<?php
/**
 * AI Content Signals - Robots.txt Handler
 *
 * Manages both physical and virtual robots.txt files.
 *
 * @package AIContentSignals
 * @since   1.0.0
 */

// Exit if accessed directly.
if ( ! defined( 'ABSPATH' ) ) {
	exit;
}

/**
 * Add Content Signals to virtual robots.txt
 *
 * @param string $output Current robots.txt output.
 * @param string $public Blog visibility setting.
 * @return string Modified robots.txt output.
 */
function ayudawp_acs_modify_robots( $output, $public ) {
	if ( '0' === $public ) {
		return $output;
	}

	$content_signals = ayudawp_acs_generate_content_signals();

	if ( ! empty( $content_signals ) ) {
		$output .= "\n" . $content_signals;
	}

	return $output;
}
add_filter( 'robots_txt', 'ayudawp_acs_modify_robots', 999, 2 );

/**
 * Initialize WP_Filesystem
 *
 * @return bool True if filesystem is available.
 */
function ayudawp_acs_init_filesystem() {
	global $wp_filesystem;

	if ( empty( $wp_filesystem ) ) {
		require_once ABSPATH . 'wp-admin/includes/file.php';
		WP_Filesystem();
	}

	return ! empty( $wp_filesystem );
}

/**
 * Generate Content Signals text
 *
 * @return string Generated Content Signals block for robots.txt.
 */
function ayudawp_acs_generate_content_signals() {
	$settings = get_option( 'ayudawp_acs_settings', array() );
	$output   = "\n# BEGIN AI Content Signals\n";
	$output  .= '# Generated on ' . gmdate( 'Y-m-d H:i:s' ) . " UTC\n\n";

	// Add legal text if enabled.
	if ( ( $settings['include_legal_text'] ?? 'yes' ) === 'yes' ) {
		$output .= ayudawp_acs_get_legal_text();
	}

	// Global settings.
	$signals = array();
	if ( ! empty( $settings['global_search'] ) ) {
		$signals[] = 'search=' . $settings['global_search'];
	}
	if ( ! empty( $settings['global_ai_input'] ) ) {
		$signals[] = 'ai-input=' . $settings['global_ai_input'];
	}
	if ( ! empty( $settings['global_ai_train'] ) ) {
		$signals[] = 'ai-train=' . $settings['global_ai_train'];
	}

	if ( ! empty( $signals ) ) {
		$output .= "User-Agent: *\n";
		$output .= 'Content-Signal: ' . implode( ', ', $signals ) . "\n";
		$output .= "Allow: /\n\n";
	}

	// Bot-specific settings.
	$bot_settings = $settings['bot_settings'] ?? array();

	foreach ( $bot_settings as $bot => $bot_data ) {
		$bot_signals = array();

		if ( ! empty( $bot_data['search'] ) ) {
			$bot_signals[] = 'search=' . $bot_data['search'];
		}
		if ( ! empty( $bot_data['ai_input'] ) ) {
			$bot_signals[] = 'ai-input=' . $bot_data['ai_input'];
		}
		if ( ! empty( $bot_data['ai_train'] ) ) {
			$bot_signals[] = 'ai-train=' . $bot_data['ai_train'];
		}

		if ( ! empty( $bot_signals ) ) {
			$output .= 'User-Agent: ' . $bot . "\n";
			$output .= 'Content-Signal: ' . implode( ', ', $bot_signals ) . "\n";
			$output .= "Allow: /\n\n";
		}
	}

	// Custom bots.
	$custom_bots = $settings['custom_bots'] ?? array();
	foreach ( $custom_bots as $bot ) {
		if ( empty( $bot['name'] ) ) {
			continue;
		}

		$custom_signals = array();

		// Use specific setting if set, otherwise use global.
		$search_value   = ! empty( $bot['search'] ) ? $bot['search'] : ( $settings['global_search'] ?? 'yes' );
		$ai_input_value = ! empty( $bot['ai_input'] ) ? $bot['ai_input'] : ( $settings['global_ai_input'] ?? 'yes' );
		$ai_train_value = ! empty( $bot['ai_train'] ) ? $bot['ai_train'] : ( $settings['global_ai_train'] ?? 'yes' );

		if ( ! empty( $search_value ) ) {
			$custom_signals[] = 'search=' . $search_value;
		}
		if ( ! empty( $ai_input_value ) ) {
			$custom_signals[] = 'ai-input=' . $ai_input_value;
		}
		if ( ! empty( $ai_train_value ) ) {
			$custom_signals[] = 'ai-train=' . $ai_train_value;
		}

		if ( ! empty( $custom_signals ) ) {
			$output .= 'User-Agent: ' . $bot['name'] . "\n";
			$output .= 'Content-Signal: ' . implode( ', ', $custom_signals ) . "\n";
			$output .= "Allow: /\n\n";
		}
	}

	return $output . "# END AI Content Signals\n";
}

/**
 * Get legal text for Content Signals Policy
 *
 * @return string Legal text block as robots.txt comments.
 */
function ayudawp_acs_get_legal_text() {
	$text  = "# As a condition of accessing this website, you agree to abide by the following content signals:\n";
	$text .= "# (a) If a content-signal = yes, you may collect content for the corresponding use.\n";
	$text .= "# (b) If a content-signal = no, you may not collect content for the corresponding use.\n";
	$text .= "# (c) If the website operator does not include a content signal for a corresponding use,\n";
	$text .= "#     the website operator neither grants nor restricts permission via content signal\n";
	$text .= "#     with respect to the corresponding use.\n";
	$text .= "#\n";
	$text .= "# The content signals and their meanings are:\n";
	$text .= "# search: building a search index and providing search results\n";
	$text .= "#         (e.g., returning hyperlinks and short excerpts from your website's contents).\n";
	$text .= "#         Search does not include providing AI-generated search summaries.\n";
	$text .= "# ai-input: inputting content into one or more AI models\n";
	$text .= "#          (e.g., retrieval augmented generation, grounding, or other real-time\n";
	$text .= "#          taking of content for generative AI search answers).\n";
	$text .= "# ai-train: training or fine-tuning AI models.\n";
	$text .= "#\n";
	$text .= "# ANY RESTRICTIONS EXPRESSED VIA CONTENT SIGNALS ARE EXPRESS RESERVATIONS OF RIGHTS\n";
	$text .= "# UNDER ARTICLE 4 OF THE EUROPEAN UNION DIRECTIVE 2019/790 ON COPYRIGHT AND RELATED\n";
	$text .= "# RIGHTS IN THE DIGITAL SINGLE MARKET.\n\n";

	return $text;
}

/**
 * Create physical robots.txt with basic WordPress rules
 *
 * @return bool True on success, false on failure.
 */
function ayudawp_acs_create_physical_robots() {
	if ( ! ayudawp_acs_init_filesystem() ) {
		return false;
	}

	global $wp_filesystem;
	$robots_file = ABSPATH . 'robots.txt';

	// Don't create if already exists.
	if ( $wp_filesystem->exists( $robots_file ) ) {
		return false;
	}

	// Check if root directory is writable.
	if ( ! $wp_filesystem->is_writable( ABSPATH ) ) {
		return false;
	}

	// Basic WordPress robots.txt content.
	$content  = "# Basic WordPress robots.txt\n";
	$content .= "# Created by AI Content Signals\n\n";
	$content .= "User-agent: *\n";
	$content .= "Disallow: /wp-admin/\n";
	$content .= "Allow: /wp-admin/admin-ajax.php\n";
	$content .= "Disallow: /wp-includes/\n";
	$content .= "Disallow: /wp-content/plugins/\n";
	$content .= "Disallow: /wp-content/themes/\n";
	$content .= "Disallow: /wp-content/cache/\n";
	$content .= "Disallow: /readme.html\n";
	$content .= "Disallow: /license.txt\n\n";

	// Add sitemap if available.
	$sitemap_url = get_sitemap_url( 'index' );
	if ( $sitemap_url ) {
		$content .= 'Sitemap: ' . esc_url( $sitemap_url ) . "\n";
	}

	// Write to file.
	return $wp_filesystem->put_contents( $robots_file, $content, FS_CHMOD_FILE );
}

/**
 * Update physical robots.txt file if it exists
 *
 * @param array|null $settings Optional settings array; fetched from DB if null.
 * @return bool True on success, false on failure.
 */
function ayudawp_acs_update_physical_robots( $settings = null ) {
	if ( ! ayudawp_acs_init_filesystem() ) {
		return false;
	}

	global $wp_filesystem;
	$robots_file = ABSPATH . 'robots.txt';

	// If settings not provided, get from database.
	if ( null === $settings ) {
		$settings = get_option( 'ayudawp_acs_settings', array() );
	}

	// Create physical robots.txt if option is enabled and file doesn't exist.
	if ( ! $wp_filesystem->exists( $robots_file ) ) {
		if ( ( $settings['create_physical_robots'] ?? 'no' ) === 'yes' ) {
			ayudawp_acs_create_physical_robots();
		} else {
			return false;
		}
	}

	if ( ! $wp_filesystem->exists( $robots_file ) || ! $wp_filesystem->is_writable( $robots_file ) ) {
		return false;
	}

	// Read current content.
	$current_content = $wp_filesystem->get_contents( $robots_file );

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

	// Remove ALL Content Signals blocks using BEGIN/END markers.
	$pattern         = '/\n*# BEGIN AI Content Signals\n.*?# END AI Content Signals\n*/s';
	$current_content = preg_replace( $pattern, '', $current_content );

	// Trim any extra newlines.
	$current_content = rtrim( $current_content );

	// Generate new Content Signals.
	$content_signals = ayudawp_acs_generate_content_signals();

	// Combine content.
	$new_content = $current_content . "\n" . $content_signals;

	// Write to file.
	return $wp_filesystem->put_contents( $robots_file, $new_content, FS_CHMOD_FILE );
}

/**
 * Hook to update robots.txt after settings are saved
 *
 * @param array $old_value Previous settings.
 * @param array $new_value New settings.
 */
function ayudawp_acs_update_after_save( $old_value, $new_value ) {
	ayudawp_acs_update_physical_robots( $new_value );
}
add_action( 'update_option_ayudawp_acs_settings', 'ayudawp_acs_update_after_save', 10, 2 );

/**
 * Hook to update robots.txt after settings are added for the first time
 *
 * @param string $option Option name.
 * @param mixed  $value  Option value.
 */
function ayudawp_acs_update_after_add( $option, $value ) {
	ayudawp_acs_update_physical_robots( $value );
}
add_action( 'add_option_ayudawp_acs_settings', 'ayudawp_acs_update_after_add', 10, 2 );

/**
 * Remove Content Signals from physical robots.txt
 *
 * Called from uninstall.php during plugin removal.
 *
 * @return bool True on success, false on failure.
 */
function ayudawp_acs_remove_from_physical_robots() {
	if ( ! ayudawp_acs_init_filesystem() ) {
		return false;
	}

	global $wp_filesystem;
	$robots_file = ABSPATH . 'robots.txt';

	if ( ! $wp_filesystem->exists( $robots_file ) || ! $wp_filesystem->is_writable( $robots_file ) ) {
		return false;
	}

	// Read current content.
	$current_content = $wp_filesystem->get_contents( $robots_file );

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

	// Remove ALL Content Signals blocks using BEGIN/END markers.
	$pattern     = '/\n*# BEGIN AI Content Signals\n.*?# END AI Content Signals\n*/s';
	$new_content = preg_replace( $pattern, '', $current_content );

	// Trim any extra newlines.
	$new_content = rtrim( $new_content );

	// Write to file.
	return $wp_filesystem->put_contents( $robots_file, $new_content, FS_CHMOD_FILE );
}