<?php
/**
 * URL inventory and target cache of a scan (S-SCAN).
 *
 * @package ForensixDigital
 */

declare(strict_types=1);

namespace ForensixDigital\Services\Scan;

defined( 'ABSPATH' ) || exit;

// phpcs:disable WordPress.DB.DirectDatabaseQuery.DirectQuery, WordPress.DB.DirectDatabaseQuery.NoCaching -- Our own tables; scan rows change constantly and must not be cached.

use ForensixDigital\Core\Schema;

/**
 * Stores the pages a scan will visit (with their coverage status) and the result of every URL
 * it checked, so each target is requested once per scan however many pages link to it.
 * Rows of older scans are purged when a new scan starts.
 */
final class ScanQueue {

	public const PENDING = 'pending';
	public const SCANNED = 'scanned';
	public const SKIPPED = 'skipped';
	public const FAILED  = 'failed';

	private const BATCH = 100;

	/**
	 * Table names.
	 *
	 * @var array{urls: string, targets: string}
	 */
	private array $t;

	/**
	 * Creates the queue.
	 */
	public function __construct() {
		$this->t = array(
			'urls'    => Schema::table( 'forensix_scan_urls' ),
			'targets' => Schema::table( 'forensix_scan_targets' ),
		);
	}

	/**
	 * Adds pages. A page added twice keeps its highest importance.
	 *
	 * @param int   $scan_id Scan id.
	 * @param array $rows    Pages.
	 * @phpstan-param array<int, array{url: string, origin: string, importance: int, status?: string, reason?: string}> $rows
	 */
	public function add( int $scan_id, array $rows ): void {
		global $wpdb;
		foreach ( array_chunk( $rows, self::BATCH ) as $chunk ) {
			$values = array();
			$params = array( $this->t['urls'] );
			foreach ( $chunk as $row ) {
				$values[] = '(%d, %s, %s, %s, %d, %s, %s)';
				array_push(
					$params,
					$scan_id,
					$row['url'],
					sha1( $row['url'] ),
					substr( $row['origin'], 0, 64 ),
					max( 1, min( 255, $row['importance'] ) ),
					$row['status'] ?? self::PENDING,
					substr( $row['reason'] ?? '', 0, 64 )
				);
			}
			// phpcs:disable WordPress.DB.PreparedSQLPlaceholders.ReplacementsWrongNumber, WordPress.DB.PreparedSQL.InterpolatedNotPrepared, WordPress.DB.PreparedSQL.NotPrepared -- Placeholders are built above, one group per row.
			$wpdb->query(
				$wpdb->prepare(
					'INSERT INTO %i (scan_id, url, url_hash, origin, importance, status, reason) VALUES ' . implode( ', ', $values ) . '
					ON DUPLICATE KEY UPDATE importance = GREATEST(importance, VALUES(importance))',
					$params
				)
			);
			// phpcs:enable WordPress.DB.PreparedSQLPlaceholders.ReplacementsWrongNumber, WordPress.DB.PreparedSQL.InterpolatedNotPrepared, WordPress.DB.PreparedSQL.NotPrepared
		}
	}

	/**
	 * Number of pages in the inventory.
	 *
	 * @param int $scan_id Scan id.
	 */
	public function count( int $scan_id ): int {
		global $wpdb;
		return (int) $wpdb->get_var( $wpdb->prepare( 'SELECT COUNT(*) FROM %i WHERE scan_id = %d', $this->t['urls'], $scan_id ) );
	}

	/**
	 * Next page still to visit, in the order they were added.
	 *
	 * @param int $scan_id Scan id.
	 * @return array{id: int, url: string, importance: int}|null
	 */
	public function next( int $scan_id ): ?array {
		global $wpdb;
		$row = $wpdb->get_row(
			$wpdb->prepare( 'SELECT id, url, importance FROM %i WHERE scan_id = %d AND status = %s ORDER BY id LIMIT 1', $this->t['urls'], $scan_id, self::PENDING ),
			ARRAY_A
		);
		if ( ! is_array( $row ) ) {
			return null;
		}
		return array(
			'id'         => (int) $row['id'],
			'url'        => (string) $row['url'],
			'importance' => (int) $row['importance'],
		);
	}

	/**
	 * Records the outcome of visiting a page.
	 *
	 * @param int    $id          Row id.
	 * @param string $status      SCANNED, SKIPPED or FAILED.
	 * @param string $reason      Why it was skipped or failed.
	 * @param int    $http_status HTTP status, 0 when none.
	 */
	public function mark( int $id, string $status, string $reason = '', int $http_status = 0 ): void {
		global $wpdb;
		$wpdb->update(
			$this->t['urls'],
			array(
				'status'      => $status,
				'reason'      => substr( $reason, 0, 64 ),
				'http_status' => max( 0, min( 65535, $http_status ) ),
			),
			array( 'id' => $id ),
			array( '%s', '%s', '%d' ),
			array( '%d' )
		);
	}

	/**
	 * Coverage totals: discovered, scanned, skipped, failed, pending, and skip/fail reasons.
	 *
	 * @param int $scan_id Scan id.
	 * @return array{discovered: int, scanned: int, skipped: int, failed: int, pending: int, reasons: array<string, int>}
	 */
	public function coverage( int $scan_id ): array {
		global $wpdb;
		$rows = $wpdb->get_results(
			$wpdb->prepare( 'SELECT status, reason, COUNT(*) AS n FROM %i WHERE scan_id = %d GROUP BY status, reason', $this->t['urls'], $scan_id ),
			ARRAY_A
		);

		$counts  = array(
			self::SCANNED => 0,
			self::SKIPPED => 0,
			self::FAILED  => 0,
			self::PENDING => 0,
		);
		$reasons = array();
		$total   = 0;
		foreach ( (array) $rows as $row ) {
			$n      = (int) $row['n'];
			$status = (string) $row['status'];
			$reason = (string) $row['reason'];
			$total += $n;
			if ( isset( $counts[ $status ] ) ) {
				$counts[ $status ] += $n;
			}
			if ( '' !== $reason ) {
				$reasons[ $reason ] = ( $reasons[ $reason ] ?? 0 ) + $n;
			}
		}
		return array(
			'discovered' => $total,
			'scanned'    => $counts[ self::SCANNED ],
			'skipped'    => $counts[ self::SKIPPED ],
			'failed'     => $counts[ self::FAILED ],
			'pending'    => $counts[ self::PENDING ],
			'reasons'    => $reasons,
		);
	}

	/**
	 * Pages of the inventory with one status.
	 *
	 * @param int    $scan_id  Scan id.
	 * @param string $status   Status, or '' for all.
	 * @param int    $page     Page number, from 1.
	 * @param int    $per_page Rows per page.
	 * @return array{items: array<int, array{url: string, status: string, reason: string, http_status: int, origin: string}>, total: int}
	 */
	public function urls( int $scan_id, string $status = '', int $page = 1, int $per_page = 50 ): array {
		global $wpdb;
		$per_page = max( 1, min( 200, $per_page ) );
		$offset   = ( max( 1, $page ) - 1 ) * $per_page;
		if ( '' === $status ) {
			$total = (int) $wpdb->get_var( $wpdb->prepare( 'SELECT COUNT(*) FROM %i WHERE scan_id = %d', $this->t['urls'], $scan_id ) );
			$rows  = $wpdb->get_results(
				$wpdb->prepare( 'SELECT url, status, reason, http_status, origin FROM %i WHERE scan_id = %d ORDER BY id LIMIT %d OFFSET %d', $this->t['urls'], $scan_id, $per_page, $offset ),
				ARRAY_A
			);
		} else {
			$total = (int) $wpdb->get_var( $wpdb->prepare( 'SELECT COUNT(*) FROM %i WHERE scan_id = %d AND status = %s', $this->t['urls'], $scan_id, $status ) );
			$rows  = $wpdb->get_results(
				$wpdb->prepare( 'SELECT url, status, reason, http_status, origin FROM %i WHERE scan_id = %d AND status = %s ORDER BY id LIMIT %d OFFSET %d', $this->t['urls'], $scan_id, $status, $per_page, $offset ),
				ARRAY_A
			);
		}
		$items = array();
		foreach ( (array) $rows as $row ) {
			$items[] = array(
				'url'         => (string) $row['url'],
				'status'      => (string) $row['status'],
				'reason'      => (string) $row['reason'],
				'http_status' => (int) $row['http_status'],
				'origin'      => (string) $row['origin'],
			);
		}
		return array(
			'items' => $items,
			'total' => $total,
		);
	}

	/**
	 * A checked target of this scan.
	 *
	 * @param int    $scan_id Scan id.
	 * @param string $url     Target URL.
	 */
	public function target( int $scan_id, string $url ): ?TargetResult {
		global $wpdb;
		$row = $wpdb->get_row(
			$wpdb->prepare( 'SELECT verdict, http_status, final_url, hops, error FROM %i WHERE scan_id = %d AND url_hash = %s', $this->t['targets'], $scan_id, sha1( $url ) ),
			ARRAY_A
		);
		if ( ! is_array( $row ) ) {
			return null;
		}
		return new TargetResult( (string) $row['verdict'], (int) $row['http_status'], (string) $row['final_url'], (int) $row['hops'], (string) $row['error'] );
	}

	/**
	 * Stores a checked target.
	 *
	 * @param int          $scan_id Scan id.
	 * @param string       $url     Target URL.
	 * @param TargetResult $result  Result.
	 */
	public function save_target( int $scan_id, string $url, TargetResult $result ): void {
		global $wpdb;
		$wpdb->query(
			$wpdb->prepare(
				'INSERT INTO %i (scan_id, url, url_hash, verdict, http_status, final_url, hops, error) VALUES (%d, %s, %s, %s, %d, %s, %d, %s)
				ON DUPLICATE KEY UPDATE verdict = VALUES(verdict), http_status = VALUES(http_status), final_url = VALUES(final_url), hops = VALUES(hops), error = VALUES(error)',
				$this->t['targets'],
				$scan_id,
				$url,
				sha1( $url ),
				$result->verdict,
				max( 0, min( 65535, $result->status ) ),
				$result->final_url,
				max( 0, min( 255, $result->hops ) ),
				substr( $result->error, 0, 191 )
			)
		);
	}

	/**
	 * Removes the rows of every scan before this one.
	 *
	 * @param int $scan_id Scan id to keep, and newer.
	 */
	public function purge_before( int $scan_id ): void {
		global $wpdb;
		$wpdb->query( $wpdb->prepare( 'DELETE FROM %i WHERE scan_id < %d', $this->t['urls'], $scan_id ) );
		$wpdb->query( $wpdb->prepare( 'DELETE FROM %i WHERE scan_id < %d', $this->t['targets'], $scan_id ) );
	}

	/**
	 * Removes all rows. Used when the Site Scan data is deleted.
	 */
	public function delete_all(): void {
		global $wpdb;
		$wpdb->query( $wpdb->prepare( 'DELETE FROM %i', $this->t['urls'] ) );
		$wpdb->query( $wpdb->prepare( 'DELETE FROM %i', $this->t['targets'] ) );
	}
}