidf_settings', array() );

		if ( empty( $saved['crawler_logging_enabled'] ) ) {
			return;
		}

		$ua = isset( $_SERVER['HTTP_USER_AGENT'] )
			? sanitize_text_field( wp_unslash( $_SERVER['HTTP_USER_AGENT'] ) )
			: '';

		if ( empty( $ua ) ) {
			return;
		}

		// Build the enabled-only bot subset from the registry.
		$enabled_keys = isset( $saved['crawler_enabled_bots'] ) ? (array) $saved['crawler_enabled_bots'] : array();
		$all_bots     = AIDF_Crawler_Registry::get_bots();

		$enabled_bots = array();
		foreach ( $enabled_keys as $key ) {
			if ( isset( $all_bots[ $key ] ) ) {
				$enabled_bots[ $key ] = $all_bots[ $key ];
			}
		}

		$matched_key = AIDF_Crawler_Registry::match_user_agent( $ua, $enabled_bots );

		if ( false === $matched_key ) {
			return;
		}

		// Build the URL path from the known file type metadata.
		$file_types = AIDF_Plugin::get_file_types();
		$url_path   = isset( $file_types[ $file_slug ] )
			? '/' . $file_types[ $file_slug ]['filename']
			: '/' . $file_slug;

		global $wpdb;

		$table       = $wpdb->prefix . 'aidf_crawler_log';
		$status_code = 200;
		$now         = current_time( 'mysql', true );

		$wpdb->insert( // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery, WordPress.DB.DirectDatabaseQuery.NoCaching
			$table,
			array(
				'bot_name'    => $all_bots[ $matched_key ]['name'],
				'bot_ua'      => mb_substr( $ua, 0, 500 ),
				'url_path'    => mb_substr( $url_path, 0, 500 ),
				'status_code' => $status_code,
				'created_at'  => $now,
			),
			array( '%s', '%s', '%s', '%d', '%s' )
		);

		/**
		 * Fires after a crawler hit has been written to the log.
		 *
		 * @since 1.3.0
		 *
		 * @param string $bot_name    Human-readable bot name.
		 * @param string $url_path    Requested URL path (e.g. /llms.txt).
		 * @param int    $status_code HTTP response status code (always 200).
		 */
		do_action( 'aidf_crawler_hit_logged', $all_bots[ $matched_key ]['name'], $url_path, $status_code );
	}

	/**
	 * Schedule the daily cron events used by the crawler analytics feature.
	 *
	 * Should be called when crawler logging is enabled (e.g. on settings
	 * save). Uses `wp_next_scheduled` guards so calling it multiple times
	 * is safe.
	 *
	 * @since 1.3.0
	 * @return void
	 */
	public static function schedule_cron() {
		if ( ! wp_next_scheduled( 'aidf_rebuild_crawler_summary' ) ) {
			wp_schedule_event( time(), 'daily', 'aidf_rebuild_crawler_summary' );
		}

		if ( ! wp_next_scheduled( 'aidf_prune_crawler_log' ) ) {
			wp_schedule_event( time(), 'daily', 'aidf_prune_crawler_log' );
		}
	}

	/**
	 * Remove the daily cron events registered by this logger.
	 *
	 * Called on plugin deactivation or when logging is disabled in settings
	 * so no orphaned cron events remain in the scheduler.
	 *
	 * @since 1.3.0
	 * @return void
	 */
	public static function clear_cron() {
		wp_clear_scheduled_hook( 'aidf_rebuild_crawler_summary' );
		wp_clear_scheduled_hook( 'aidf_prune_crawler_log' );
	}

	/**
	 * Delete log rows that fall outside the configured retention window.
	 *
	 * Called daily via the `aidf_prune_crawler_log` cron event. Removes
	 * entries from both the raw log table and the pre-aggregated summary
	 * table that are older than the `crawler_log_retention` setting (days).
	 * Defaults to 90 days if the setting is below 30.
	 *
	 * @since 1.3.0
	 * @return void
	 */
	public static function prune_log() {
		global $wpdb;

		$settings  = AIDF_Plugin::get_settings();
		$retention = absint( $settings['crawler_log_retention'] );

		if ( $retention < 30 ) {
			$retention = 30;
		}

		$cutoff = gmdate( 'Y-m-d H:i:s', strtotime( "-{$retention} days" ) );

		$wpdb->query( // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery, WordPress.DB.DirectDatabaseQuery.NoCaching
			$wpdb->prepare(
				"DELETE FROM {$wpdb->prefix}aidf_crawler_log WHERE created_at < %s",
				$cutoff
			)
		);

		$wpdb->query( // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery, WordPress.DB.DirectDatabaseQuery.NoCaching
			$wpdb->prepare(
				"DELETE FROM {$wpdb->prefix}aidf_crawler_summary WHERE log_date < %s",
				gmdate( 'Y-m-d', strtotime( "-{$retention} days" ) )
			)
		);
	}
}