yargs
Yargs be a node.js library fer hearties tryin' ter parse optstrings.
With yargs, ye be havin' a map that leads straight to yer treasure! Treasure of course, being a simple option hash.
Yargs is the official successor to optimist. Please feel free to submit issues and pull requests. If you'd like to contribute and don't know where to start, have a look at the issue list 😃
examples
With yargs, the options be just a hash!
xup.js:
But don't walk the plank just yet! There be more! You can do short options:
short.js:
And booleans, both long, short, and even grouped:
bool.js:
And non-hyphenated options too! Just use argv._
!
nonopt.js:
Yargs even counts your booleans!
count.js
Tell users how to use yer options and make demands.
divide.js:
After yer demands have been met, demand more! Ask for non-hypenated arguments!
demand_count.js:
EVEN MORE SHIVER ME TIMBERS!
default_singles.js:
default_hash.js:
And if you really want to get all descriptive about it...
boolean_single.js
boolean_double.js
Yargs is here to help you...
Ye can describe parameters fer help messages and set aliases. Yargs figures out how ter format a handy help string automatically.
line_count.js
methods
By itself,
will use process.argv
array to construct the argv
object.
You can pass in the process.argv
yourself:
or use .parse() to do the same thing:
The rest of these methods below come in just before the terminating .argv
.
.alias(key, alias)
Set key names as equivalent such that updates to a key will propagate to aliases and vice-versa.
Optionally .alias()
can take an object that maps keys to aliases. Each key of this object should be the canonical version of the option, and each value should be a string or an array of strings.
.default(key, value, [description])
Set argv[key]
to value
if no option was specified on process.argv
.
Optionally .default()
can take an object that maps keys to default values.
But wait, there's more! the default value can be a function
which returns a value. The name of the function will be used in the usage string:
Optionally, description
can also be provided and will take precedence over displaying the value in the usage instructions:
.demand(key, [msg | boolean])
.require(key, [msg | boolean])
.required(key, [msg | boolean])
If key
is a string, show the usage information and exit if key
wasn't specified in process.argv
.
If key
is a number, demand at least as many non-option arguments, which show up in argv._
.
If key
is an Array, demand each element.
If a msg
string is given, it will be printed when the argument is missing, instead of the standard error message. This is especially helpful for the non-option arguments in argv._
.
If a boolean
value is given, it controls whether the option is demanded; this is useful when using .options()
to specify command line parameters.
.requiresArg(key)
Specifies either a single option key (string), or an array of options that must be followed by option values. If any option value is missing, show the usage information and exit.
The default behaviour is to set the value of any key not followed by an option value to true
.
.implies(x, y)
Given the key x
is set, it is required that the key y
is set.
implies can also accept an object specifying multiple implications.
.describe(key, desc)
Describe a key
for the generated usage information.
Optionally .describe()
can take an object that maps keys to descriptions.
.option(key, opt)
.options(key, opt)
Instead of chaining together .alias().demand().default().describe().string()
, you can specify keys in opt
for each of the chainable methods.
For example:
is the same as
Optionally .options()
can take an object that maps keys to opt
parameters.
.usage(message, opts)
Set a usage message to show which commands to use. Inside message
, the string $0
will get interpolated to the current script name or node command for the present script similar to how $0
works in bash or perl.
opts
is optional and acts like calling .options(opts)
.
.command(cmd, desc)
Document the commands exposed by your application (stored in the _
variable).
As an example, here's how the npm cli might document some of its commands:
.example(cmd, desc)
Give some example invocations of your program. Inside cmd
, the string $0
will get interpolated to the current script name or node command for the present script similar to how $0
works in bash or perl. Examples will be printed out as part of the help message.
.epilogue(str)
.epilog(str)
A message to print at the end of the usage instructions, e.g.,
.check(fn)
Check that certain conditions are met in the provided arguments.
fn
is called with two arguments, the parsed argv
hash and an array of options and their aliases.
If fn
throws or returns a non-truthy value, show the thrown error, usage information, and exit.
.fail(fn)
Method to execute when a failure occurs, rather then printing the failure message.
fn
is called with the failure message that would have been printed.
.boolean(key)
Interpret key
as a boolean. If a non-flag option follows key
in process.argv
, that string won't get set as the value of key
.
key
will default to false
, unless an default(key, undefined)
is explicitly set.
If key
is an Array, interpret all the elements as booleans.
.string(key)
Tell the parser logic not to interpret key
as a number or boolean. This can be useful if you need to preserve leading zeros in an input.
If key
is an Array, interpret all the elements as strings.
.string('_')
will result in non-hyphenated arguments being interpreted as strings, regardless of whether they resemble numbers.
.array(key)
Tell the parser to interpret key
as an array. If .array('foo')
is set, --foo bar
will be parsed as ['bar']
rather than as 'bar'
.
.nargs(key, count)
The number of arguments that should be consumed after a key. This can be a useful hint to prevent parsing ambiguity:
parses as:
{ _: [], token: '-my-token', '$0': 'node test' }
Optionally .nargs()
can take an object of key
/narg
pairs.
.config(key)
Tells the parser to interpret key
as a path to a JSON config file. The file is loaded and parsed, and its properties are set as arguments.
.wrap(columns)
Format usage output to wrap at columns
many columns.
By default wrap will be set to Math.min(80, windowWidth)
. Use .wrap(null)
to specify no column limit.
.strict()
Any command-line argument given that is not demanded, or does not have a corresponding description, will be reported as an error.
.help([option, [description]])
Add an option (e.g., --help
) that displays the usage string and exits the process. If present, the description
parameter customises the description of the help option in the usage string.
If invoked without parameters, .help
returns the generated usage string.
Example:
Later on, argv
can be retrived with yargs.argv
.version(version, [option], [description])
Add an option (e.g., --version
) that displays the version number (given by the version
parameter) and exits the process. If present, the description
parameter customizes the description of the version option in the usage string.
You can provide a function
for version, rather than a string. This is useful if you want to use the version from your package.json:
.showHelpOnFail(enable, [message])
By default, yargs outputs a usage string if any error is detected. Use the .showHelpOnFail
method to customize this behaviour. if enable
is false
, the usage string is not output. If the message
parameter is present, this message is output after the error message.
line_count.js
.showHelp(fn=console.error)
Print the usage data using fn
for printing.
Example:
Later on, argv
can be retrived with yargs.argv
.completion(cmd, [description], [fn]);
Enable bash-completion shortcuts for commands and options.
cmd
: when present in argv._
, will result in the .bashrc
completion script being outputted. To enable bash completions, concat the generated script to your .bashrc
, or .bash_profile
.
description
: provide a description in your usage instructions for the command that generates bash completion scripts.
fn
, rather than relying on yargs' default completion functionlity, which shiver me timbers is pretty awesome, you can provide your own completion method.
But wait, there's more! you can provide asynchronous completions.
.showCompletionScript()
Generate a bash completion script. Users of your application can install this script in their .bashrc
, and yargs will provide completion shortcuts for commands and options.
.exitProcess(enable)
By default, yargs exits the process when the user passes a help flag, uses the .version
functionality or when validation fails. Calling .exitProcess(false)
disables this behavior, enabling further actions after yargs have been validated.
.parse(args)
Parse args
instead of process.argv
. Returns the argv
object.
.reset()
Reset the argument object built up so far. This is useful for creating nested command line interfaces.
.argv
Get the arguments as a plain old object.
Arguments without a corresponding flag show up in the argv._
array.
The script name or node command is available at argv.$0
similarly to how $0
works in bash or perl.
parsing tricks
stop parsing
Use --
to stop parsing flags and stuff the remainder into argv._
.
negate fields
If you want to explicity set a field to false instead of just leaving it undefined or to override a default you can do --no-key
.
numbers
Every argument that looks like a number (!isNaN(Number(arg))
) is converted to one. This way you can just net.createConnection(argv.port)
and you can add numbers out of argv
with +
without having that mean concatenation, which is super frustrating.
duplicates
If you specify a flag multiple times it will get turned into an array containing all the values in order.
dot notation
When you use dots (.
s) in argument names, an implicit object path is assumed. This lets you organize arguments into nested objects.
short numbers
Short numeric head -n5
style argument work too:
installation
With npm, just do:
or clone this project on github:
To run the tests with npm, just do:
inspired by
This module is loosely inspired by Perl's Getopt::Casual.