<?php
/**
 * LLMs.txt endpoint generator.
 *
 * Serves a /llms.txt endpoint via WordPress rewrite rules that provides
 * structured information about the store for LLM consumption.
 *
 * @package AICommerce\Crawler
 */

declare(strict_types=1);

namespace AICommerce\Crawler;

/**
 * Generates and serves the /llms.txt endpoint.
 */
class LlmsTxtGenerator {

	/**
	 * Register the rewrite rule for /llms.txt.
	 *
	 * @return void
	 */
	public function register_rewrite_rule(): void {
		add_rewrite_rule( '^llms\.txt$', 'index.php?aicommerce_llms_txt=1', 'top' );
	}

	/**
	 * Add the query var for the llms.txt endpoint.
	 *
	 * @param array $vars Existing query vars.
	 *
	 * @return array Modified query vars.
	 */
	public function add_query_var( array $vars ): array {
		$vars[] = 'aicommerce_llms_txt';
		return $vars;
	}

	/**
	 * Serve the llms.txt content if requested.
	 *
	 * @return void
	 */
	public function maybe_serve(): void {
		if ( ! get_query_var( 'aicommerce_llms_txt' ) ) {
			return;
		}

		header( 'Content-Type: text/plain; charset=utf-8' );
		header( 'Cache-Control: public, max-age=3600' );
		header( 'X-Robots-Tag: noindex' );

		echo $this->build_content(); // phpcs:ignore WordPress.Security.EscapeOutput.OutputNotEscaped -- Plain text output.
		exit;
	}

	/**
	 * Build the llms.txt content.
	 *
	 * @return string The llms.txt content.
	 */
	public function build_content(): string {
		$store_name = get_bloginfo( 'name' );
		$store_url  = home_url();
		$store_desc = get_bloginfo( 'description' );

		$lines = array();

		// Header.
		$lines[] = '# ' . $store_name;
		$lines[] = '';
		$lines[] = '> ' . ( ! empty( $store_desc ) ? $store_desc : $store_name . ' — an online store.' );
		$lines[] = '';

		// Store info.
		$lines[] = '## Store Information';
		$lines[] = '';
		$lines[] = '- Website: ' . $store_url;

		if ( function_exists( 'get_woocommerce_currency' ) ) {
			$lines[] = '- Currency: ' . get_woocommerce_currency();
		}

		if ( post_type_exists( 'product' ) ) {
			$product_count = wp_count_posts( 'product' );
			if ( $product_count && isset( $product_count->publish ) ) {
				$lines[] = '- Products: ' . (int) $product_count->publish;
			}
		}

		$lines[] = '';

		// Feeds.
		$lines[] = '## Product Feeds';
		$lines[] = '';

		$feed_types = array(
			'google'     => array(
				'name' => 'Google Shopping Feed',
				'file' => 'google-feed.xml',
			),
			'openai'     => array(
				'name' => 'OpenAI Product Feed',
				'file' => 'openai-feed.json',
			),
			'bing'       => array(
				'name' => 'Bing Shopping Feed',
				'file' => 'bing-feed.xml',
			),
			'perplexity' => array(
				'name' => 'Perplexity Product Feed',
				'file' => 'perplexity-feed.json',
			),
		);

		$upload_dir = wp_upload_dir();
		$feed_dir   = $upload_dir['basedir'] . '/aicommerce/feeds/';

		foreach ( $feed_types as $type => $info ) {
			$file_path = $feed_dir . $info['file'];
			if ( file_exists( $file_path ) ) {
				$feed_url = $upload_dir['baseurl'] . '/aicommerce/feeds/' . $info['file'];
				$lines[]  = '- [' . $info['name'] . '](' . $feed_url . ')';
			}
		}

		$lines[] = '';

		// API discovery.
		$lines[] = '## API Discovery';
		$lines[] = '';
		$lines[] = '- [AI Commerce Manifest](' . home_url( '/.well-known/ai-commerce.json' ) . ')';
		$lines[] = '- [MCP Server](' . rest_url( 'aicommerce/v1/mcp' ) . ')';
		$lines[] = '- [MCP Discovery](' . home_url( '/.well-known/mcp.json' ) . ')';

		$sitemap_url = home_url( '/wp-sitemap.xml' );
		$lines[]     = '- [Sitemap](' . $sitemap_url . ')';

		$lines[] = '';

		// Top categories — helps LLMs discover the catalog structure.
		$category_lines = $this->build_category_lines();
		if ( ! empty( $category_lines ) ) {
			$lines[] = '## Top Categories';
			$lines[] = '';
			foreach ( $category_lines as $line ) {
				$lines[] = $line;
			}
			$lines[] = '';
		}

		// Crawling policy — reflects the store's actual configuration.
		$lines[] = '## Crawling Policy';
		$lines[] = '';
		$lines[] = 'This store welcomes AI agents and LLM crawlers. Product data is';
		$lines[] = 'available via structured feeds and schema markup on product pages.';
		$lines[] = '';

		$blocked = get_option( 'aicommerce_blocked_bots', array() );
		$blocked = is_array( $blocked ) ? array_values( array_filter( array_map( 'strval', $blocked ) ) ) : array();

		if ( ! empty( $blocked ) ) {
			$lines[] = 'Disallowed AI crawler groups: ' . implode( ', ', $blocked );
		} else {
			$lines[] = 'All major AI crawlers are permitted, subject to robots.txt.';
		}
		$lines[] = '';
		$lines[] = 'See ' . home_url( '/robots.txt' ) . ' for per-crawler rules.';
		$lines[] = '';
		$lines[] = 'For automated purchasing, see the API discovery links above.';
		$lines[] = '';

		// Contact — never exposes the WordPress admin email automatically.
		$contact_email = sanitize_email( (string) get_option( 'aicommerce_llms_contact_email', '' ) );

		/**
		 * Filters the public contact email published in /llms.txt.
		 *
		 * Empty by default so the site admin email is never leaked. Set the
		 * aicommerce_llms_contact_email option or this filter to publish one.
		 *
		 * @param string $contact_email The contact email (may be empty).
		 */
		$contact_email = (string) apply_filters( 'aicommerce_llms_contact_email', $contact_email );

		if ( ! empty( $contact_email ) && is_email( $contact_email ) ) {
			$lines[] = '## Contact';
			$lines[] = '';
			$lines[] = '- Email: ' . $contact_email;
			$lines[] = '';
		}

		$content = implode( "\n", $lines );

		/**
		 * Filters the complete /llms.txt content before output.
		 *
		 * @param string $content The generated llms.txt content.
		 */
		return (string) apply_filters( 'aicommerce_llms_txt_content', $content );
	}

	/**
	 * Build markdown links for the top product categories.
	 *
	 * @param int $limit Maximum number of categories to include.
	 *
	 * @return array<int, string> Markdown list lines.
	 */
	private function build_category_lines( int $limit = 20 ): array {
		if ( ! taxonomy_exists( 'product_cat' ) ) {
			return array();
		}

		$terms = get_terms(
			array(
				'taxonomy'   => 'product_cat',
				'hide_empty' => true,
				'number'     => $limit,
				'orderby'    => 'count',
				'order'      => 'DESC',
			)
		);

		if ( is_wp_error( $terms ) || empty( $terms ) ) {
			return array();
		}

		$lines = array();
		foreach ( $terms as $term ) {
			$link = get_term_link( $term );
			if ( is_wp_error( $link ) ) {
				continue;
			}
			$lines[] = '- [' . $term->name . '](' . $link . ') — ' . (int) $term->count . ' products';
		}

		return $lines;
	}
}