0.0.1 • Published 2 years ago

@standardize/optional v0.0.1

Weekly downloads
-
License
MIT
Repository
-
Last release
2 years ago

@standardize/optional

AKA Optional(Java/OOP) Option(Scala/OCAML) Maybe(Haskell)

API Docs

Why is this different from other libraries?

Javascripts primary strength is its simple asynchronous code. This optional implementation is able to map into and out of the underlying AsyncOptional and Optional interfaces without the user needing to keep track of the async state of execution.

Other libraries like async-optional (also no TS definitions) require you to define that the optional will be async up front instead of being able to map into an async version at will.

Why?

Applying functions to potentially null/missing values has been the bane of many programs. The optional helps refactor potentially dangerous or verbose code into a check free style.

import { Optional } from 'designed'
import { storeUser, hashPassword, updateTimestamps, emptyUser, findUser } from './UserRepo'

function updateUserPassword(userId?: string, password: string) {
await Optional.of(userId)
		.mapAsync(findUser)
		.map(addHashedPasswordToUser(password))  // Perform a sync transformation
		.map(updateTimestamps)                   // Add a value
		.mapAsync(storeUser)                     // Store the user asynchronously
		.orElse(emptyUser())
}

updateUserPassword(undefined, '123')   // Resolves just fine
updateUserPassword(userId, '123')      // Also resolves!

Public Interface

isPresent / isAbsent

Acts as a type guard to check whether the value is present. This should always be used before issuing Optional.get() to prevent throwing exceptions.

if (maybeUser.isPresent()) {
  maybeUser; // PresentOptional<User>
}

filter / filterNot

If the predicate is true the value will remain a PresentOptional.

Optional.of("ABC").filter((s) => s.length > 3); // AbsentOptional<'ABC'>
Optional.of("ABC").filterNot((s) => s.length > 3); // PresentOptional<'ABC'>

Type narrowing is supported by supplying a User Defined Type Guard as the predicate.

const optional =
  Optional.of<string | number>(123) // Optional<string | number>

const isString =
  (v: Object): v is string => typeof v === 'string'

optional.filter(isString) // Optional<string>
optional.filterNot(isString) // Optional<number>

map

Applies a function to the wrapped value if it is present. If the return value is absent (null or undefined) it will return an AbsentOptional

Optional.of("ABC")
  .map((s) => s.toLowerCase())
  .get(); // 'abc'

Optional.of < string > null.map((s) => s.toLowerCase()).orElse("not there"); // 'not there'

flatMap

Applies a function to the wrapped value that returns another Optional

const readFileDataAtPathIfExists = (filepath: string): Optional<Buffer> => {
  /* implement */
};

Optional.of("./realfile.md").map(readFileDataAtPathIfExists); // Optional<Optional<Buffer>>
Optional.of("./realfile.md").flatMap(readFileDataAtPathIfExists); // Optional<Buffer>
Optional.of("not a file").flatMap(readFileDataAtPathIfExists); // Optional<Buffer>

toResult

Converts the Optional<T> to a Result<T, null>

Zip Methods

404 Documentation not complete

Unwrapping Methods

orElse

Return the wrapped value or "else" the value passed to the function.

Optional.of("a string").orElse(123) // Typed: string | number Result: "a string"
Optional.empty().orElse(123) // Result: 123

orGet

Invokes the passed method and returns its value if the optional value is absent.

Optional.of("a string").orElse(() => 123) // Typed: string | number Result: "a string"

This is particularly useful in async methods. As any value can have await applied to it, you can easily setup fallback behaviour without requiring branching.

await Optional.of(userId).mapAsync(findUserAsync).orElse(createUserAsync); // Will always return an `User`. createUserAsync is only invoked if the previous value is empty

orThrow

If the value is missing, throws the error supplied by the passed function. If an error is not returned from orThrow a TypeError will be thrown instead.

Optional.empty().orThrow(() => new Error("was empty")); // throws Error
Optional.of(1).orThrow(() => new Error("was empty")); // 1

Async Methods

Optionals can traverse back and forth between being in an async and sync state. AsyncOptional is a PromiseLike class that can be await'ed to make your optionals easier to use with async code.

All methods available on sync Optional's are also available when the optional is async.

mapAsync

Performs a similar function to the above but returns an AsyncOptional depending on whether the value existed. This object is await'able and is used instead of a promise.

const fetchUserFromAPIById = (id: string) => fetch(/* details */);

await Optional.of("123").mapAsync(fetchUserFromAPIById).orElse("notThere"); // Fetch response

await Optional.empty().mapAsync(fetchUserFromAPIById).orElse("notThere"); // "notThere"

flatMapAsync

Performs a similar function to mapAsync and flatMap. It expects a function that returns a Promise<Optional<T>> and will flat map the Optional<T> into itself.

const searchForBlogPostByTitle = async (
  title: string
): Promise<Optional<BlogPost>> => {
  /* implement */
};

await Optional.of("Big Software").flatMapAsync(searchForBlogPostByTitle); // Resolves to Optional<BlogPost>

await Optional.empty().flatMapAsync(searchForBlogPostByTitle); // Also resolves to Optional<BlogPost>
0.0.1

2 years ago

0.0.0

2 years ago