Endo Monoid

newtype Endo a = Endo { appEndo :: a -> a }
instance Monoid (Endo a) where
mempty = Endo id
Endo f `mappend` Endo g = Endo (f . g)

Endo 是个 newtype,也就是对现有类型的封装。

Endo a 封装的是一个自反射的函数,即 a->a。通过 appEndo 字段可以取出这个函数。

Endo a 在结合时结合两个函数,因此它本质上是对函数合成运算符 (.) 的封装。

Endo a 是一个幺半群。这是因为自反射函数在函数合成时满足结合律。

Dual Monoid

newtype Dual a = Dual { getDual :: a }
instance Monoid a => Monoid (Dual a) where
mempty = Dual mempty
Dual x `mappend` Dual y = Dual (y `mappend` x)

Dual 是个 newtype,也就是对现有类型的封装。

Dual a 封装的是一个值,即 a。通过 getDual 字段可以取出这个值。

Dual a 是一个幺半群,前提是 a 是个幺半群。

Dual a 在结合时交换两个操作数的值,因此它本质上是对 flip 函数的封装。

证明 Dual a 满足结合律。
(Dual x <> Dual y) <> Dual z
= Dual (y <> x) <> Dual z
= Dual (z <> (y <> x))
Dual x <> (Dual y <> Dual z)
= Dual x <> Dual (z <> y)
= Dual ((z <> y) <> x)
由于 a 是 Monoid 类型, 满足结合律
所以 (z <> y) <> x = z <> (y <> x)
=> Dual (z <> (y <> x)) = Dual ((z <> y) <> x)
=> Dual x <> (Dual y <> Dual z) = (Dual x <> Dual y) <> Dual z

Foldable 的法则

class Foldable t where
fold :: Monoid m => t m -> m
fold = foldMap id foldMap :: Monoid m => (a -> m) -> t a -> m
foldMap f = foldr (mappend . f) mempty foldr :: (a -> b -> b) -> b -> t a -> b
foldr f z t = appEndo (foldMap (Endo #. f) t) z foldl :: (b -> a -> b) -> b -> t a -> b
foldl f z t = appEndo (getDual (foldMap (Dual . Endo . flip f) t)) z

Foldable 类型类的实例都必须符合以下法则。

1. foldr f z t = appEndo (foldMap (Endo . f) t ) z
2. foldl f z t = appEndo (getDual (foldMap (Dual . Endo . flip f) t)) z
3. fold = foldMap id

[] 是个 Foldable

instance Foldable [] where
foldl = List.foldl
foldr = List.foldr
证明 Foldable 类型类的实例 [] 符合 Foldable 的法则
1. foldr f z t = appEndo (foldMap (Endo . f) t ) z
foldr f z t
= List.foldr f z [x1, x2, ..., xn]
= x1 `f` (x2 `f` (...(xn `f` z)...))
appEndo (foldMap (Endo . f) t ) z
= appEndo (foldr (mappend . (Endo . f)) mempty [x1, x2, ..., xn]) z
= appEndo (foldr (mappend . (Endo . f)) (Endo id) [x1, x2, ..., xn]) z
= appEndo ( (mappend . (Endo . f)) x1 ( (mappend . (Endo . f)) x2 ( ... ( (mappend . (Endo . f)) xn (Endo id) ) ... ) ) ) z
= appEndo ( (Endo . (x1 `f`)) <> ( (Endo . (x2 `f`)) <> ( ... <> ( (Endo . (f . id)) xn ) ... ) ) ) z
= appEndo ( Endo . ((x1 `f`) . (x2 `f`) . ... . (xn `f`)) ) $ z
= (x1 `f`) . (x2 `f`) . ... . (xn `f`) $ z
= x1 `f` (x2 `f` (...(xn `f` z)...))
2. foldl f z t = appEndo (getDual (foldMap (Dual . Endo . flip f) t)) z
foldl f z t
= List.foldl f z [x1, x2, ..., xn]
= (...((z `f` x1) `f` x2)...) `f` xn
appEndo (getDual (foldMap (Dual . Endo . flip f) t)) z
= appEndo (getDual (foldr (mappend . (Dual . Endo . flip f)) mempty [x1, x2, ..., xn])) z
= appEndo (getDual (foldr (mappend . (Dual . Endo . flip f)) (Dual . Endo . id) [x1, x2, ..., xn])) z
= appEndo ( getDual ( (Dual . Endo . flip f) x1 <> ( (Dual . Endo . flip f) x2 <> ( ... <> ( (Dual . Endo . (flip f . id)) xn ) ... ) ) ) ) z
= appEndo ( getDual ( (Dual . Endo . (`f` x1)) <> ( (Dual . Endo . (`f` x2)) <> ( ... <> ( (Dual . Endo . (`f` xn)) ) ... ) ) ) ) z
= appEndo ( getDual ( (Dual . Endo . ((`f` xn) . ... . (`f` x2) . (`f` x1)) ) ) z
= (`f` xn) . ... . (`f` x2) . (`f` x1) $ z
= (...((z `f` x1) `f` x2)...) `f` xn

build 与 (++)

build   :: forall a. (forall b. (a -> b -> b) -> b -> b) -> [a]
build g = g (:) [] (++) xs ys = foldr (:) ys xs

foldMap 与 concatMap

foldMap :: Monoid m => (a -> m) -> t a -> m

foldMap f = foldr (mappend . f) mempty

当 m := [b] 时 foldMap := concatMap

concatMap :: Foldable t => (a -> [b]) -> t a -> [b]
concatMap f xs = build (\c n -> foldr (\x b -> foldr c b (f x)) n xs)
concatMap f xs
= build (\c n -> foldr (\x b -> foldr c b (f x)) n xs)
= (\c n -> foldr (\x b -> foldr c b (f x)) n xs) (:) []
= foldr (\x b -> foldr (:) b (f x)) [] xs foldMap f xs
= foldr (mappend . f) mempty xs
= foldr ((++) . f) [] xs
= foldr (\x b -> foldr (:) b (f x)) [] xs

fold 与 concat

fold :: Monoid m => t m -> m

fold = foldMap id

当 m := [a] 时 fold := concat

concat :: Foldable t => t [a] -> [a]
concat xs = build (\c n -> foldr (\x y -> foldr c y x) n xs)
concat xs
= build (\c n -> foldr (\x y -> foldr c y x) n xs)
= (\c n -> foldr (\x y -> foldr c y x) n xs) (:) xs
= foldr (\x y -> foldr (:) y x) [] xs fold xs
= foldMap id xs
= foldr (mappend . id) mempty xs
= foldr (++) [] xs
= foldr (\x y -> foldr (:) y x) [] xs

reverse foldl foldr

用 foldl foldr 实现 reverse

How can I write reverse by foldr efficiently in Haskell?

reverse' xs
= foldl (flip (:)) [] xs
= appEndo . getDual $ foldMap (Dual . Endo . flip (flip (:))) xs $ []
= appEndo . getDual $ foldr (mappend . Dual . Endo . (:)) mempty xs $ []
= appEndo . foldr (flip mappend . Endo . (:)) mempty $ xs $ []
= foldr (flip (.) . (:)) id xs []
= flip (foldr (flip (.) . (:)) id) [] xs
reverse' = flip (foldr (flip (.) . (:)) id) []

将 reverse' 的推导过程一般化,可以得到

foldl f
= flip (foldr (flip (.) . flip f) id)

下面证明

foldl f z xs
= foldr (flip (.) . flip f) id xs z
= foldr step id xs z
where step x g a = g (f a x) step x g a = g (f a x)
step x g a = g ((flip f) x a)
step x g a = g . (flip f) x $ a
step x g = g . (flip f) x
step x g = (.) g ((flip f) x)
step x g = (flip (.)) ((flip f) x) g
step x = (flip (.)) ((flip f) x)
step x = flip (.) . flip f $ x
step = flip (.) . flip f

Haskell语言学习笔记(27)Endo, Dual, Foldable的更多相关文章

  1. Haskell语言学习笔记(88)语言扩展(1)

    ExistentialQuantification {-# LANGUAGE ExistentialQuantification #-} 存在类型专用的语言扩展 Haskell语言学习笔记(73)Ex ...

  2. Haskell语言学习笔记(79)lambda演算

    lambda演算 根据维基百科,lambda演算(英语:lambda calculus,λ-calculus)是一套从数学逻辑中发展,以变量绑定和替换的规则,来研究函数如何抽象化定义.函数如何被应用以 ...

  3. Haskell语言学习笔记(69)Yesod

    Yesod Yesod 是一个使用 Haskell 语言的 Web 框架. 安装 Yesod 首先更新 Haskell Platform 到最新版 (Yesod 依赖的库非常多,版本不一致的话很容易安 ...

  4. Haskell语言学习笔记(20)IORef, STRef

    IORef 一个在IO monad中使用变量的类型. 函数 参数 功能 newIORef 值 新建带初值的引用 readIORef 引用 读取引用的值 writeIORef 引用和值 设置引用的值 m ...

  5. Haskell语言学习笔记(39)Category

    Category class Category cat where id :: cat a a (.) :: cat b c -> cat a b -> cat a c instance ...

  6. Haskell语言学习笔记(58)Bifoldable

    Bifoldable class Bifoldable p where bifold :: Monoid m => p m m -> m bifold = bifoldMap id id ...

  7. Haskell语言学习笔记(84)Concurrent

    Control.Concurrent Prelude> import Control.Concurrent Prelude Control.Concurrent> Control.Conc ...

  8. Haskell语言学习笔记(72)Free Monad

    安装 free 包 $ cabal install free Installed free-5.0.2 Free Monad data Free f a = Pure a | Free (f (Fre ...

  9. Haskell语言学习笔记(44)Lens(2)

    自定义 Lens 和 Isos -- Some of the examples in this chapter require a few GHC extensions: -- TemplateHas ...

随机推荐

  1. tyvj1172自然数拆分

    题目:http://www.joyoi.cn/problem/tyvj-1172 非常水的完全背包.物品就是1~n这n个数. 第6行有橙色的警告:this decimal constant is un ...

  2. opencv中的缩放函数

    图像处理里面缩放操作是比较常见的: 最近邻插值:类似简单映射的处理方式,目标图像w1,h1,原始图像w0,h0,则在目标图像上的点(x,y)的像素点实际对应原始图上(x*w0/w1,y*h0/h1)的 ...

  3. ios获取安装的app

    http://www.iphonedevsdk.com/forum/iphone-sdk-development/22289-possible-retrieve-these-information.h ...

  4. mysql视图 新手的问答

    ☺ζั͡ޓއއއ๓º♥双٩(•(365335093) 16:03:02 创建的视图能保存多长时间,保存在哪啊 上天&宠儿(961431958) 16:08:59 数据库 上天&宠儿(9 ...

  5. python处理excel(二):写

    代码参考自zhoujie.函数接口可参考该blog. 基本的write函数接口很简单: 新建一个excel文件 file = xlwt.Workbook() (注意这里的Workbook首字母是大写) ...

  6. bzoj3685 普通veb树

    Description 设计数据结构支持: 1 x  若x不存在,插入x 2 x  若x存在,删除x 3    输出当前最小值,若不存在输出-1 4    输出当前最大值,若不存在输出-1 5 x   ...

  7. Eclipse: the import java.util cannot be resolved

    the import java.util cannot be resolved 导入JRE System Library. 右键项目 Build Path Configure Build Path.. ...

  8. cpu监控之三:mpstat命令

    mpstat是MultiProcessor Statistics的缩写,是实时系统监控工具.报告CPU的一些统计信息,这些信息存放在/proc/stat文件中.在多CPUs系统里,其不但能查看所有CP ...

  9. COMMON INTERVIEW QUESTIONS

    1. What do you see yourself doing five years from now? 2. What motivates you to put forth your great ...

  10. JAVA URI URL 区别

    String urlString = "http://192.168.21.77:8080/swp/mainPage?aa=11&bb%3D22"; URI uri = U ...