Underscore.go

Move Fast; Optimize Late

A useful collection of Go utilities. Designed for programmer happiness.

TL;DR Sort-of like underscore.js, but for Go

API Documentation

GoDoc is WorkInProgress

:warning: Warning

This package is in heavy flux at the moment as I work to incorporate feedback from various sources.

:squirrel: Todo

• godoc
• contains
• indexOf
• worker pools
• parallel each
• parallel map with worker pool
• refactor to make functions first parameter (eg Each func(func(A), []A))
• handle maps & slices
• all
• any
• none

• Notes on Typed Functions
• Any
• Each
• Every
• Map
• None

Typed Functions

Any

Each

Each func(func(A int), []A) Each func(func(A B), []A)

Applies the given iterator function to each element of a collection (slice or map).

If the collection is a Slice, the iterator function arguments are value, index

If the collection is a Map, the iterator function arguments are value, key

EachP is a Parallel implementation of Each and concurrently applies the given iterator function to each element of a collection (slice or map).

  // var Each func(func(value interface{}, i interface{}), interface{})

  var buffer bytes.Buffer

  fn := func(s, i interface{}) {
    buffer.WriteString(s.(string))
  }

  s := []string{"a", "b", "c", "d", "e"}
  Each(fn, s)

  expect := "abcde"

  e := un.Each(fn, s)

  fmt.Printf("%#v\n", e) //"abcde"

Typed Each can be defined using a function type and the MakeEach helper.

Using a Typed Slice

  var EachInt func(func(value, i int), []int)
  MakeEach(&EachInt)

  var sum int

  fn := func(v, i int) {
    sum += v
  }

  i := []int{1, 2, 3, 4, 5}
  EachInt(fn, i)

  fmt.Printf("%#v\n", sum) //15

Using a Typed Map

  var EachStringInt func(func(key string, value int), map[string]int)
  var sum int

  fn := func(v int, k string) {
    sum += v
  }

  m := map[string]int{"a": 1, "b": 2, "c": 3, "d": 4, "e": 5}
  EachStringInt(fn, m)

  fmt.Printf("%#v\n", sum) //15

Of note is the ability to close over variables within the calling scope.

Every

Map

Map func([]A, func(A) B) []B

Applies the given function to each element of a slice, returning a slice of results

The base Map function accepts interface{} types and returns []interface{}

  // Map func(interface{}, func(interface{}) interface{}) []interface{}

  s := []string{"a", "b", "c", "d"}

  fn := func(s interface{}) interface{} {
    return s.(string) + "!"
  }

  m := un.Map(ToI(s), fn)
  fmt.Println(m) //["a!", "b!", "c!", "d!"]

Typed Maps can be defined using a function type and the MakeMap helper.

  Map func([]A, func(A) B) []B

  var SMap func([]string, func(string) string) []string
  un.MakeMap(&SMap)

  m := un.SMap(s, fn)
  fmt.Println(m) //["a!", "b!", "c!", "d!"]

Of note is the return value of Map is a slice of the return type of the applied function.

Partition

Partition func([]A, func(A) bool) ([]A []A)

Partition splits a slice or map based on the evaluation of the supplied function

The base Partition function accepts interface{} types and returns []interface{}

  // Partition func(interface{}, func(interface{}) bool) ([]interface{}, []interface{})

  s := []int{1, 2, 3, 4, 5, 6, 7, 8, 9, 10}

  fn := func(i interface{}) bool {
    return (i.(int) % 2) == 1
  }

  odd, even := un.Partition(s, fn)

  fmt.Println(odd)  //[1, 3, 5, 7, 9]
  fmt.Println(even) //[2, 4, 6, 8, 10]

Typed Partitions can be defined using a function type and the MakePartition helper.

  // Partition func([]A, func(A) bool) ([]A []A)

  var IPartition func([]int, func(int) bool) ([]int, []int)

  un.MakePartition(&IPartition)

  s := []int{1, 2, 3, 4, 5, 6, 7, 8, 9, 10}

  fn := func(i int) bool {
    return (i % 2) == 1
  }

  odd, even := un.IPartition(s, fn)

  fmt.Println(odd)  //[1, 3, 5, 7, 9]
  fmt.Println(even) //[2, 4, 6, 8, 10]

Contains returns true if an object is in a slice.

  o := "a"
  s := []string{"a", "b", "c"}

  b := un.Contains(s, o)
  fmt.Println(b) //true

ToI converts a slice of arbitrary type []T into a slice of []interfaces{}

  s := []int{1, 1, 3, 5, 8, 13}
  i := un.ToI(s)

Notes

I am aware that the whole idea is not particularly very TheGoWay™, but it is useful as a learning exercise, and it is useful for moving fast and optimising later.