The typical Redux Reducer is function that takes in the previous state and an action and uses a switch case to determine how to transition the provided State. We can take advantage of the Action Names being Strings and replace the typical switch case with a JavaScript Object that pairs our State ADT reducers with their actions.

We can also take care of the case in which a reducer does not have an implementation for a given action, by reaching for the Maybe data type and updating our main Redux reducer to handle the case for us, by providing the previous state untouched. And all of this can be captured in a simple helper function we can use in all of our reducer files.

This post will focus on how to do reducer pattern, not the functionalities details, all the functionalities codes are listed here:

const {prop, isSameType, State, when, assign, omit, curry, converge,map, composeK, liftA2, equals, constant,option, chain, mapProps, find, propEq, isNumber, compose, safe} = require('crocks');
const {get, modify, of} = State; const state = {
cards: [
{id: 'green-square', color: 'green', shape: 'square'},
{id: 'orange-square', color: 'orange', shape: 'square'},
{id: 'blue-triangle', color: 'blue', shape: 'triangle'}
],
hint: {
color: 'green',
shape: 'square'
},
isCorrect: null,
rank: ,
left: ,
moves:
}
const inc = x => x + ;
const dec = x => x - ;
const incOrDec = b => b ? dec : inc;
const clamp = (min, max) => x => Math.min(Math.max(min, x), max);
const clampAfter = curry((min, max, fn) => compose(clamp(min, max), fn))
const limitRank = clampAfter(, );
const over = (key, fn) => modify(mapProps({[key]: fn}))
const getState = key => get(prop(key));
const liftState = fn => compose(
of,
fn
)
const limitMoves = clampAfter(, );
const decLeft = () => over("left", limitMoves(dec));
const incMoves = () => over("moves", limitMoves(inc));
const assignBy = (pred, obj) =>
when(pred, assign(obj));
const applyMove =
composeK(decLeft, incMoves) const getCard = id => getState('cards')
.map(chain(find(propEq('id', id))))
.map(option({}))
const getHint = () => getState('hint')
.map(option({}))
const cardToHint = composeK(
liftState(omit(['id'])),
getCard
)
const validateAnswer = converge(
liftA2(equals),
cardToHint,
getHint
)
const setIsCorrect = b => over('isCorrect', constant(b));
const adjustRank = compose(limitRank, incOrDec);
const updateRank = b => over('rank', adjustRank(b));
const applyFeedback = converge(
liftA2(constant),
setIsCorrect,
updateRank
)
const markSelected = id =>
assignBy(propEq('id', id), { selected: true })
const selectCard = id =>
over('cards', map(markSelected(id)))
const answer = composeK(
applyMove,
selectCard
)
const feedback = composeK(
applyFeedback,
validateAnswer
)

For now, the reducer pattern was implemented like this:

// Action a :: {type: string, payload: a}
// createAction :: String -> a -> Action a
const createAction = type => payload => ({type, payload});
const SELECT_CARD = 'SELECT_CARD';
const SHOW_FEEDBACK = 'SHOW_FEEDBACK';
const selectCardAction = createAction(SELECT_CARD);
const showFeedbackAction = createAction(SHOW_FEEDBACK); // reducer :: (State, a) -> (State AppState ()) | Null
const reducer = (prevState, {type, payload}) => {
let result;
switch(type) {
case SELECT_CARD:
result = answer(payload);
break;
case SHOW_FEEDBACK:
result = feedback(payload);
break;
default:
result = null;
} return isSameType(State, result) ? result.execWith(prevState): prevState;
} console.log(
reducer(
state,
showFeedbackAction('green-square')
)
)

Instead of using 'Switch', we can use Object to do the reducer:

const actionReducer= {
SELECT_CARD: answer,
SHOW_FEEDBACK: feedback
} const turn = ({type, payload}) => (actionReducer[type] || Function.prototype)(payload); const reducer = (prevState, action) => {
const result = turn(action);
return isSameType(State, result) ? result.execWith(prevState) : prevState;
}

'Function.prototype' makes sure that it always return a function can accept payload as param, just it do nothing and return undefined.

And code:

(actionReducer[type] || Function.prototype)

it is prefect for Maybe type, so we can continue with refactoring with Maybe:

const createReducer = actionReducer => ({type, payload}) =>
prop(type, actionReducer) // safe check type exists on actionReducer
.map(applyTo(payload)) // we get back a function need to call with payload, using applyTo
const turn = createReducer({
SELECT_CARD: answer,
SHOW_FEEDBACK: feedback
})
const reducer = (prevState, action) => {
return turn(action)
.chain(safe(isSameType(State))) // check result is the same type as State
.map(execWith(prevState)) // run with execWith
.option(prevState); // unwrap Just and provide default value };

---

 const {prop, execWith, applyTo, isSameType, State, when, assign, omit, curry, converge,map, composeK, liftA2, equals, constant,option, chain, mapProps, find, propEq, isNumber, compose, safe} = require('crocks');
const {get, modify, of} = State; const state = {
cards: [
{id: 'green-square', color: 'green', shape: 'square'},
{id: 'orange-square', color: 'orange', shape: 'square'},
{id: 'blue-triangle', color: 'blue', shape: 'triangle'}
],
hint: {
color: 'green',
shape: 'square'
},
isCorrect: null,
rank: ,
left: ,
moves:
}
const inc = x => x + ;
const dec = x => x - ;
const incOrDec = b => b ? dec : inc;
const clamp = (min, max) => x => Math.min(Math.max(min, x), max);
const clampAfter = curry((min, max, fn) => compose(clamp(min, max), fn))
const limitRank = clampAfter(, );
const over = (key, fn) => modify(mapProps({[key]: fn}))
const getState = key => get(prop(key));
const liftState = fn => compose(
of,
fn
)
const limitMoves = clampAfter(, );
const decLeft = () => over("left", limitMoves(dec));
const incMoves = () => over("moves", limitMoves(inc));
const assignBy = (pred, obj) =>
when(pred, assign(obj));
const applyMove =
composeK(decLeft, incMoves) const getCard = id => getState('cards')
.map(chain(find(propEq('id', id))))
.map(option({}))
const getHint = () => getState('hint')
.map(option({}))
const cardToHint = composeK(
liftState(omit(['id'])),
getCard
)
const validateAnswer = converge(
liftA2(equals),
cardToHint,
getHint
)
const setIsCorrect = b => over('isCorrect', constant(b));
const adjustRank = compose(limitRank, incOrDec);
const updateRank = b => over('rank', adjustRank(b));
const applyFeedback = converge(
liftA2(constant),
setIsCorrect,
updateRank
)
const markSelected = id =>
assignBy(propEq('id', id), { selected: true })
const selectCard = id =>
over('cards', map(markSelected(id)))
const answer = composeK(
applyMove,
selectCard
)
const feedback = composeK(
applyFeedback,
validateAnswer
)
// Action a :: {type: string, payload: a}
// createAction :: String -> a -> Action a
const createAction = type => payload => ({type, payload});
const SELECT_CARD = 'SELECT_CARD';
const SHOW_FEEDBACK = 'SHOW_FEEDBACK';
const selectCardAction = createAction(SELECT_CARD);
const showFeedbackAction = createAction(SHOW_FEEDBACK); const createReducer = actionReducer => ({type, payload}) =>
prop(type, actionReducer) // safe check type exists on actionReducer
.map(applyTo(payload)) // we get back a function need to call with payload, using applyTo
const turn = createReducer({
SELECT_CARD: answer,
SHOW_FEEDBACK: feedback
}) const reducer = (prevState, action) => {
return turn(action)
.chain(safe(isSameType(State))) // check result is the same type as State
.map(execWith(prevState)) // run with execWith
.option(prevState); // unwrap Just and provide default value }; const sillyVerb = createAction('SILLY_VERB'); console.log(
reducer(
state,
selectCardAction('green-square')
)
)

[Functional Programming ADT] Create State ADT Based Reducers (applyTo, Maybe)的更多相关文章

  1. [Functional Programming] Combine Multiple State ADT Instances with the Same Input (converge(liftA2(constant)))

    When combining multiple State ADT instances that depend on the same input, using chain can become qu ...

  2. [Functional Programming] Compose Simple State ADT Transitions into One Complex Transaction

    State is a lazy datatype and as such we can combine many simple transitions into one very complex on ...

  3. [React + Functional Programming ADT] Connect State ADT Based Redux Actions to a React Application

    With our Redux implementation lousy with State ADT based reducers, it is time to hook it all up to a ...

  4. [Functional Programming] Define Discrete State Transitions using the State ADT

    We build our first state transactions as two discrete transactions, each working on a specific porti ...

  5. [Functional Programming] Reader with Async ADT

    ReaderT is a Monad Transformer that wraps a given Monad with a Reader. This allows the interface of ...

  6. [Functional Programming] Introduction to State, thinking in State

    Recently, I am learning Working with ADT. Got some extra thought about State Monad. Basiclly how to ...

  7. [Functional Programming Monad] Substitute State Using Functions With A State Monad (get, evalWith)

    We take a closer look at the get construction helper and see how we can use it to lift a function th ...

  8. [Functional Programming ADT] Create a Redux Store for Use with a State ADT Based Reducer

    With a well defined demarcation point between Redux and our State ADT based model, hooking up to a R ...

  9. [Functional Programming ADT] Combine Multiple State ADT Based Redux Reducers

    Redux provides a convenient helper for combining many reducers called combineReducer, but it focuses ...

随机推荐

  1. c# 发送web请求

    我们目前涉及到的现有的接收请求方式有三种, 第一种: 页面式的Form表单 第二种: 服务的webservice形式的xml 第三个: restful风格的post包体json 第一种比较老,博客园的 ...

  2. ubuntu要安装新软件,已有deb安装包

    如果ubuntu要安装新软件,已有deb安装包(例如:iptux.deb),但是无法登录到桌面环境.那该怎么安装?答案是:使用dpkg命令.dpkg命令常用格式如下:sudo dpkg -I iptu ...

  3. 微信JSSDK分享功能实现

    <script src="http://res.wx.qq.com/open/js/jweixin-1.2.0.js"></script> <scri ...

  4. AtCoder Regular Contest 103 Problem D Robot Arms (构造)

    题目链接  Problem D 给定$n$个坐标,然后让你构造一个长度为$m$的序列, 然后给每个坐标规定一个长度为$m$的序列,ULRD中的一个,意思是走的方向, 每次从原点出发按照这个序列方向,每 ...

  5. Java多线程-run方法与start方法的区别

    package com.interview; /** * java多线程的两种实现方式以及run.start方法的区别 * @author MEI.LIU * */ public class Thre ...

  6. javascrip异步问题

    for ( var i = 1; i <= 3; i++) { setTimeout( function (){ console.log(i); }, 0); };     一般人会以为输出结果 ...

  7. 【推导】【凸包】MIPT-2016 Pre-Finals Workshop, Taiwan NTU Contest, Sunday, March 27, 2016 Problem D. Drawing Hell

    平面上n个点,两个人交替决策,用线段连接两个点,但不能跨越其他点或者已经存在的线段.不能做的人算输,问你谁赢. 实际上,跟两个人的决策无关,n个点将平面三角剖分,只需要算出有几条边即可. 凸包上如果有 ...

  8. 【动态规划】POJ3616-Milking Time

    [题目大意] 奶牛Bessie在0~N时间段产奶.农夫约翰有M个时间段可以挤奶,时间段[a,b]Bessie能挤到的牛奶量v.奶牛产奶后需要休息R小时才能继续下一次产奶,求Bessie最大的挤奶量. ...

  9. ArrayList之扩容与迭代器

    1. 扩容 ArrayList扩容包括ensureCapacity(对外开放)和ensureCapacityInternal(内部隐式调用)两个接口: 1' 两者都调用ensureExplicitCa ...

  10. Codeforces Round #344 (Div. 2) E. Product Sum 维护凸壳

    E. Product Sum 题目连接: http://www.codeforces.com/contest/631/problem/E Description Blake is the boss o ...