ts-iterable-functions

Tree-shakeable, functional, type-safe library for lazy transformation of iterable sequences in TypeScript.

Usage no npm install needed!

<script type="module">
  import tsIterableFunctions from 'https://cdn.skypack.dev/ts-iterable-functions';
</script>

README

ts-iterable-functions

A collection of type-safe functions for operating over iterable sequences, with specialized versions that generate unary functions for use in pipes. Will feel immediately familiar for users of MS LINQ-to-objects.

npm version Build Status

Usage

First, import pipeInto from ts-functional-pipe:

import { pipeInto as pp } from 'ts-functional-pipe'

Let's make a collection of cars

const cars = [
  {
    manufacturer: 'Ford',
    model: 'Escort',
  },
  {
    manufacturer: 'Ford',
    model: 'Cortina',
  },
  {
    manufacturer: 'Renault',
    model: 'Clio',
  },
  {
    manufacturer: 'Vauxhall',
    model: 'Corsa',
  },
  {
    manufacturer: 'Ford',
    model: 'Fiesta',
  },
  {
    manufacturer: 'Fiat',
    model: '500',
  },
]

...and sort them by manufacturer, and then by model:

const orderedCars = pp(
  cars,
  orderBy((c) => c.manufacturer),
  thenBy((c) => c.model),
  toArray()
)

Or we could count the number of cars for each manufacturer:

const carsPerManufacturer = pp(
  cars,
  groupBy((c) => c.manufacturer),
  map((g) => ({
    count: _count(g),
    manufacturer: g.key,
  })),
  orderByDescending((c) => c.count),
  thenBy((c) => c.manufacturer)
)
for (var c of carsPerManufacturer) {
  console.log(`${c.manufacturer} : ${c.count}`)
}

to give

Ford : 3
Fiat : 1
Renault : 1
Vauxhall : 1

Functions for iterable sequences

Almost every function in this collection is designed to work over Iterable<T>.

An example

Let's start with map (which is also aliased to select) to see how it works.

Here's an Iterable<number>

const src = [1, 2, 3]

We can use the _map function to transform this as follows:

const times2 = _map(src, (x) => x + x)

Two forms of the same function

All of the functions that transform iterables in this library exist in two forms.

The first form is the one we used above and looks like this:

function _someOperator<T, A, B, R>(src: T, a: A, b: B): R

and by convention is prefixed with an _underscore. While handy in their own way, composing these functions is ugly.

//this looks awful
const times2squared = _map(
  _map(src, (x) => x + x),
  (x) => x * x
)

The pipeable function

If, instead, we had functions that look like this

function someOperator<T, A, B, R>(a: A, b: B): (src: T) => R

where the function returns a unary function of the form (src: T) => R, we can use them in pipes (where the output of one function is fed in to the input of the next function).

In fact, we can transform _someOperator into someOperator (preserving all type information) with the deferP0 function (from ts-functional-pipe):

const deferP0 = <P0, A extends any[], R>(fn: (src: P0, ...args: A) => R) => (...args: A) => (
  src: P0
): R => fn(src, ...args)

so, we could take the _map function above and transform it into the pipeable form with a simple call to deferP0(_map).

All functions that transform Iterable<T> in the library exist in the two forms. So, for instance the map function exists as _map and map. Moving forward, we'll be avoiding the _underscored functions.

Usage with pipes

The functions in this library are designed to be composed. Package ts-functional-pipe offers excellent type-inference for this purpose. There is good information about to use the pipe/pipeInto/compose functions it contains in the README over there.

Making times2squared readable

Let's use pipeInto (imported above as pp) to pipe our iterable into a chain of unary functions, generated (in this case) using the map function discussed above:

const src = [1, 2, 3]
const times2squared = pp(
  src,
  map((x) => x + x),
  map((x) => x * x)
)

Types in the pipe are preserved

Due to some funky type-definitions in ts-functional-pipe, types flow through the pipe nicely:

const src = [1, 2, 3]
const toStringRepeated = pp(
  src,
  map((x) => x.toString()), // here x is number
  map((s) => s + s) // here s is string
) // returns a string

and all types are correctly inferred.

More coming soon.

Generators

range, repeat, repeatGenerate

Transformers

aggregate, all/every, append, average, concat, count, defaultIfEmpty, distinctBy, distinct, elementAt, except, firstOrDefault, first, flatten, forEach, fullOuterGroupJoin, fullOuterJoin, groupAdjacent, groupBy, groupJoin, intersect, isSubsetOf, isSupersetOf, join, lastOrDefault, last, leftOuterJoin, maxBy, max, minBy, min, orderByDescending, orderBy, preprend, reduce, reduceRight, reverse, selectMany/flapMap, select/map, sequenceEqual, singleOrDefault, single, skip, skipWhile, some, sum, take, takeWhile, thenByDescending, thenBy, toArray, toLookup, toMap, toSet, union, where/filter, zipAll, zip, zipMap

acknowledgements

Created using the wonderful https://github.com/alexjoverm/typescript-library-starter.