Pairs nicely with libraries like minimist and yargs-parser that are great for parsing command line arguments but don't generate help messages.

Installation

Install with npm:

npm install help-output

Or with yarn:

yarn add help-output

Usage

Import the helpOutput function and pass it a configuration object with the name of your program, positional arguments, and options:

const helpOutput = require('help-output')

const message = helpOutput({
    name: 'mycli',
    description: 'A cli program that does something useful',
    positional: [{
        name: 'arg',
        description: 'A required positional argument',
    }],
    options: [{
        name: 'log',
        arg: 'lvl',
        conflicts: ['q', 'd'],
        description: "Set logging level ('error', 'warning', 'debug', or 'silent')",
    }, {
        name: 'quiet',
        alias: 'q',
        description: "Silence console output, equivalent to setting --log='silent'",
    }, {
        name: 'debug',
        alias: 'd',
        conflicts: 'quiet',
        description: "Log verbose output, equivalent to setting --log='debug'",
    }, {
        name: 'help',
        alias: 'h',
        preferAlias: true,
        description: 'Display this help message',
    }, {
        name: 'version',
        alias: 'v',
        preferAlias: true,
        description: 'Display program version',
    }],
}, { width: 80 });

console.log(message);

Positional arguments and option flags are formatted into columns, and a git style usage section is automatically generated based on the information provided about them:

API

helpOutput(config, [options])

A function that takes a config object and options object as arguments, and returns a help output message string. The config object is where you specify the arguments and options that will be displayed in the help output message, while the optional options object gives you some control over how the message will be formatted.

Configuration

Note: configuration validation is strict - an error will be thrown if any properties do not conform to the schema detailed below.

config.name

Type: string

The name of your cli program. If left unspecified, a name will be inferred from process.argv.

config.title

Type: string

An optional title to display at the top of the outputted help message. It can it can include two placeholder strings, %name and %version, which will be replaced by the values of config.name and config.version, respectively.

config.version

Type: string

The programs current version. This field only necessary if config.title is set and includes the %version placeholder string.

config.description

Type: string

An optional description of what the program does. If specified, it will be displayed at the top of the outputted help message, just below config.title.

config.positional

Type: Object[]

An array of objects specifying your programs positional arguments. They will be displayed in the order they are specified. Each argument spec object can contain the following properties:

• name - The name of the positional argument (required).
• description - A description of the argument. This property is not strictly required, but is strongly encouraged.
• required - A boolean (defaults to true) indicating whether the argument is required.
• variadic - A boolean (defaults to false) indicating whether the argument is variadic.

config.options

Type: Object[]

An array of objects specifying your programs option flags. They will be displayed in the order they are specified. Each option spec object can contain the following properties:

• name - The name of the option (required).
• description - A description of the option. This property is not strictly required, but is strongly encouraged.
• required - A boolean (defaults to false) indicating whether the option is required.
• arg - A string, object, or array of either specifying one or more positional arguments the option takes. Object specs use the same structure as those specified in config.positional (minus the description field). String specs are equivalent to specifiying an object containing only a name property.
• alias - A string or array of strings specifying any aliases for the option.
• preferAlias - A boolean or string indicating that the options alias name should be used in the program usage section that gets generated. If value is a string, it must be one of the alias names specified for the option. Defaults to false.
• dependsOn - The name of another option that this option depends on. Make sure that the option referenced here has been configured, otherwise an error will be thrown. References to alias names are allowed.
• conflicts - Another option name or array of names that this option conflicts with. Make sure that these option names specified here reference other options that have been configured, otherwise an error will be thrown. References to alias names are allowed.

Note: relationships specified by the dependsOn and conflicts fields are reflected in the program usage section that is automatically generated. Check out the examples section below for more information.

Options

options.width

Type: number

Character width to wrap the help output message to, defaults to the width of process.stdout.

options.spacing

Type: number

Size of the gap between table columns, default is 2.

options.indent

Type: number

How much to indent the content of each section on the left side, defaults to the value of options.spacing.

options.color

Type: boolean

Whether the returned help message string should be colorized, default is true.

options.styles

Type: Object

Ansi colorization customizations This option is only relevant if options.color is true. Must be an object that maps style selector ids to ansi style values. The following are valid style selector ids that target different components of the help output string:

• 'positional' - selector id for positional arguments (default style is 'yellow').
• 'option' - selector id for option names (default style is 'green').
• 'header' - selector id for section headers (default style is 'bold.underline').

Specifying null or '' for a selector id will result in no style being applied. Check out ansi-styles for a list of valid style values, (multiple values must be separated by a '.', or supplied as an array).

Examples

Additional information specified about a program option, such as whether it is required or if it depends on another option, is reflected in the program usage section that is included in the returned help message.

The following example demonstrates how using the required and dependsOn fields will affect the resulting message.

const message = helpOutput({
    name: 'mycli',
    options: [{
        name: 'foo',
        required: true,
        description: 'A required flag',
    }, {
        name: 'bar',
        description: 'An optional flag',
    }, {
        name: 'baz',
        description: 'Another optional flag',
    }, {
        name: 'qux',
        dependsOn: 'baz',
        description: 'A flag that can only be specified if --baz is present',
    }],
});

The generated usage will indicate that the --foo flag must be present, the --bar and --baz flags are optional, and the --qux flag is allowed only if the --baz flag is present:

Mutually Exclusive Groups

Option conflicts specified via the conflicts field will be indicated in the generated usage in form of mutually exclusive groups. Groups with required members are enclosed in parentheses, while groups with optional members are enclosed in brackets.

Note: all options that make up a mutually exclusive group must be either optional or required - attempting to have a mutually exclusive group with a mixture of optional and required members will result in an error being thrown.

In the following example, the --foo and --bar flags form one group, and the --baz and --qux flags form another.

const message = helpOutput({
    name: 'mycli',
    options: [{
        name: 'foo',
        required: true,
        conflicts: 'bar',
        description: 'Part of a required mutually exclusive group with --bar',
    }, {
        name: 'bar',
        required: true,
        description: 'Part of a required mutually exclusive group with --foo',
    }, {
        name: 'baz',
        conflicts: 'qux',
        description: 'Part of an optional mutually exclusive group with --qux',
    }, {
        name: 'qux',
        description: 'Part of an optional mutually exclusive group with --baz',
    }],
});

The generated usage will indicate that either the --foo or the --bar flag must be preset, and that either the --baz or the --qux flag may optionally be present:

Partially Exclusive Groups

It is not a requirement that every member of a mutually exclusive group conflict with every other option in that group.

In the following example --foo conflicts with the --bar, --baz, and --qux flags, but the latter three flags do not conflict with each other.

const message = helpOutput({
    name: 'mycli',
    options: [{
        name: 'foo',
        conflicts: ['bar', 'baz', 'qux'],
        description: 'Cannot be specified alongside --bar, --baz, or --qux',
    }, {
        name: 'bar',
        description: 'Can be specified alongside --baz and --qux, but not --foo',
    }, {
        name: 'baz',
        description: 'Can be specified alongside --bar and --qux, but not --foo',
    }, {
        name: 'qux',
        description: 'Can be specified alongside --bar and --baz, but not --foo',
    }],
});

The generated usage will indicate that either the --foo flag may be present, or any combination of the --bar, --baz, and --qux flags may be present:

Development

Contributions are welcome!

To report a bug or request a feature, please open a new issue.

Running Tests

Install project dependencies and run the test suite with the following command:

yarn && yarn test

To generate coverage reports, run:

yarn test --coverage

License

MIT