<?php
/**
 * File scanner — enumerates wp-content files (honoring scope + excludes) into a
 * newline-delimited list on disk, so the runner can archive them with a resumable
 * line cursor and never hold the whole tree in memory.
 *
 * Paths are stored relative to WP_CONTENT_DIR (e.g. "uploads/2026/x.jpg").
 *
 * @package AI_SiteArk
 * Copyright (c) 2026 AtoZ INFOWAY INC.
 * Licensed under the GNU General Public License, version 2 or later.
 */

namespace AtoZ\SiteArk\Backup;

use AtoZ\SiteArk\Helpers\Local_Store;

defined( 'ABSPATH' ) || exit;

/**
 * Class File_Scanner
 */
class File_Scanner {

	/**
	 * Content-dir subfolders per scope key.
	 *
	 * @return array<string,string>
	 */
	private static function roots() {
		return array(
			'uploads'   => 'uploads',
			'themes'    => 'themes',
			'plugins'   => 'plugins',
			'muplugins' => 'mu-plugins',
		);
	}

	/**
	 * Write the file list; returns [count, total_bytes].
	 *
	 * @param array  $scope     Scope flags.
	 * @param array  $excludes  Extra path fragments to skip.
	 * @param int    $max_bytes Skip files bigger than this (0 = no cap).
	 * @param string $list_path Where to write the newline-delimited list.
	 * @return array{count:int,bytes:int}
	 */
	public static function build_list( array $scope, array $excludes, $max_bytes, $list_path ) {
		$res = self::build_indexed( $scope, $excludes, $max_bytes, $list_path );
		return array( 'count' => $res['count'], 'bytes' => $res['bytes'] );
	}

	/**
	 * G1 — scan with optional indexing / differential filtering.
	 *
	 * - `$index_out` (base backups): also writes "rel\tsize\tmtime" for every file, so
	 *   later differentials can diff against this cycle's snapshot.
	 * - `$base_index` (differentials): a map rel => "size:mtime"; only new/changed files
	 *   (size or mtime differs) are written to the list, and paths present in the base
	 *   but gone now are returned as `deletions`.
	 *
	 * @param array  $scope      Scope flags.
	 * @param array  $excludes   Extra path fragments to skip.
	 * @param int    $max_bytes  Skip files bigger than this (0 = no cap).
	 * @param string $list_path  Where to write the newline-delimited list.
	 * @param string $index_out  Optional index file to write (full snapshot).
	 * @param array  $base_index Optional base map rel => "size:mtime" (differential mode).
	 * @return array{count:int,bytes:int,deletions:array,total:int}
	 */
	public static function build_indexed( array $scope, array $excludes, $max_bytes, $list_path, $index_out = '', array $base_index = array() ) {
		$content = rtrim( WP_CONTENT_DIR, '/\\' );

		$backup_dir = Local_Store::dir();
		$backup_dir = is_wp_error( $backup_dir ) ? '' : $backup_dir;

		// phpcs:ignore WordPress.WP.AlternativeFunctions.file_system_operations_fopen
		$fh = @fopen( $list_path, 'wb' );
		if ( ! $fh ) {
			return array( 'count' => 0, 'bytes' => 0, 'deletions' => array(), 'total' => 0 );
		}
		$ifh = '';
		if ( '' !== $index_out ) {
			// phpcs:ignore WordPress.WP.AlternativeFunctions.file_system_operations_fopen
			$ifh = @fopen( $index_out, 'wb' );
		}

		$differential = ! empty( $base_index );
		$seen         = array();

		$count = 0;
		$bytes = 0;
		$total = 0;

		// Directories to walk: the four named wp-content sub-trees that are selected…
		$targets = array();
		foreach ( self::roots() as $scope_key => $sub ) {
			if ( ! empty( $scope[ $scope_key ] ) && is_dir( $content . '/' . $sub ) ) {
				$targets[] = array( $content . '/' . $sub, $sub );
			}
		}

		// …plus EVERYTHING ELSE in wp-content when the "other" bucket is on: loose
		// root files (drop-ins like object-cache.php) + any non-standard folder
		// (languages, custom plugin data, etc.). This is what a full-site backup must
		// include; without it the archive silently omits real site data.
		$top_files = array();
		if ( ! empty( $scope['other'] ) ) {
			$named   = array( 'uploads' => 1, 'themes' => 1, 'plugins' => 1, 'mu-plugins' => 1 );
			$entries = @scandir( $content ); // phpcs:ignore WordPress.PHP.NoSilencedErrors.Discouraged
			foreach ( (array) $entries as $entry ) {
				if ( '.' === $entry || '..' === $entry ) {
					continue;
				}
				$abs = $content . '/' . $entry;
				if ( is_dir( $abs ) ) {
					if ( isset( $named[ $entry ] ) ) {
						continue; // Already covered by the named roots above.
					}
					$targets[] = array( $abs, $entry );
				} elseif ( is_file( $abs ) ) {
					$top_files[] = array( $abs, $entry );
				}
			}
		}

		foreach ( $targets as $target ) {
			$abs = $target[0];
			$sub = $target[1];
			try {
				// Do NOT follow symlinks: a symlinked directory under wp-content would pull an
				// unintended external tree into the backup (or loop forever on a symlink cycle).
				// The callback filter returns false for any symlink, so the iterator neither
				// includes it nor descends into it.
				$filter   = new \RecursiveCallbackFilterIterator(
					new \RecursiveDirectoryIterator( $abs, \FilesystemIterator::SKIP_DOTS ),
					static function ( $current ) {
						return ! $current->isLink();
					}
				);
				$iterator = new \RecursiveIteratorIterator( $filter, \RecursiveIteratorIterator::LEAVES_ONLY );
			} catch ( \Exception $e ) {
				continue;
			}

			foreach ( $iterator as $file ) {
				if ( ! $file->isFile() ) {
					continue;
				}
				$abspath = $file->getPathname();
				$rel     = $sub . '/' . substr( $abspath, strlen( $abs ) + 1 );
				$rel     = str_replace( '\\', '/', $rel );
				self::emit( $fh, $ifh, $abspath, $rel, (int) $file->getSize(), (int) $file->getMTime(), $max_bytes, $differential, $base_index, $excludes, $backup_dir, $seen, $count, $bytes, $total );
			}
		}

		// Loose files directly under wp-content (drop-ins, index.php, custom includes).
		foreach ( $top_files as $tf ) {
			self::emit( $fh, $ifh, $tf[0], $tf[1], (int) @filesize( $tf[0] ), (int) @filemtime( $tf[0] ), $max_bytes, $differential, $base_index, $excludes, $backup_dir, $seen, $count, $bytes, $total );
		}

		fclose( $fh ); // phpcs:ignore WordPress.WP.AlternativeFunctions.file_system_operations_fclose
		if ( $ifh ) {
			fclose( $ifh ); // phpcs:ignore WordPress.WP.AlternativeFunctions.file_system_operations_fclose
		}

		$deletions = array();
		if ( $differential ) {
			foreach ( $base_index as $rel => $sig ) {
				if ( empty( $seen[ $rel ] ) ) {
					$deletions[] = $rel;
				}
			}
		}

		return array( 'count' => $count, 'bytes' => $bytes, 'deletions' => $deletions, 'total' => $total );
	}

	/**
	 * Process one file: honor excludes + max size, write it to the list (and index),
	 * and in differential mode skip unchanged files while tracking what was seen.
	 *
	 * @param resource      $fh           List handle.
	 * @param resource|null $ifh          Index handle (or '').
	 * @param string        $abspath      Absolute path.
	 * @param string        $rel          wp-content-relative path.
	 * @param int           $size         File size.
	 * @param int           $mtime        Modified time.
	 * @param int           $max_bytes    Size cap (0 = none).
	 * @param bool          $differential Differential mode.
	 * @param array         $base_index   Base index map (differential).
	 * @param array         $excludes     User excludes.
	 * @param string        $backup_dir   Our backups dir.
	 * @param array         $seen         (by ref) seen set.
	 * @param int           $count        (by ref) written count.
	 * @param int           $bytes        (by ref) written bytes.
	 * @param int           $total        (by ref) total scanned.
	 * @return void
	 */
	private static function emit( $fh, $ifh, $abspath, $rel, $size, $mtime, $max_bytes, $differential, array $base_index, array $excludes, $backup_dir, array &$seen, &$count, &$bytes, &$total ) {
		if ( self::excluded( $rel, $abspath, $excludes, $backup_dir ) ) {
			return;
		}
		if ( $max_bytes > 0 && $size > $max_bytes ) {
			return;
		}
		++$total;
		if ( $ifh ) {
			// phpcs:ignore WordPress.WP.AlternativeFunctions.file_system_operations_fwrite -- streaming append to the file-index; WP_Filesystem cannot stream.
			fwrite( $ifh, $rel . "\t" . $size . "\t" . $mtime . "\n" );
		}
		if ( $differential ) {
			$seen[ $rel ] = true;
			if ( isset( $base_index[ $rel ] ) && $base_index[ $rel ] === $size . ':' . $mtime ) {
				return; // Unchanged — skip.
			}
		}
		// phpcs:ignore WordPress.WP.AlternativeFunctions.file_system_operations_fwrite -- streaming append to the file-list; WP_Filesystem cannot stream.
		fwrite( $fh, $rel . "\n" );
		++$count;
		$bytes += $size;
	}

	/**
	 * Whether a content-relative path should be skipped.
	 *
	 * @param string $rel        Content-relative path.
	 * @param string $abspath    Absolute path.
	 * @param array  $excludes   User exclude fragments.
	 * @param string $backup_dir AI-SiteArk backups dir (never back up ourselves).
	 * @return bool
	 */
	private static function excluded( $rel, $abspath, array $excludes, $backup_dir ) {
		if ( $backup_dir && 0 === strpos( str_replace( '\\', '/', $abspath ), str_replace( '\\', '/', $backup_dir ) ) ) {
			return true;
		}
		$default = array( '/node_modules/', '/.git/', '/.svn/' );
		$needle  = '/' . $rel . '/';
		foreach ( $default as $seg ) {
			if ( false !== strpos( $needle, $seg ) ) {
				return true;
			}
		}
		if ( '.DS_Store' === basename( $rel ) ) {
			return true;
		}
		foreach ( $excludes as $ex ) {
			$ex = trim( (string) $ex );
			if ( '' !== $ex && false !== strpos( $rel, $ex ) ) {
				return true;
			}
		}
		return false;
	}
}