@poppinss/utils

Handy utilities for repetitive work

Usage no npm install needed!

<script type="module">
  import poppinssUtils from 'https://cdn.skypack.dev/@poppinss/utils';
</script>

README

Utils

Collection of reusable scripts used by AdonisJS core team

gh-workflow-image typescript-image npm-image license-image synk-image

This module exports a collection of re-usable utilties to avoid re-writing the same code in every other package. We also include a handful of Lodash utilities, which are used across the AdonisJS packages eco-system.

Version 3.0 breaking changes

The version 3.0 re-format the exports to expose an "helpers" subpath to be used within the AdonisJS apps as well.

The idea is to separate helpers that we need to share with AdonisJS core inside its own module, accessible as ‌@poppinss/utils/build/helpers.

Inside helpers subpath

Following modules are now moved to a subpath.

  • MessageBuilder
  • base64
  • compose
  • fsReadAll
  • interpolate
  • requireAll
  • resolveDir
  • resolveFrom
// Earlier
import {
  MessageBuilder,
  base64,
  compose,
  fsReadAll,
  interpolate,
  requireAll,
  resolveDir,
  resolveFrom,
  safeEqual
} from '@poppinss/utils'

// After version 3.0
import {
MessageBuilder,
base64,
compose,
fsReadAll,
interpolate,
requireAll,
resolveDir,
resolveFrom,
safeEqual
} from '@poppinss/utils/build/helpers'

randomString

The randomString is now part of the string helpers.

// Earlier
import { randomString } from '@poppinss/utils'
randomString(32)

// After version 3.0
import { string } from '@poppinss/utils/build/helpers'
string.generateRandom(32)

lodash

The following lodash functions have been removed with new alternatives.

  • snakeCase
  • camelCase
  • startCase
// Earlier
import { lodash } from '@poppinss/utils'

lodash.snakeCase()
lodash.camelCase()
lodash.startCase()

// After version 3.0
import { string } from '@poppinss/utils/build/helpers'

string.snakeCase()
string.camelCase()
string.titleCase()

Table of contents

Installation

Install the package from npm registry as follows:

npm i @poppinss/utils

# yarn
yarn add @poppinss/utils

and then use it as follows:

import { requireAll } from '@poppinss/utils'
requireAll(__dirname)

Exception

A custom exception class that extends the Error class to add support for defining status and error codes.

import { Exception } from '@poppinss/utils'

throw new Exception('Something went wrong', 500, 'E_RUNTIME_EXCEPTION')
throw new Exception('Route not found', 404, 'E_ROUTE_NOT_FOUND')

esmRequire

Utility to require script files wihtout worrying about CommonJs and ESM exports. This is how it works.

  • Returns the exported value for module.exports.
  • Returns the default value is an ESM module has export default.
  • Returns all exports if is an ESM module and doesn't have export default.

foo.js

module.exports = {
  greeting: 'Hello world',
}

foo.default.js

export default {
  greeting: 'Hello world',
}

foo.esm.js

export const greeting = {
  greeting: 'hello world',
}

import { esmRequire } from '@poppinss/utils'

esmRequire('./foo.js') // { greeting: 'hello world' }
esmRequire('./foo.default.js') // { greeting: 'hello world' }
esmRequire('./foo.esm.js') // { greeting: { greeting: 'hello world' } }

esmResolver

The esmResolver method works similar to esmRequire. However, instead of requiring the file, it accepts the object and returns the exported as per the same logic defined above.

import { esmRequire } from '@poppinss/utils'

esmResolver({ greeting: 'hello world' }) // { greeting: 'hello world' }

esmResolver({
  default: { greeting: 'hello world' },
  __esModule: true,
}) // { greeting: 'hello world' }

esmResolver({
  greeting: { greeting: 'hello world' },
  __esModule: true,
}) // { greeting: { greeting: 'hello world' } }

Lodash utilities

Lodash itself is a bulky library and most of the times, we don't need all the functions from it.

Also, all of the lodash functions are published as individual modules on npm. However, most of those individual packages are outdated and using them is not an option.

Instead, we decided to pick some individual utilities that we need inside AdonisJS ecosystem and export them from the lodash package, as each function is exposed in its own separate file.

import { lodash } from '@poppinss/utils'
lodash.get({ name: 'virk' }, 'name') // virk

Exported methods

Following is the list of exported helpers.

Safe stringify

Similar to JSON.stringify, but also handles Circular references by removing them.

import { safeStringify } from '@poppinss/utils'

const o = { b: 1, a: 0 }
o.o = o

console.log(safeStringify(o))
// { "b":1,"a":0 }

console.log(JSON.stringify(o))
// TypeError: Converting circular structure to JSON

Safe parse

Similar to JSON.parse, but protects against Prototype Poisoning

import { safeParse } from '@poppinss/utils'

const input = '{ "user": { "__proto__": { "isAdmin": true } } }'

JSON.parse(input)
// { user: { __proto__: { isAdmin: true } } }

safeParse(input)
// { user: {} }

defineStaticProperty

Explicitly define static properties on a class by checking for hasOwnProperty. In case of inheritance, the properties from the parent class are cloned vs following the prototypal inheritance.

We use/need this copy from parent class behavior a lot in AdonisJS. Here's an example of Lucid models

You create an application wide base model

class AppModel extends BaseModel {
  @column.datetime()
  public createdAt: DateTime
}

AdonisJS will create the $columnDefinitions property on the AppModel class, that holds all the columns

AppModel.$columnDefinitions // { createdAt: { columName: created_at } }

Now, lets create another model inheriting the AppModel

class User extends AppModel {
  @column()
  public id: number
}

As per the Javascript prototypal inheritance. The User model will not contain the columns from the AppModel, because we just re-defined the $columnDefinitions property. However, we don't want this behavior and instead want to copy the columns from the AppModel and then add new columns to it.

Voila! Use the defineStaticProperty helper from this class.

class LucidBaseModel {
  static boot() {
    defineStaticProperty(this, LucidBaseModel, {
      propertyName: '$columnDefinitions',
      defaultValue: {},
      strategy: 'inherit',
    })
  }
}

The defineStaticProperty takes a total of three arguments.

  • The first argument is always this.
  • The second argument is the root level base class. This will usually be the class exported by your package or module.
  • The third argument takes the propertyName, defaultValue (in case, there is nothing to copy), and the strategy.
  • The inherit strategy will copy the properties from the base class.
  • The define strategy will always use the defaultValue to define the property on the class. In other words, there is no copy behavior, but prototypal inheritance chain is also breaked by explicitly re-defining the property.

flatten

Flatten an object/array. The method wraps the flattie package.

import { flatten } from '@poppinss/utils'

flatten({
  a: 'hi',
  b: {
    a: null,
    b: ['foo', '', null, 'bar'],
    d: 'hello',
    e: {
      a: 'yo',
      b: undefined,
      c: 'sup',
      d: 0,
      f: [
        { foo: 123, bar: 123 },
        { foo: 465, bar: 456 },
      ]
    }
  },
  c: 'world'
});
// {
//   'a': 'hi',
//   'b.b.0': 'foo',
//   'b.b.1': '',
//   'b.b.3': 'bar',
//   'b.d': 'hello',
//   'b.e.a': 'yo',
//   'b.e.c': 'sup',
//   'b.e.d': 0,
//   'b.e.f.0.foo': 123,
//   'b.e.f.0.bar': 123,
//   'b.e.f.1.foo': 465,
//   'b.e.f.1.bar': 456,
//   'c': 'world'
// }

Helpers

The helpers module is also available in AdonisJS applications as follows:

import { fsReadAll, string, types } from '@ioc:Adonis/Core/Helpers'

The @poppinss/utils exposes this module as follows

import { fsReadAll, string, types } from '@poppinss/utils/build/helpers'

fsReadAll

A utility to recursively read all script files for a given directory. This method is equivalent to readdir + recursive + filter (.js, .json, .ts).

import { fsReadAll } from '@poppinss/utils/build/helpers'

const files = fsReadAll(__dirname) // array of strings

You can also define your custom filter function. The filter function must return true for files to be included.

const files = fsReadAll(__dirname, (file) => {
  return file.endsWith('.foo.js')
})

requireAll

Same as fsReadAll, but instead require the files. Helpful when you want to load all the config files inside a directory on app boot.

import { requireAll } from '@poppinss/utils/build/helpers'
const config = requireAll(join(__dirname, 'config'))

{
  file1: {}, // exported object
  file2: {} // exported object
}

The method also accepts the following options

requireAll(join(__dirname, 'config'), recursive, optional, filter)
  • recursive Load all files recursively. Defaults to true.
  • optional Do not raise exception when the root directory is missing. Defaults to false.
  • filter Cherry pick files to require. By default, all JavaScript, TypeScript and JSON files are required.

resolveFrom

Works similar to require.resolve, however it handles the absolute paths properly.

import { resolveFrom } from '@poppinss/utils/build/helpers'

resolveFrom(__dirname, 'npm-package') // returns path to package "main" file
resolveFrom(__dirname, './foo.js') // returns path to `foo.js` (if exists)
resolveFrom(__dirname, join(__dirname, './foo.js')) // returns path to `foo.js` (if exists)

resolveDir

The require.resolve or resolveFrom method can only resolve paths to a given file and not the directory. For example: If you pass path to a directory, then it will search for index.js inside it and in case of a package, it will be search for main entry point.

On the other hand, the resolveDir method can also resolve path to directories using following resolution.

  • Absolute paths are returned as it is.
  • Relative paths starting with ./ or .\ are resolved using path.join.
  • Path to packages inside node_modules are resolved as follows: - Uses require.resolve to resolve the package.json file. - Then replace the package-name with the absolute resolved package path.
import { resolveDir } from '@poppinss/utils/build/helpers'

resolveDir(__dirname, './database/migrations')
// __dirname + /database/migrations

resolveDir(__dirname, 'some-package/database/migrations')
// {path-to-package}/database/migrations

resolveDir(__dirname, '@some/package/database/migrations')
// {path-to-package}/database/migrations

interpolate

A small utility function to interpolate values inside a string.

import { interpolate } from '@poppinss/utils/build/helpers'

interpolate('hello {{ username }}', {
  username: 'virk',
})

interpolate('hello {{ users.0.username }}', {
  users: [{ username: 'virk' }],
})

If value is missing, it will be replaced with an 'undefined' string.

Use the \ to escape a mustache block from getting evaluated.

import { interpolate } from '@poppinss/utils/build/helpers'

interpolate('\\{{ username }} expression evaluates to {{ username }}', {
  username: 'virk',
})
// Output: {{ username }} expression evaluates to virk

Base 64 Encode/Decode

Following helpers for base64 encoding/decoding also exists.

encode

import { base64 } from '@poppinss/utils/build/helpers'

base64.encode('hello world')
base64.encode(Buffer.from('hello world', 'binary'))

decode

import { base64 } from '@poppinss/utils/build/helpers'

base64.decode(base64.encode('hello world'))
base64.decode(base64.encode(Buffer.from('hello world', 'binary')), 'binary')

urlEncode

Same as encode, but safe for URLS and Filenames

urlDecode

Same as decode, but decodes the urlEncode output values

Safe equal

Compares two values by avoid timing attack. Accepts any input that can be passed to Buffer.from

import { safeValue } from '@poppinss/utils/build/helpers'

if (safeValue('foo', 'foo')) {
}

Message Builder

Message builder provides a sane API for stringifying objects similar to JSON.stringify but has a few advantages.

  • It is safe from JSON poisoning vulnerability.
  • You can define expiry and purpose for the encoding. The verify method will respect these values.

The message builder alone may seem useless, since anyone can decode the object and change its expiry or purpose. However, you can generate an hash of the stringified object and verify the tampering by validating the hash. This is what AdonisJS does for cookies.

import { MessageBuilder } from '@poppinss/utils/build/helpers'

const builder = new MessageBuilder()
const encoded = builder.build({ username: 'virk' }, '1 hour', 'login')

Now verify it

builder.verify(encoded) // returns null, no purpose defined
builder.verify(encoded, 'register') // returns null, purpose mismatch.
builder.verify(encoded, 'login') // return { username: 'virk' }

compose

Javascript doesn't have a concept of inherting multiple classes together and neither does Typescript. However, the official documentation of Typescript does talks about the concept of mixins.

As per the Typescript docs, you can create and apply mixins as follows.

type Constructor = new (...args: any[]) => any

const UserWithEmail = <T extends Constructor>(superclass: T) => {
  return class extends superclass {
    public email: string
  }
}

const UserWithPassword = <T extends Constructor>(superclass: T) => {
  return class extends superclass {
    public password: string
  }
}

class BaseModel {}
class User extends UserWithPassword(UserWithEmail(BaseModel)) {}

Mixins are close to a perfect way of inherting multiple classes. I recommend reading this article for same.

However, the syntax of applying multiple mixins is kind of ugly, as you have to apply mixins over mixins, creating a nested hierarchy as shown below.

UserWithAttributes(UserWithAge(UserWithPassword(UserWithEmail(BaseModel))))

The compose method is a small utility to improve the syntax a bit.

import { compose } from '@poppinss/utils/build/helpers'

class User extends compose(
  BaseModel,
  UserWithPassword,
  UserWithEmail,
  UserWithAge,
  UserWithAttributes
) {}

Mixins gotchas

Typescript has an open issue related to the constructor arguments of the mixin class or the base class.

Typescript expects all classes used in the mixin chain to have a constructor with only one argument of ...args: any[]. For example: The following code will work fine at runtime, but the typescript compiler complains about it.

class BaseModel {
  constructor(name: string) {}
}

const UserWithEmail = <T extends typeof BaseModel>(superclass: T) => {
  return class extends superclass {
    // ERROR: A mixin class must have a constructor with a single rest parameter of type 'any[]'.ts(2545)
    public email: string
  }
}

class User extends compose(BaseModel, UserWithEmail) {}

You can work around this by overriding the constructor of the base class.

import { NormalizeConstructor, compose } from '@poppinss/utils/build/helpers'

const UserWithEmail = <T extends NormalizeConstructor<typeof BaseModel>>(superclass: T) => {
  return class extends superclass {
    public email: string
  }
}

string

The string module includes a bunch of helper methods to work with strings.

camelCase

Convert a string to its camelCase version.

import { string } from '@poppinss/utils/build/helpers'

string.camelCase('hello-world') // helloWorld

snakeCase

Convert a string to its snake_case version.

import { string } from '@poppinss/utils/build/helpers'

string.snakeCase('helloWorld') // hello_world

dashCase

Convert a string to its dash-case version. Optionally, you can also capitalize the first letter of each segment.

import { string } from '@poppinss/utils/build/helpers'

string.dashCase('helloWorld') // hello-world
string.dashCase('helloWorld', { capitalize: true }) // Hello-World

pascalCase

Convert a string to its PascalCase version.

import { string } from '@poppinss/utils/build/helpers'

string.pascalCase('helloWorld') // HelloWorld

capitalCase

Capitalize a string

import { string } from '@poppinss/utils/build/helpers'

string.capitalCase('helloWorld') // Hello World

sentenceCase

Convert string to a sentence

import { string } from '@poppinss/utils/build/helpers'

string.sentenceCase('hello-world') // Hello world

dotCase

Convert string to its dot.case version.

import { string } from '@poppinss/utils/build/helpers'

string.dotCase('hello-world') // hello.world

noCase

Remove all sorts of casing

import { string } from '@poppinss/utils/build/helpers'

string.noCase('hello-world') // hello world
string.noCase('hello_world') // hello world
string.noCase('helloWorld') // hello world

titleCase

Convert a sentence to title case

import { string } from '@poppinss/utils/build/helpers'

string.titleCase('Here is a fox') // Here Is a fox

pluralize

Pluralize a word.

import { string } from '@poppinss/utils/build/helpers'

string.pluralize('box') // boxes
string.pluralize('i') // we

You can also define your own irregular rules using the string.defineIrregularRule method.

  • The first argument is the singular variation
  • The second argument is the plural variation
import { string } from '@poppinss/utils/build/helpers'

string.defineIrregularRule('auth', 'auth')
string.plural('auth') // auth

You can also define your own uncountable rules using the string.defineUncountableRule method.

import { string } from '@poppinss/utils/build/helpers'

string.defineUncountableRule('login')
string.plural('login') // home

truncate

Truncate a string after a given number of characters

import { string } from '@poppinss/utils/build/helpers'

string.truncate('This is a very long, maybe not that long title', 12) // This is a ve...

By default, the string is truncated exactly after the given characters. However, you can instruct the method to wait for the words to complete.

string.truncate('This is a very long, maybe not that long title', 12, {
  completeWords: true,
}) // This is a very...

Also, it is possible to customize the suffix.

string.truncate('This is a very long, maybe not that long title', 12, {
  completeWords: true,
  suffix: ' <a href="/1"> Read more </a>',
}) // This is a very <a href="/1"> Read more </a>

excerpt

The excerpt method is same as the truncate method. However, it strips the HTML from the string.

import { string } from '@poppinss/utils/build/helpers'

string.excerpt('<p>This is a <strong>very long</strong>, maybe not that long title</p>', 12) // This is a very...

condenseWhitespace

Condense whitespaces from a given string. The method removes the whitespace from the left, right and multiple whitespace in between the words.

import { string } from '@poppinss/utils/build/helpers'

string.condenseWhitespace(' hello  world ')
// hello world

escapeHTML

Escape HTML from the string

import { string } from '@poppinss/utils/build/helpers'

string.escapeHTML('<p> foo © bar </p>')
// &lt;p&gt; foo © bar &lt;/p&gt;

Additonally, you can also encode non-ascii symbols

import { string } from '@poppinss/utils/build/helpers'

string.escapeHTML('<p> foo © bar </p>', {
  encodeSymbols: true,
})
// &lt;p&gt; foo &#xA9; bar &lt;/p&gt;

encodeSymbols

Encode symbols. Checkout he for available options

import { string } from '@poppinss/utils/build/helpers'

string.encodeSymbols('foo © bar')
// foo &#xA9; bar

toSentence

Join an array of words with a separator.

import { string } from '@poppinss/utils/build/helpers'

string.toSentence(['route', 'middleware', 'controller']) // route, middleware, and controller

string.toSentence(['route', 'middleware']) // route and middleware

You can also customize

  • separator: The value between two words except the last one
  • pairSeparator: The value between the first and the last word. Used, only when there are two words
  • lastSeparator: The value between the second last and the last word. Used, only when there are more than two words
string.toSentence(['route', 'middleware', 'controller'], {
  separator: '/ ',
  lastSeparator: '/or ',
}) // route/ middleware/or controller

prettyBytes

Convert bytes value to a human readable string. For options, recommend the bytes package.

import { string } from '@poppinss/utils/build/helpers'

string.prettyBytes(1024) // 1KB
string.prettyBytes(1024, { unitSeparator: ' ' }) // 1 KB

toBytes

Convert human readable string to bytes. This method is the opposite of the prettyBytes method.

import { string } from '@poppinss/utils/build/helpers'

string.toBytes('1KB') // 1024

prettyMs

Convert time in milliseconds to a human readable string

import { string } from '@poppinss/utils/build/helpers'

string.prettyMs(60000) // 1min
string.prettyMs(60000, { long: true }) // 1 minute

toMs

Convert human readable string to milliseconds. This method is the opposite of the prettyMs method.

import { string } from '@poppinss/utils/build/helpers'

string.toMs('1min') // 60000

ordinalize

Ordinalize a string or a number value

import { string } from '@poppinss/utils/build/helpers'

string.ordinalize(1) // 1st
string.ordinalize(99) // 99th

generateRandom

Generate a cryptographically strong random string

import { string } from '@poppinss/utils/build/helpers'

string.generateRandom(32)

isEmpty

Find if a value is empty. Also checks for empty strings with all whitespace

import { string } from '@poppinss/utils/build/helpers'

string.isEmpty('') // true
string.isEmpty('      ') // true

Types

The types module allows distinguishing between different Javascript datatypes. The typeof returns the same type for many different values. For example:

typeof {} // object
typeof [] // object
typeof null // object

WHAT??? Yes, coz everything is an object in Javascript. To have better control, you can make use of the types.lookup method.

lookup

Returns a more accurate type for a given value.

import { types } from '@poppinss/utils/build/helpers'

types.lookup({}) // object
types.lookup([]) // array
types.lookup(Object.create(null)) // object
types.lookup(null) // null
types.lookup(function () {}) // function
types.lookup(class Foo {}) // class
types.lookup(new Map()) // map

isNull

Find if the given value is null

import { types } from '@poppinss/utils/build/helpers'

types.isNull(null)) // true

isBoolean

Find if the given value is a boolean

import { types } from '@poppinss/utils/build/helpers'

types.isBoolean(true)) // true

isBuffer

Find if the given value is a buffer

import { types } from '@poppinss/utils/build/helpers'

types.isBuffer(new Buffer())) // true

isNumber

Find if the given value is a number

import { types } from '@poppinss/utils/build/helpers'

types.isNumber(100)) // true

isString

Find if the given value is a string

import { types } from '@poppinss/utils/build/helpers'

types.isString('hello')) // true

isArguments

Find if the given value is an arguments object

import { types } from '@poppinss/utils/build/helpers'

function foo() {
  types.isArguments(arguments)) // true
}

isObject

Find if the given value is a plain object

import { types } from '@poppinss/utils/build/helpers'

types.isObject({})) // true

isDate

Find if the given value is a date object

import { types } from '@poppinss/utils/build/helpers'

types.isDate(new Date())) // true

isArray

Find if the given value is an array

import { types } from '@poppinss/utils/build/helpers'

types.isArray([1, 2, 3])) // true

isRegexp

Find if the given value is an regular expression

import { types } from '@poppinss/utils/build/helpers'

types.isRegexp(/[a-z]+/)) // true

isError

Find if the given value is an instance of the error object

import { types } from '@poppinss/utils/build/helpers'
import { Exception } from '@poppinss/utils'

types.isError(new Error('foo'))) // true
types.isError(new Exception('foo'))) // true

isFunction

Find if the given value is a function

import { types } from '@poppinss/utils/build/helpers'

types.isFunction(function foo() {})) // true

isClass

Find if the given value is a class constructor. Uses regex to distinguish between a function and a class.

import { types } from '@poppinss/utils/build/helpers'

class User {}

types.isClass(User) // true
types.isFunction(User) // true

isInteger

Find if the given value is an integer.

import { types } from '@poppinss/utils/build/helpers'

types.isInteger(22.0) // true
types.isInteger(22) // true
types.isInteger(-1) // true
types.isInteger(-1.0) // true

types.isInteger(22.1) // false
types.isInteger(0.3) // false
types.isInteger(-0.3) // false

isFloat

Find if the given value is an float number.

import { types } from '@poppinss/utils/build/helpers'

types.isFloat(22.1) // true
types.isFloat(-22.1) // true
types.isFloat(0.3) // true
types.isFloat(-0.3) // true

types.isFloat(22.0) // false
types.isFloat(-22.0) // false
types.isFloat(-22) // false

isDecimal

Find if the given value has a decimal. The value can be a string or a number. The number values are casted to a string by calling the toString() method on the value itself.

The string conversion is peformed to test the value against a regex. Since, there is no way to natively find a decimal value in Javascript.

import { types } from '@poppinss/utils/build/helpers'

types.isDecimal('22.10') // true
types.isDecimal(22.1) // true

types.isDecimal('-22.10') // true
types.isDecimal(-22.1) // true

types.isDecimal('.3') // true
types.isDecimal(0.3) // true

types.isDecimal('-.3') // true
types.isDecimal(-0.3) // true

types.isDecimal('22.00') // true
types.isDecimal(22.0) // false (gets converted to 22)

types.isDecimal('-22.00') // true
types.isDecimal(-22.0) // false (gets converted to -22)

types.isDecimal('22') // false
types.isDecimal(22) // false

types.isDecimal('0.0000000000001') // true
types.isDecimal(0.0000000000001) // false (gets converted to 1e-13)

ObjectBuilder

A very simple class to conditionally builder an object. Quite often, I create a new object from an existing one and wants to avoid writing undefined values to it. For example

const obj = {
  ...(user.username ? { username: user.username } : {}),
  ...(user.id ? { id: user.id } : {}),
  ...(user.createdAt ? { createdAt: user.createdAt.toString() } : {}),
}

Not only the above code is harder to write. It is performance issues as well, since we are destructuring too many objects.

To address this use case, you can make use of the ObjectBuilder class as follows

import { ObjectBuilder } from '@poppinss/utils/build/helpers'

const obj = new ObjectBuilder()
  .add('username', user.username)
  .add('id', user.id)
  .add('createdAt', user.createdAt && user.createdAt.toString()).value // returns the underlying object

The add method ignores the value if its undefined. So it never gets added to the object at all. You can also ignore null properties by passing a boolean flag to the constructor.

new ObjectBuilder(true) // ignore null as well