Add filter for skipping files earlier in the pipeline than 'wp_parser_pre_import_file'

[pbiron]

\\WP_Parser\Importer::import_file() provides the wp_parser_pre_import_file filter to skip importing a particular file.

It would improve efficiency if there were a similar filter earlier in the pipeline, e.g., in \\WP_Parser\get_wp_files(). For example,

function get_wp_files( $directory ) {
	$iterableFiles = new \RecursiveIteratorIterator(
		new \RecursiveDirectoryIterator( $directory )
	);
	$files         = array();

	try {
		foreach ( $iterableFiles as $file ) {
			if ( 'php' !== $file->getExtension() ) {
				continue;
			}

			if ( ! apply_filters( 'wp_parser_pre_get_wp_file', true, $file->getPathname(), $directory ) ) {
				continue;
			}

			$files[] = $file->getPathname();
		}
	} catch ( \UnexpectedValueException $exc ) {
		return new \WP_Error(
			'unexpected_value_exception',
			sprintf( 'Directory [%s] contained a directory we can not recurse into', $directory )
		);
	}

	return $files;
}

I use this plugin to produce developer documentation for plugins/themes I create for clients, and currently hook into wp_parser_pre_import_file to skip importing anything in tests/ or vendor/ sub-dirs of a plugin/theme. Having the proposed wp_parser_pre_get_wp_file hook would allow avoiding the overhead of parsing all those files, only to skip them later in the pipeline.

Here is an example of a func I would hook into the proposed filter, to show why the proposed filter takes the additional $directory param that wp_parser_pre_import_file does not:

add_filter( 'wp_parser_pre_get_wp_file', 'shc_wp_parser_skip_files', 10, 3 );

function shc_wp_parser_skip_files( $default, $file, $directory ) {
	$file = str_replace( $directory . DIRECTORY_SEPARATOR, '', $file );
	$paths = explode( DIRECTORY_SEPARATOR, $file );
	if ( in_array( $paths[1], array( 'tests', 'vendor' ) ) ) {
		return false;
	}

	return $default;
}

[pbiron]

I'd also recommend adding something like the following to \\WP_Parser\Importer::import_file():

if ( ! apply_filters( 'wp_parser_pre_import_file', true, $file ) ) {
	$this->logger->info( sprintf( 'Skipping  file %1$s', $file['path'] ) );

	return;
}

whether this proposal is accepted or not.

[keesiemeijer]

+1 for this feature. I also have a plugin parsing plugins where I exclude directories. For filtering directories we should use a RecursiveCallbackFilterIterator so the directories (and its files) are not traversed recursively. This saves a lot of computing power.

Here is how I did it for my plugin
https://github.com/keesiemeijer/wp-plugin-parser/blob/master/includes/class-file-parser.php#L109

[sebastienserre]

Hi All,
The @keesiemeijer solution is working great, what about merging it ?