iterable-typescript

iterable-ts

This library is still a work-in-progress!

Documentation: https://github.com/baileympearson/iterable-ts

What is iterable-ts?

iterable-ts is a simple, flexible library for working with syncronous iterators in Javascript. The intention is to provide a small set of functions that can operate on all iterators and iterables. There are other libraries that provide similar functionality, but I found none that satisfied the following design goals (so naturally, I reinvented the wheel)

Prioritizes reusability and composability

Many libraries, such as wu.js, rely on method chaining. This is inherently less composible than free functions. For example, consider the following using wu.js:

const source = [1, 2, 3]
const result = wu(source)
  .map(x => x * 2)
  .filter(x => x > 5)
for (const value in result) {
  console.log(result)
}

The source (wu(source)) is inherently coupled to the logic (map(x => x * 2).filter(x => x > 5)). Suppose we wanted to reuse the logic of filtering out all doubled values that are greater than 5. In wu, we'd have to abstract this to a function

function doubledGreaterThan5(source) {
  return wu(source)
    .map(x => x * 2)
    .filter(x => x > 5)
}

Compare this to the curried, free function version:

// declare the same function as before
const doubledGreaterThan5 = compose(
  filter(x => x > 5),
  map(x => x * 2)
)
const source = [1, 2, 3]
const result = doubledGreaterThan5(source)
for (const value in source) {
  console.log(result)
}

• Uses the concept of iterators as its base abstraction

Some libraries, such as rambda.js, use more complicated functional programming constructs as the base abstraction. For example, the map function in rambda.js takes a Functor and returns a "Functor of the same shape". In a flexible, interpreted language like Javascript, I wanted to model the library off of the approach of Clojure. Every standard function returns an iterable, which can be turned into native Javascript structures as desired.

map(x => x * 2, range(5)) // Generator<number>
into(
  'array',
  map(x => x * 2, range(5))
) // [0, 2, 4, 6, 8]

Focuses on simplicity

Many libraries attempt to solve multiple problems at once or introduce multiple abstractions. For example, IxJS provides operators for both synchronous and asynchronous iterators, as well as providing multiple functions such as of, to allow iterator operations on primitive values. There is nothing wrong with this approach. Instead, iterable-ts is focused purely on synchronous iterators.
It also does not provide functions to elevate non-iterable values into iterators (although, feel free to write these yourself if you'd like. They're easy to write using generator functions).

Installation

notes to follow here

Why use iterable-ts?

iterable-ts is intended to be more flexible than the native iterator operations. All the functions included operate solely on Iterables and therefore work on all iterable types. For example:

const array = [1, 2, 3, 4]
const set = new Set([1, 2, 3, 4])
function* iterable() {
  yield 1
  yield 2
  yield 3
  yield 4
}

const double = (x: number) => x * 2

// all of the following are valid
map(double, array)
map(double, set)
map(double, iterable)

Since all the core operations produce iterables, a utility function is provided to easily convert iterables into native Javascript data structures:

function* iterable() {
  yield 1
  yield 2
  yield 3
}
into('array', iterable()) // [1,2,3]
into('set', iterable()) // Set(1,2,3)

// coming soon as well
function* key_value_pairs() {
  yield ['name', 'john']
  yield ['age', 23]
  yield ['occupation', 'space ship pilot']
}

into('object', key_value_pairs()) // { name: 'john', age: 23, occupation: "space ship pilot" }

Examples

Here I attempt to illustrate the power versitility and power of functional-ts. Not all of these examples are as clean or immediatley understandable as their imperitive counterparts.

Print the first 50 fibonacci numbers

const fib = iterate(([a, b]) => [b, a + b], [0, 1])
const first_50_fib = take(50, fib)
for (const value of first_50_fib) {
  console.log(value)
}

Build testable, composible logic pipelines

/**
 * print the square of the first 10 even fibonacci numbers after the first 10 odd fibonacci numbers
 */

// `compose` composes functions backwards - so in this case the `map` is composed inside the `skipWhile`, and the `skipWhile` is composed inside the `scan`
const skipTenOdd = compose(
  map(([value, _]) => value),
  skipWhile(([value, count]) => count < 10),
  scan((n: number, accum: number) => {
    if (n % 2 === 0) {
      return accum
    }
    return accoum + 1
  }, 0)
)

const isEven = (x: number) => x % 2 === 0

// don't like how `compose` works backwards?  Pipe does the same thing as compose, but does it top-to-bottom
const pipeline = pipe(
  skipTenOdd,
  filter(isEven),
  take(10),
  map((x: number) => x * x)
)

const fib = iterate(([a, b]) => [b, a + b], [0, 1])

for (const value of pipeline(fib)) {
  console.log(value)
}

Build logic pipelines

interface User {
  name: string
  age: number
}

const firstOver23 = find((user: User) => user.age > 21)

// print NAME:age for all users over 21.  useful? probably not.
//    illustrative of the flexibility of compose? probably.
const pipeline = compose(
  map(({ name, age}: User) => `${name}:${age}`)
  map((user: User) => { ...user, name: user.name.toUpperCase() })
  firstOver32
)