<?php
/**
 * Logs visits from known AI bots to a local table.
 *
 * @package AgentAisle
 */

defined( 'ABSPATH' ) || exit;

class AgentAisle_Logger {

	const DB_VERSION = '1';
	const CRON_HOOK  = 'agentaisle_purge_logs';

	/**
	 * Table name.
	 *
	 * @return string
	 */
	public static function table() {
		global $wpdb;
		return $wpdb->prefix . 'agentaisle_visits';
	}

	/**
	 * Hooks.
	 */
	public static function init() {
		add_action( 'template_redirect', array( __CLASS__, 'maybe_log' ), 1 );
		add_action( 'do_robots', array( __CLASS__, 'maybe_log' ), 1 );
		add_action( self::CRON_HOOK, array( __CLASS__, 'purge' ) );
		if ( get_option( 'agentaisle_db_version' ) !== self::DB_VERSION ) {
			self::create_table();
		}
	}

	/**
	 * Activation: create table and schedule purge.
	 */
	public static function activate() {
		self::create_table();
		if ( ! wp_next_scheduled( self::CRON_HOOK ) ) {
			wp_schedule_event( time() + HOUR_IN_SECONDS, 'daily', self::CRON_HOOK );
		}
	}

	/**
	 * Deactivation: clear cron.
	 */
	public static function deactivate() {
		wp_clear_scheduled_hook( self::CRON_HOOK );
	}

	/**
	 * Create or update the visits table.
	 */
	public static function create_table() {
		global $wpdb;
		require_once ABSPATH . 'wp-admin/includes/upgrade.php';
		$table   = self::table();
		$charset = $wpdb->get_charset_collate();
		$sql     = "CREATE TABLE {$table} (
			id bigint(20) unsigned NOT NULL AUTO_INCREMENT,
			bot varchar(64) NOT NULL,
			path varchar(255) NOT NULL,
			object_id bigint(20) unsigned NOT NULL DEFAULT 0,
			status smallint(5) unsigned NOT NULL DEFAULT 200,
			ip_hash char(16) NOT NULL DEFAULT '',
			visited_at datetime NOT NULL,
			PRIMARY KEY  (id),
			KEY bot_time (bot, visited_at),
			KEY visited_at (visited_at)
		) {$charset};";
		dbDelta( $sql );
		update_option( 'agentaisle_db_version', self::DB_VERSION, false );
	}

	/**
	 * Log the current request if it comes from a known AI bot.
	 */
	public static function maybe_log() {
		static $done = false;
		if ( $done ) {
			return;
		}
		$done = true;

		$settings = agentaisle_settings();
		if ( empty( $settings['logging'] ) ) {
			return;
		}

		$ua  = isset( $_SERVER['HTTP_USER_AGENT'] ) ? sanitize_text_field( wp_unslash( $_SERVER['HTTP_USER_AGENT'] ) ) : '';
		$bot = AgentAisle_Bots::match( $ua );
		if ( ! $bot ) {
			return;
		}

		$path = isset( $_SERVER['REQUEST_URI'] ) ? esc_url_raw( wp_unslash( $_SERVER['REQUEST_URI'] ) ) : '/';
		$path = substr( (string) wp_parse_url( $path, PHP_URL_PATH ), 0, 255 );

		$object_id = 0;
		if ( function_exists( 'is_singular' ) && is_singular() ) {
			$object_id = (int) get_queried_object_id();
		}

		// Anonymised: a truncated salted hash, never the raw IP.
		$ip      = isset( $_SERVER['REMOTE_ADDR'] ) ? sanitize_text_field( wp_unslash( $_SERVER['REMOTE_ADDR'] ) ) : '';
		$ip_hash = $ip ? substr( hash( 'sha256', $ip . wp_salt( 'auth' ) ), 0, 16 ) : '';

		global $wpdb;
		$wpdb->insert( // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery -- custom table, write-only log.
			self::table(),
			array(
				'bot'        => $bot,
				'path'       => $path ? $path : '/',
				'object_id'  => $object_id,
				'status'     => is_404() ? 404 : 200,
				'ip_hash'    => $ip_hash,
				'visited_at' => current_time( 'mysql', true ),
			),
			array( '%s', '%s', '%d', '%d', '%s', '%s' )
		);

		/**
		 * Fires after a known AI bot visit is logged.
		 *
		 * @param string $bot       Bot token.
		 * @param string $path      Request path.
		 * @param int    $object_id Queried object ID or 0.
		 */
		do_action( 'agentaisle_bot_visit', $bot, $path, $object_id );
	}

	/**
	 * Delete logs older than the retention period.
	 */
	public static function purge() {
		global $wpdb;
		$days = max( 1, (int) agentaisle_settings()['retention_days'] );
		$wpdb->query( $wpdb->prepare( 'DELETE FROM %i WHERE visited_at < %s', self::table(), self::since( $days ) ) ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery, WordPress.DB.DirectDatabaseQuery.NoCaching -- custom table, scheduled cleanup.
	}

	/**
	 * Visits per bot over the last N days.
	 *
	 * @param int $days Days.
	 * @return array
	 */
	public static function counts_by_bot( $days ) {
		global $wpdb;
		$key  = 'bots_' . (int) $days;
		$rows = wp_cache_get( $key, 'agentaisle' );
		if ( false === $rows ) {
			$rows = $wpdb->get_results( $wpdb->prepare( 'SELECT bot, COUNT(*) AS visits, MAX(visited_at) AS last_seen FROM %i WHERE visited_at >= %s GROUP BY bot ORDER BY visits DESC', self::table(), self::since( $days ) ) ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery -- custom table, result cached.
			wp_cache_set( $key, $rows, 'agentaisle', 5 * MINUTE_IN_SECONDS );
		}
		return $rows;
	}

	/**
	 * Most-crawled paths.
	 *
	 * @param int $days  Days.
	 * @param int $limit Limit.
	 * @return array
	 */
	public static function top_paths( $days, $limit = 10 ) {
		global $wpdb;
		$key  = 'paths_' . (int) $days . '_' . (int) $limit;
		$rows = wp_cache_get( $key, 'agentaisle' );
		if ( false === $rows ) {
			$rows = $wpdb->get_results( $wpdb->prepare( 'SELECT path, object_id, COUNT(*) AS visits FROM %i WHERE visited_at >= %s GROUP BY path, object_id ORDER BY visits DESC LIMIT %d', self::table(), self::since( $days ), (int) $limit ) ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery -- custom table, result cached.
			wp_cache_set( $key, $rows, 'agentaisle', 5 * MINUTE_IN_SECONDS );
		}
		return $rows;
	}

	/**
	 * Daily totals.
	 *
	 * @param int $days Days.
	 * @return array date => visits
	 */
	public static function daily( $days ) {
		global $wpdb;
		$key  = 'daily_' . (int) $days;
		$rows = wp_cache_get( $key, 'agentaisle' );
		if ( false === $rows ) {
			$rows = $wpdb->get_results( $wpdb->prepare( 'SELECT DATE(visited_at) AS d, COUNT(*) AS visits FROM %i WHERE visited_at >= %s GROUP BY d', self::table(), self::since( $days ) ) ); // phpcs:ignore WordPress.DB.DirectDatabaseQuery.DirectQuery -- custom table, result cached.
			wp_cache_set( $key, $rows, 'agentaisle', 5 * MINUTE_IN_SECONDS );
		}
		$out = array();
		for ( $i = $days - 1; $i >= 0; $i-- ) {
			$out[ gmdate( 'Y-m-d', time() - $i * DAY_IN_SECONDS ) ] = 0;
		}
		foreach ( (array) $rows as $row ) {
			if ( isset( $out[ $row->d ] ) ) {
				$out[ $row->d ] = (int) $row->visits;
			}
		}
		return $out;
	}

	/**
	 * Cutoff datetime string.
	 *
	 * @param int $days Days.
	 * @return string
	 */
	private static function since( $days ) {
		return gmdate( 'Y-m-d H:i:s', time() - (int) $days * DAY_IN_SECONDS );
	}
}