args 0.12.2

Library for defining parsers for parsing raw command-line arguments into a set of options and values using GNU and POSIX style options.

Parses raw command-line arguments into a set of options and values.

This library supports GNU and POSIX style options, and it works in both server-side and client-side apps.

Defining options

First create an ArgParser:

var parser = new ArgParser();

Then define a set of options on that parser using addOption() and addFlag(). Here's the minimal way to create an option named "name":

parser.addOption('name');

When an option can only be set or unset (as opposed to taking a string value), use a flag:

parser.addFlag('name');

Flag options, by default, accept a 'no-' prefix to negate the option. You can disable the 'no-' prefix using the negatable parameter:

parser.addFlag('name', negatable: false);

Note: From here on out, "option" refers to both regular options and flags. In cases where the distinction matters, we'll use "non-flag option."

Options can have an optional single-character abbreviation, specified with the abbr parameter:

parser.addOption('mode', abbr: 'm');
parser.addFlag('verbose', abbr: 'v');

Options can also have a default value, specified with the defaultsTo parameter. The default value is used when arguments don't specify the option.

parser.addOption('mode', defaultsTo: 'debug');
parser.addFlag('verbose', defaultsTo: false);

The default value for non-flag options can be any string. For flags, it must be a bool.

To validate a non-flag option, you can use the allowed parameter to provide an allowed set of values. When you do, the parser throws a [FormatException] if the value for an option is not in the allowed set. Here's an example of specifying allowed values:

parser.addOption('mode', allowed: ['debug', 'release']);

You can use the callback parameter to associate a function with an option. Later, when parsing occurs, the callback function is invoked with the value of the option:

parser.addOption('mode', callback: (mode) => print('Got mode $mode'));
parser.addFlag('verbose', callback: (verbose) {
  if (verbose) print('Verbose');
});

The callbacks for all options are called whenever a set of arguments is parsed. If an option isn't provided in the args, its callback is passed the default value, or null if no default value is set.

Parsing arguments

Once you have an ArgParser set up with some options and flags, you use it by calling ArgParser.parse() with a set of arguments:

var results = parser.parse(['some', 'command', 'line', 'args']);

These arguments usually come from the arguments to main(), but you can pass in any list of strings. The parse() method returns an instance of ArgResults, a map-like object that contains the values of the parsed options.

var parser = new ArgParser();
parser.addOption('mode');
parser.addFlag('verbose', defaultsTo: true);
var results = parser.parse(['--mode', 'debug', 'something', 'else']);

print(results['mode']); // debug
print(results['verbose']); // true

By default, the parse() method stops as soon as it reaches -- by itself or anything that the parser doesn't recognize as an option, flag, or option value. If arguments still remain, they go into ArgResults.rest.

print(results.rest); // ['something', 'else']

To continue to parse options found after non-option arguments, pass allowTrailingOptions: true when creating the ArgParser.

Specifying options

To actually pass in options and flags on the command line, use GNU or POSIX style. Consider this option:

parser.addOption('name', abbr: 'n');

You can specify its value on the command line using any of the following:

--name=somevalue
--name somevalue
-nsomevalue
-n somevalue

Consider this flag:

parser.addFlag('name', abbr: 'n');

You can set it to true using one of the following:

--name
-n

You can set it to false using the following:

--no-name

Multiple flag abbreviations can be collapsed into a single argument. Say you define these flags:

parser.addFlag('verbose', abbr: 'v');
parser.addFlag('french', abbr: 'f');
parser.addFlag('iambic-pentameter', abbr: 'i');

You can set all three flags at once:

-vfi

By default, an option has only a single value, with later option values overriding earlier ones; for example:

var parser = new ArgParser();
parser.addOption('mode');
var results = parser.parse(['--mode', 'on', '--mode', 'off']);
print(results['mode']); // prints 'off'

If you need multiple values, set the allowMultiple parameter. In that case the option can occur multiple times, and the parse() method returns a list of values:

var parser = new ArgParser();
parser.addOption('mode', allowMultiple: true);
var results = parser.parse(['--mode', 'on', '--mode', 'off']);
print(results['mode']); // prints '[on, off]'

Defining commands

In addition to options, you can also define commands. A command is a named argument that has its own set of options. For example, consider this shell command:

$ git commit -a

The executable is git, the command is commit, and the -a option is an option passed to the command. You can add a command using the addCommand method:

var parser = new ArgParser();
var command = parser.addCommand('commit');

It returns another ArgParser, which you can then use to define options specific to that command. If you already have an ArgParser for the command's options, you can pass it in:

var parser = new ArgParser();
var command = new ArgParser();
parser.addCommand('commit', command);

The ArgParser for a command can then define options or flags:

command.addFlag('all', abbr: 'a');

You can add multiple commands to the same parser so that a user can select one from a range of possible commands. When parsing an argument list, you can then determine which command was entered and what options were provided for it.

var results = parser.parse(['commit', '-a']);
print(results.command.name);   // "commit"
print(results.command['all']); // true

Options for a command must appear after the command in the argument list. For example, given the above parser, "git -a commit" is not valid. The parser tries to find the right-most command that accepts an option. For example:

var parser = new ArgParser();
parser.addFlag('all', abbr: 'a');
var command = parser.addCommand('commit');
command.addFlag('all', abbr: 'a');

var results = parser.parse(['commit', '-a']);
print(results.command['all']); // true

Here, both the top-level parser and the "commit" command can accept a "-a" (which is probably a bad command line interface, admittedly). In that case, when "-a" appears after "commit", it is applied to that command. If it appears to the left of "commit", it is given to the top-level parser.

Dispatching Commands

If you're writing a command-based application, you can use the CommandRunner and Command classes to help structure it. CommandRunner has built-in support for dispatching to Commands based on command-line arguments, as well as handling --help flags and invalid arguments. For example:

var runner = new CommandRunner("git", "Distributed version control.");
runner.addCommand(new CommitCommand());
runner.addCommand(new StashCommand());
runner.run(['commit', '-a']); // Calls [CommitCommand.run()]

Custom commands are defined by extending the Command class. For example:

class CommitCommand extends Command {
  // The [name] and [description] properties must be defined by every
  // subclass.
  final name = "commit";
  final description = "Record changes to the repository.";

  CommitCommand() {
    // [argParser] is automatically created by the parent class.
    argParser.addFlag('all', abbr: 'a');
  }

  // [run] may also return a Future.
  void run() {
    // [options] is set before [run()] is called and contains the options
    // passed to this command.
    print(options['all']);
  }
}

Commands can also have subcommands, which are added with addSubcommand. A command with subcommands can't run its own code, so run doesn't need to be implemented. For example:

class StashCommand extends Command {
  final String name = "stash";
  final String description = "Stash changes in the working directory.";

  StashCommand() {
    addSubcommand(new StashSaveCommand());
    addSubcommand(new StashListCommand());
  }
}

CommandRunner automatically adds a help command that displays usage information for commands, as well as support for the --help flag for all commands. If it encounters an error parsing the arguments or processing a command, it throws a UsageError; your main() method should catch these and print them appropriately. For example:

runner.run(arguments).catchError((error) {
  if (error is! UsageError) throw error;
  print(error);
  exit(64); // Exit code 64 indicates a usage error.
});

Displaying usage

You can automatically generate nice help text, suitable for use as the output of --help. To display good usage information, you should provide some help text when you create your options.

To define help text for an entire option, use the help: parameter:

parser.addOption('mode', help: 'The compiler configuration',
    allowed: ['debug', 'release']);
parser.addFlag('verbose', help: 'Show additional diagnostic info');

For non-flag options, you can also provide a help string for the parameter:

parser.addOption('out', help: 'The output path', valueHelp: 'path',
    allowed: ['debug', 'release']);

For non-flag options, you can also provide detailed help for each expected value by using the allowedHelp: parameter:

parser.addOption('arch', help: 'The architecture to compile for',
    allowedHelp: {
      'ia32': 'Intel x86',
      'arm': 'ARM Holding 32-bit chip'
    });

To display the help, use the getUsage() method:

print(parser.getUsage());

The resulting string looks something like this:

--mode            The compiler configuration
                  [debug, release]

--out=<path>      The output path
--[no-]verbose    Show additional diagnostic info
--arch            The architecture to compile for
      [arm]       ARM Holding 32-bit chip
      [ia32]      Intel x86