helmsman

node-helmsman

Easily make command line interfaces using git style subcommand executables

So what does helmsman actually do?

A common setup for command line applications is <command> <subcommand> <arguments/options> (for example: git commit -m 'message'). Rather than having a giant file that switches or if elses over each potential subcommand, it's much neater to store each subcommand in it's own file (bin/command,bin/command-subcomand, bin/command-subcommand2, etc). Helmsman makes it easy to add, modify or delete subcommands without having to do housekeeping steps in your root command file or package.json

Features

• Helmsman is automatically aware of all the <command>-<subcommand> files in your modules bin/ (or any folder you tell it to look at)
• <command> --help automatically generates help output, telling you all the subcommands that are available to you
• <command> --version prints the version from package.json of the module requiring helmsman
• Running <command> <subcommand> automatically executes the <command>-<subcommand> file, passing along all the arguments & options
• Helmsman is capable of smart command completion including dynamic shorthands and spelling correction (eg: <command> st => <command> status or <command> isntall => <command> install )
• Use whatever option parsing library you want for your subcommands (optimist, commander, etc)
• Helmsman is minimally intrusive in your subcommands

Installation & Setup

In your command line application folder:

npm install helmsman --save

Setting up your main executable: <command>

In your main executable, add helmsman:

#!/usr/bin/env node
 
var helmsman = require('helmsman');
 
helmsman().parse();

Want to append in additional help messaging or modify the arguments that are parsed?

#!/usr/bin/env node
 
var helmsman = require('helmsman');
 
var cli = helmsman()
 
cli.on('--help', function(){
  console.log('EXTRA HELPFUL!');
});
 
var argv = process.argv;
 
argv.push('--pizza');
 
// parse() can accept modified arguments, otherwise it defaults to process.argv
cli.parse(argv);

Setting up your sub-commands: <command>-<subcommand>

For your sub-executables to work with helmsman you need to do two things: 1. Expose metadata about the task, like its description and 2. Make sure the meat & potatoes of the script only runs when it's directly called

#!/usr/bin/env node
 
// 1. Expose the metadata
exports.command = {
  description: 'Show current worker counts and their pids'
};
 
// 2. Make sure it only runs when it's directly called:
if (require.main === module) {
  // Parse options and run the magic
}

Note: If you're not putting each script in package.json's bin object, make sure that the sub-commands are executable by running `chmod +x bin/-

API

helmsman([options]) or new Helmsman([options])

• options {Object}

Create an instance of helmsman. It is an EventEmitter and will also begin searching for files once it's instantiated.

Events

• --help: Emitted when --help is passed as the first option or no commands or options are passed

Options

• localDir: The local module folder where to search for executable files. Defaults to the directory of the executable (eg: If you execute <module folder>/bin/<command> the localDir will be <module folder>/bin)
• prefix: The prefix of the subcommands to search for. Defaults to the executed file (eg: If you run <command> it will search for files in the localDir that start with <command>-
• metadata: An object containing keys of command names and sub-objects containing the keys description and optionally arguments
• usePath: If true helmsman will search the PATH for commands matching the prefix
• fillCommandData: An optional function to use to retrieve metadata from a command file; takes a defaults object, a filename, and an extension
• fallbackCommandData: If true helmsman will use its default function to retrieve metadata from a command file if the user-specified fuction returns a falsy value
• ignoreRequireFail: If true helmsman will ignore failures to require an extensionless or .js-extensioned command file
• nodePath: The path to the node executable on Windows, defaults to 'node'

Methods

• parse([argv]) Parse argv or process.argv if there is no argv and either display the help or run the subcommand

exports.command

• description: A one line description of the command. Required.
• arguments: A shorthand for options the subcommand accepts. Generated help will include it next to command. See help <command>"

TODO

• Allow for automatically including npm installed libraries

Thanks

Much of this was inspired by TJ Holowaychuk's commander and component