<?php
/**
 * Known AI crawlers and agents.
 *
 * @package AgentAisle
 */

defined( 'ABSPATH' ) || exit;

class AgentAisle_Bots {

	/**
	 * Known AI bots keyed by user-agent token.
	 * type: training (collects data for model training), search (powers AI answers/search),
	 * agent (fetches on behalf of a user in real time).
	 *
	 * @return array
	 */
	public static function all() {
		$bots = array(
			'GPTBot'             => array( 'operator' => 'OpenAI', 'type' => 'training' ),
			'OAI-SearchBot'      => array( 'operator' => 'OpenAI', 'type' => 'search' ),
			'ChatGPT-User'       => array( 'operator' => 'OpenAI', 'type' => 'agent' ),
			'ClaudeBot'          => array( 'operator' => 'Anthropic', 'type' => 'training' ),
			'Claude-SearchBot'   => array( 'operator' => 'Anthropic', 'type' => 'search' ),
			'Claude-User'        => array( 'operator' => 'Anthropic', 'type' => 'agent' ),
			'PerplexityBot'      => array( 'operator' => 'Perplexity', 'type' => 'search' ),
			'Perplexity-User'    => array( 'operator' => 'Perplexity', 'type' => 'agent' ),
			'Google-Extended'    => array( 'operator' => 'Google', 'type' => 'training' ),
			'Applebot-Extended'  => array( 'operator' => 'Apple', 'type' => 'training' ),
			'Meta-ExternalAgent' => array( 'operator' => 'Meta', 'type' => 'training' ),
			'Meta-ExternalFetcher' => array( 'operator' => 'Meta', 'type' => 'agent' ),
			'Amazonbot'          => array( 'operator' => 'Amazon', 'type' => 'search' ),
			'Bytespider'         => array( 'operator' => 'ByteDance', 'type' => 'training' ),
			'CCBot'              => array( 'operator' => 'Common Crawl', 'type' => 'training' ),
			'MistralAI-User'     => array( 'operator' => 'Mistral', 'type' => 'agent' ),
			'DuckAssistBot'      => array( 'operator' => 'DuckDuckGo', 'type' => 'search' ),
			'cohere-ai'          => array( 'operator' => 'Cohere', 'type' => 'training' ),
		);

		/**
		 * Filter the list of known AI bots.
		 *
		 * @param array $bots Bots keyed by user-agent token.
		 */
		return apply_filters( 'agentaisle_bots', $bots );
	}

	/**
	 * Match a user agent string to a known bot.
	 *
	 * @param string $user_agent User agent.
	 * @return string|null Bot token or null.
	 */
	public static function match( $user_agent ) {
		if ( '' === $user_agent ) {
			return null;
		}
		foreach ( array_keys( self::all() ) as $token ) {
			if ( false !== stripos( $user_agent, $token ) ) {
				return $token;
			}
		}
		return null;
	}

	/**
	 * Human label for a bot type.
	 *
	 * @param string $type Type key.
	 * @return string
	 */
	public static function type_label( $type ) {
		$labels = array(
			'training' => __( 'Training', 'agentaisle' ),
			'search'   => __( 'AI search', 'agentaisle' ),
			'agent'    => __( 'User agent', 'agentaisle' ),
		);
		return $labels[ $type ] ?? $type;
	}
}