npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

koa-bouncer

v6.0.4

Published

A parameter validation library for Koa routes

Downloads

1,743

Readme

koa-bouncer

Build Status NPM version Dependency Status

NPM

An http parameter validation library for Koa web apps.

npm install --save koa-bouncer

Inspired by RocksonZeta.

Works best with koa-router for routing.

If you'd like to see how koa-bouncer looks in a real (demo) Koa application, check out my koa-skeleton repository.

Example

Using koa-router for routing

const Koa = require('koa');
const Router = require('koa-router')
const bouncer = require('koa-bouncer');
const app = new Koa();
const router = new Router();

// extends the Koa context with some methods
app.use(bouncer.middleware());

// POST /users - create user endpoint
router.post('/users', async (ctx) => {

  // validate input

  ctx.validateBody('uname')
    .required('Username required')
    .isString()
    .trim()

  ctx.validateBody('email')
    .optional()
    .isString()
    .trim()
    .isEmail('Invalid email format')

  ctx.validateBody('password1')
    .required('Password required')
    .isString()
    .isLength(6, 100, 'Password must be 6-100 chars')

  ctx.validateBody('password2')
    .required('Password confirmation required')
    .isString()
    .eq(ctx.vals.password1, 'Passwords must match')

  // running database query last to give the other validations a chance to fail
  ctx.validateBody('uname')
    .check(await db.findUserByUname(ctx.vals.uname), 'Username taken')

  // if we get this far, then validation succeeded.

  // the validation populates a `ctx.vals` object with validated values
  //=> { uname: 'foo', password1: 'secret', password2: 'secret' }
  console.log(ctx.vals)

  const user = await db.insertUser({
    uname: ctx.vals.uname,
    email: ctx.vals.email,
    password: ctx.vals.password1
  })

  ctx.redirect(`/users/${user.id}`)
})

app
  .use(router.routes())
  .use(router.allowedMethods());
  

The general idea

The idea is that koa-bouncer exposes methods that transform and assert against user-input (form submissions, request bodies, query strings) within your routes.

If an assertion fails, then koa-bouncer throws a bouncer.ValidationError that you can catch upstream. For example, maybe you want to redirect back to the form, show a tip, and repopulate the form with the user's progress.

If validation succeeds, then you can access the validated/transformed parameters in a ctx.vals map that gets populated during validation.

Usage

First, you need to inject bouncer's middleware:

bouncer.middleware(opts)

This extends the Koa context with these methods for you to use in routes, the bulk of the koa-bouncer abstraction:

  • ctx.validateParam(key) => Validator
  • ctx.validateQuery(key) => Validator
  • ctx.validateBody(key) => Validator
  • ctx.check(value, [tip]) => throws ValidationError if falsey
  • ctx.checkNot(value, [tip]) => throws ValidationError if truthy

The first three methods return a validator that targets the value in the url param, query param, or body param that you specified with 'key'.

When you spawn a validator, it immediately populates ctx.vals[key] with the initial value of the parameter. You can then chain methods like .toString().trim().isEmail() to transform the value in ctx.vals and make assertions against it.

Just by calling these methods, they will begin populating ctx.vals:

router.get('/search', async (ctx) => {
  ctx.validateQuery('keyword')
  ctx.validateQuery('sort')
  ctx.body = JSON.stringify(ctx.vals)
})
curl http://localhost:3000/search
=> {}

curl http://localhost:3000/search?sort=age
=> { "sort": "age" }

We can use .required() to throw a ValidationError when the parameter is undefined. For example, we can decide that you must always supply a ?keyword= to our search endpoint.

And we can use .optional() to only run the chained validations/assertions if the parameter is undefined (not given by user) or if it is an empty string.

router.get('/search', async (ctx) => {
  ctx.validateQuery('keyword').required().isString().trim()
  ctx.validateQuery('sort').toArray()
  ctx.body = JSON.stringify(ctx.vals)
})
curl http://localhost:3000/search
=> Uncaught ValidationError

curl http://localhost:3000/search?keyword=hello
=> { "keyword": "hello", "sort": [] }

curl http://localhost:3000/search?keyword=hello&sort=age
=> { "keyword": "hello", "sort": ["age"] }

curl http://localhost:3000/search?keyword=hello&sort=age&sort=height
=> { "keyword": "hello", "sort": ["age", "height"] }

If a validation fails, then the validator throws a bouncer.ValidationError that we can catch with upstream middleware.

For example, we can decide that upon validation error, we redirect the user back to whatever the previous page was and populate a temporary flash object with the error and their parameters so that we can repopulate the form.

app.use(async (ctx, next) => {
  try {
    await next();
  } catch(err) {
    if (err instanceof bouncer.ValidationError) {
      ctx.flash = {
        message: ['danger', err.message],
        params: ctx.request.body
      };
      return ctx.redirect('back');
    }
    throw err;
  }
});

router.post('/users', async (ctx) => {
  ctx.validateBody('username')
    .required('Username is required')
    .isString()
    .trim()
    .isLength(3, 15, 'Username must be 3-15 chars');

  const user = await database.insertUser(ctx.vals.username);
  ctx.body = 'You successfully registered';
});
http --form POST localhost:3000/users
=> 302 Redirect to GET /users, message='Username is required'

http --form POST localhost:3000/users username=bo
=> 302 Redirect to GET /users, message='Username must be 3-15 chars'

http --form POST localhost:3000/users username=freeman
=> 200 OK, You successfully registered

You can pass options into the bouncer.middleware() function.

Here are the default ones:

app.use(bouncer.middleware({
  getParams({params}) { return params; },
  getQuery({query}) { return query; },
  getBody({request}) { return request.body; }
}));

You can override these if the validators need to look in a different place to fetch the respective keys when calling the validateParam, validateQuery, and validateBody methods.

You can always define custom validators via Validator.addMethod:

const Validator = require('koa-bouncer').Validator;

Validator.addMethod('isValidBitcoinAddress', function(tip = 'Invalid Bitcoin address') {
  // Will thread the tip through the nested assertions
  this
    .isString(tip)
    .trim()
    // Must be alphanumeric from start to finish
    .match(/^[a-z0-9]+$/i, tip)
    // But must not contain any of these chars
    .notMatch(/[0O1l]/, tip);

  return this;
});

Maybe put that in a custom_validations.js file and remember to load it.

Now you can use the custom validator method in a route or middleware:

ctx.validateBody('address')
  .required()
  .isValidBitcoinAddress();

These chains always return the underlying validator instance. You can access its value at any instant with .val().

const validator = ctx.validateBody('address')
  .required()
  .isValidBitcoinAddress();

console.log("current value of ctx.vals['address'] is", validator.val());

Here's how you'd write a validator method that transforms the underlying value:

Validator.addMethod('add10', function() {
  this.tap(val => val + 10);
  return this;
});

In other words, just use this.set(newVal) to update the object of validated params. And remember to return this so that you can continue chaining things on to the validator.

Validator methods

.val()

Returns the current value currently inside the validator.

router.get('/search', async (ctx) => {
  const validator1 = ctx.validateQuery('q').required();
  const validator2 = ctx.validateQuery('sort').optional();

  ctx.body = JSON.stringify([validator1.val(), validator2.val()]);
});
curl http://localhost:3000/search?q=hello&sort=created_at
// 200 OK ["hello", "created_at"]

I rarely use this method inside a route and prefer to access values from the ctx.vals object. So far I only use it internally when implementing validator functions.

.required([tip])

Only fails if val is undefined. Required the user to at least provie

ctx.validateBody('username')
  .required('Must provide username')

.optional()

If val is undefined or if it an empty string (after being trimmed) at this point, then skip over the rest of the methods.

This is so that you can validate a val only if user provided one.

ctx.validateBody('email')
  .optional()
  .isEmail('Invalid email format') // Only called if ctx.request.body is `undefined`
ctx.validateBody('email')
  .tap(x => '[email protected]')
  .optional()
  .isEmail()  // Always called since we are ensuring that val is always defined

Mutating ctx.vals to define a val inside an optional validator will turn off the validator's validator.isOptional() flag.

ctx.validateBody('email').optional();
ctx.vals.email = '[email protected]';
ctx.validateBody('email').isEmail();  // This will run

You can see the optional state of a validator with its .isOptional() method:

const validator = ctx.validateBody('email').optional();
console.log(validator.isOptional());  //=> true
ctx.vals.email = '[email protected]';
console.log(validator.isOptional());  //=> false
validator.isEmail();  // This will run

The reason koa-bouncer considers empty strings to be unset (instead of just undefined) is because the browser sends empty strings for text inputs. This is usually the behavior you want.

Also, note that .required() only fails if the value is undefined. It succeeds on empty string. This is also usually the behavior you want.

.isIn(array, [tip])

Ensure val is included in given array (=== comparison).

ctx.validateBody('role')
  .required('Must provide a role')
  .isIn(['banned', 'member', 'mod', 'admin'], 'Invalid role')

.isNotIn(array, [tip])

Ensure val is not included in given array (=== comparison).

ctx.validateBody('favorite-fruit')
  .isNotIn(['apple', 'pomegranate'], 'You cannot choose forbidden fruit')

.defaultTo(defaultVal)

If val is undefined, set it to defaultVal.

ctx.validateBody('multiplier')
  .defaultTo(1.0)
  .toFiniteFloat('multiplier must be a valid number')

.isString([tip])

Ensure val is a string.

Note: Also works with strings created via new String() where typeof new String() === 'object'.

ctx.validateBody('username')
  .isString()

It's a good practice to always call one of the .is* methods since they add explicit clarity to the validation step.

.isArray([tip])

Ensure val is an Array.

ctx.validateQuery('recipients')
  .isArray('recipients must be an array')
curl http://localhost:3000/?recipients=joey
=> ValidationError

curl http://localhost:3000/?recipients=joey&recipients=kate&recipients=max
=> 200 OK, ctx.vals => ['joey', 'kate', 'max']

Note: The previous example can be improved with .toArray.

ctx.validateQuery('recipients')
  .toArray()
  .isArray('recipients must be an array')
curl http://localhost:3000/?recipients=joey
=> 200 OK, ctx.vals.recipients => ['joey']

curl http://localhost:3000/?recipients=joey&recipients=kate&recipients=max
=> 200 OK, ctx.vals.recipients => ['joey', 'kate', 'max']

.eq(otherVal::Number, [tip])

Ensures val === otherVal.

ctx.validateBody('house-edge')
  .eq(0.01, 'House edge must be 1%')

.gt(otherVal::Number, [tip])

Ensures val > otherVal.

ctx.validateBody('hp')
  .gt(0, 'Player must have 1 or more hit points')

.gte(otherVal::Number, [tip])

Ensures val >= otherVal.

ctx.validateBody('age')
  .gte(18, 'Must be 18 or older')

.lt(otherVal::Number, [tip])

Ensures val < otherVal.

ctx.validateBody('pet-count')
  .lt(10, 'You must have fewer than 10 pets')

.lte(otherVal::Number, [tip])

Ensures val <= otherVal.

ctx.validateBody('house-edge')
  .lte(0.10, 'House edge cannot be higher than 10%')

.isLength(min:Int, max:Int, [tip])

Ensure val is a number min <= val <= max (inclusive on both sides).

ctx.validateBody('username')
  .required('Username required')
  .isString()
  .trim()
  .isLength(3, 15, 'Username must be 3-15 chars long')

.isInt([tip])

Ensures val is already an integer and that it is within integer range (Number.MIN_SAFE_INTEGER <= val <= Number.MAX_SAFE_INTEGER).

ctx.validateBody('age')
  .isInt('Age must be an integer')

.isFiniteNumber([tip])

Ensures that val is a number (float) but that it is not Infinity.

Note: This uses Number.isFinite(val) internally. Rather, it does not use the global isFinite(val) function because isFinite(val) first parses the number before checking if it is finite. isFinite('42') => true.

ctx.validateBody('num')
  .tap(n => Infinity)
  .isFiniteNumber()  // will always fail

.match(regexp::RegExp, [tip])

Ensures that val matches the given regular expression.

You must ensure that val is a string.

ctx.validateBody('username')
  .required('Username is required')
  .isString()
  .trim()
  .match(/^[a-z0-9_-]+$/i, 'Username must only contain a-z, 0-9, underscore, and hyphen')

Note: Remember to start your pattern with ^ ("start of string") and end your pattern with $ ("end of string") if val is supposed to fully match the pattern.

.notMatch(regexp::RegExp, [tip])

Ensure that val does not match the given regexp.

You must ensure that val is a string.

Note: It is often useful to chain .notMatch after a .match to refine the validation.

ctx.validateBody('username')
  .required('Username is required')
  .isString()
  .trim()
  .match(/^[a-z0-9_-]+$/i, 'Username must only contain a-z, 0-9, underscore, and hyphen')
  .notMatch(/admin/i, 'Username must not contain the word "admin" anywhere in it')
  .notMatch(/_{2,}/, 'Username must not contain consecutive underscores')
  .notMatch(/-{2,}/, 'Username must not contain consecutive hyphens')

.check(result, [tip]) and .checkNot(result, [tip])

Unlike most of the other validator methods, .check and .checkNot do not every look at the current val. They only look at the truthy/falseyness of the result you pass into them.

  • .check(result, [tip]) passes if result is truthy.
  • .checkNot(result, [tip]) passes if result is falsey.

They are a general-purpose tool for short-circuiting a validation, often based on some external condition.

Example: Ensure username is not taken:

ctx.validateBody('username')
  .required('Username required')
  .isString()
  .trim()
  .checkNot(await database.findUserByUsername(ctx.vals.uname), 'Username taken')

Example: Ensure that the email system is online only if they provide an email:

ctx.validateBody('email')
  .optional()
  .check(config.EMAIL_SYSTEM_ONLINE, 'Email system not ready, please try later')

.checkPred(fn, [tip]) and .checkPredNot(fn, [tip])

Pipes val into given fn and checks the result.

  • .checkPred(fn, [tip]) ensures that fn(val) returns truthy.
  • .checkPredNot(fn, [tip]) ensures that fn(val) returns falsey.

These methods are general-purpose tools that let you make your own arbitrary assertions on the val.

Example: Ad-hoc predicate function:

ctx.validateBody('num')
  .required()
  .toInt()
  .checkPred(n => n % 2 === 0, 'Your num must be divisible by two')

Example: Custom predicate function:

function isValidBitcoinAddress(addr) {
  // ...
}

ctx.validateBody('bitcoin-address')
  .required('Bitcoin address required')
  .isString()
  .trim()
  .checkPred(isValidBitcoinAddress, 'Invalid bitcoin address')

.isAlpha([tip])

Ensures that val is a string that contains only letters a-z (case insensitive).

ctx.validateBody('username')
  .required()
  .isString()
  .trim()
  .isAlpha()

.isAlphanumeric([tip])

Ensures that val is a string that contains only letters a-z (case insensitive) and numbers 0-9.

ctx.validateBody('username')
  .required()
  .isString()
  .trim()
  .isAlphanumeric()

.isNumeric([tip])

Ensures that val is a string that contains only numbers 0-9.

ctx.validateBody('serial-number')
  .required()
  .isString()
  .trim()
  .isNumeric()

.isAscii([tip])

Ensures that val is a string that contains only ASCII characters (https://es.wikipedia.org/wiki/ASCII).

In other words, val must only contain these characters:

! " # $ % & ' ( ) * + , - . / 0 1 2 3 4 5 6 7 8 9 : ; < = > ?
@ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z [ \ ] ^ _
` a b c d e f g h i j k l m n o p q r s t u v w x y z { | } ~
ctx.validateBody('command')
  .required()
  .isString()
  .trim()
  .isAscii()

.isBase64([tip])

Ensures that val is a base64-encoded string.

Note: An empty string ("") is considered valid.

ctx.validateBody('data')
  .required()
  .isString()
  .trim()
  .isBase64()

.isEmail([tip])

Ensures that val is a valid string email address.

ctx.validateBody('email')
  .optional()
  .isString()
  .trim()
  .isEmail()

.isHexColor([tip])

Ensures that val is a hex color string.

Accepts both 6-digit and 3-digit hex colors with and without a leading '#' char.

These are all valid: '#333333', '#333', 333333, 333.

ctx.validateBody('background-color')
  .required()
  .isString()
  .trim()
  .isHexColor()
  .tap(x => x.startsWith('#') ? x : '#' + x)

.isUuid([version::String], [tip])

Ensure that val is a valid uuid string.

version can be one of 'v3', 'v4', 'v5', 'all'. default is 'all'.

koa-bouncer can handle any of these:

.isUuid('v4', 'must be uuid v4');
.isUuid('must be any uuid');
.isUuid('v4');
.isUuid();
router.get('/things/:uuid', async (ctx) => {
  ctx.validateParam('uuid')
    .isUuid('v3')

  const thing = await database.findThing(ctx.vals.uuid);
});

.isJson([tip])

Ensures that val is a valid, well-formed JSON string.

Works by simply wrapping JSON.parse(val) with a try/catch.

ctx.validateBody('data')
  .isJson()

Methods that convert/mutate the val

.set(newVal)

Sets val to arbitrary value newVal.

Used internally by validator methods to update the value. Can't think of a reason you'd actually use it inside a route.

ctx.validateQuery('test')
  .set(42)
curl http://localhost:3000
// 200 OK, ctx.vals.test => 42

curl http://localhost:3000/?test=foo
// 200 OK, ctx.vals.test => 42

Note: .set(42) is equivalent to .tap(x => 42).

.toArray()

Converts val to an array if it is not already an array.

If val is not already an array, then it puts it into an array of one item.

If val is undefined, then sets it to empty array [].

ctx.validateQuery('friends')
  .toArray()
  .isArray()  // Always succeeds
curl http://localhost:3000/
// 200 OK, ctx.vals.friends => []

curl http://localhost:3000/?friends=joey
// 200 OK, ctx.vals.friends => ['joey']

curl http://localhost:3000/?friends=joey&friends=kate
// 200 OK, ctx.vals.friends => ['joey', 'kate']

.toInt([tip])

Parses and converts val into an integer.

Fails if val cannot be parsed into an integer or if it is out of safe integer range.

Uses parseInt(val, 10), so note that decimals and extraneous characters will be truncated off the end of the value.

ctx.validateQuery('age')
  .required('Must provide your age')
  .toInt('Invalid age')
curl http://localhost:3000/?age=42
// 200 OK, ctx.vals.age => 42

curl http://localhost:3000/?age=-42
// 200 OK, ctx.vals.age => -42 (parses negative integer)

curl http://localhost:3000/?age=42.123
// 200 OK, ctx.vals.age => 42 (truncation)

curl http://localhost:3000/?age=42abc
// 200 OK, ctx.vals.age => 42 (truncation)

curl http://localhost:3000/?age=9007199254740992
// ValidationError (out of integer range)

.toInts([tip])

Converts each string in val into an integer.

If val is undefined, sets it to empty array [].

Fails if any item cannot be parsed into an integer or if any parse into integers that are out of safe integer range.

ctx.validateQuery('guesses')
  .toInts('One of your guesses was invalid')
curl http://localhost:3000/
// 200 OK, ctx.vals.guesses => []

curl http://localhost:3000/?guesses=42
// 200 OK, ctx.vals.guesses => [42]

curl http://localhost:3000/?guesses=42&guesses=100
// 200 OK, ctx.vals.guesses => [42, 100]

curl http://localhost:3000/?guesses=42&guesses=100&guesses=9007199254740992
// ValidationError (out of safe integer range)

curl http://localhost:3000/?guesses=abc
// ValidationError (one guess does not parse into an int because it is alpha)

curl http://localhost:3000/?guesses=1.2345
// ValidationError (one guess does not parse into an int because it is a decimal)

.uniq

Removes duplicate items from val which must be an array.

You must ensure that val is already an array.

ctx.validateQuery('nums')
  .toArray()
  .toInts()
  .uniq()
curl http://localhost:3000/?nums=42
// 200 OK, ctx.vals.nums => [42]

curl http://localhost:3000/?nums=42&nums=42&nums=42
// 200 OK, ctx.vals.nums => [42]

.toBoolean()

Coerces val into boolean true | false.

Simply uses !!val, so note that these will all coerce into false:

  • Empty string ""
  • Zero 0
  • null
  • false
  • undefined
ctx.validateBody('remember-me')
  .toBoolean()

.toDecimal([tip])

Converts val to float, but ensures that it a plain ol decimal number.

In most application, you want this over .toFloat / .toFiniteFloat.

A parsed decimal will always pass a .isFiniteNumber() check.

ctx.validateBody('num')
  .toDecimal()
  .isFiniteNumber() // <-- Redundant

.toFloat([tip])

Converts val to float, throws if it fails.

Note: it uses Number.parseFloat(val) internally, so you will have to chain isFiniteNumber() after it if you don't want Infinity:

  • Number.parseFloat('Infinity') => Infinity
  • Number.parseFloat('5e3') => 5000
  • Number.parseFloat('1e+50') => 1e+50
  • Number.parseFloat('5abc') => 5
  • Number.parseFloat('-5abc') => -5
  • Number.parseFloat('5.123456789') => 5.123456789

Use .toDecimal instead of .toFloat when you only want to allow decimal numbers rather than the whole float shebang.

ctx.validateBody('num')
  .toFloat()
  .isFiniteNumber()

.toFiniteFloat([tip])

Shortcut for:

ctx.validateBody('num')
  .toFloat()
  .isFiniteNumber()

.toString()

Calls val.toString() or sets it to empty string "" if it is falsey.

Note: If val is truthy but does not have a .toString() method, like if val is Object.create(null), then koa-bouncer will break since this is undefined behavior that koa-bouncer does not want to make assumptions about.

TODO: Think of a use-case and then write an example.

.trim()

Trims whitespace off the left and right side of val which must be a string.

You almost always use this for string user-input (aside from passwords) since leading/trailing whitespace is almost always a mistake or extraneous.

You do not want to call it on the user's password since space is perfectly legal and if you trim user passwords you will hash a password that the user did not input.

koa-bouncer will break if you do not ensure that val is a string when you call .trim().

ctx.validateBody('username')
  .required()
  .isString()
  .trim();

.fromJson([tip])

Parses val into a JSON object.

Fails if it is invalid JSON or if it is not a string.

ctx.validateBody('data')
  .required()
  .fromJson()

.tap(fn, [tip])

Passes val into given fn and sets val to the result of fn(val).

General-purpose tool for transforming the val.

Almost all the validator methods that koa-bouncer provides are just convenience methods on top of .tap and .checkPred, so use these methods to implement your own logic as you please.

fn is called with this bound to the current validator instance.

tip is used if fn(val) throws a ValidationError error.

ctx.validateBody('direction')
  .required('Direction is required')
  .isString()
  .trim()
  .tap(x => x.toLowerCase())
  .isIn(['north', 'south', 'east', 'west'], 'Invalid direction')
curl http://localhost:3000/?direction=WeST
=> 200 OK, ctx.vals.direction => 'west'

.encodeBase64([tip])

Converts val string into base64 encoded string.

Empty string encodes to empty string.

ctx.vals.message = 'hello';

ctx.validateBody('message')
  .encodeBase64()
  .val(); //=> 'aGVsbG8='

.decodeBase64([tip])

Decodes val string from base64 to string.

Empty string decodes to empty string.

ctx.vals.message = 'aGVsbG8=';

ctx.validateBody('message')
  .decodeBase64()
  .val(); //=> 'hello'

.clamp(min::Number, max::Number)

Defines a number range that val is restricted to. If val exceeds this range in either direction, val is updated to the min or max of the range.

ie. If val < min, then val is set to min. If val > max, then val is set to max.

Note: You must first ensure that val is a number.

router.get('/users', async (ctx) => {
  ctx.validateQuery('per-page')
    .defaultTo(50)
    .toInt('per-page must be an integer')
    .clamp(10, 100);
});
curl http://localhost:3000/users
// 200 OK, ctx.vals['per-page'] === 50

curl http://localhost:3000/users?per-page=25
// 200 OK, ctx.vals['per-page'] === 25 (not clamped since it's in range)

curl http://localhost:3000/users?per-page=5
// 200 OK, ctx.vals['per-page'] === 10 (clamped to min)

curl http://localhost:3000/users?per-page=350
// 200 OK, ctx.vals['per-page'] === 100 (clamped to max)

Changelog

6.0.0

  • Supports Koa 2 by default instead of Koa 1.

5.1.0

  • Added .toFiniteFloat.

5.0.0

  • .optional() now considers empty strings (after trimming) to be unset instead of just undefined values.

License

MIT