redux-operations
v0.5.2
Published
Prioritized waterfall reducers for predictable, dynamic, multi reducers with a visual API
Downloads
5
Maintainers
Readme
redux-operations
Solves challenging redux problems in a clean, understandable, debuggable fasion.
##Installation
npm i -S redux-operations
What kind of problems can it solve?
Everything you'll ever encounter in the wild:
- Dynamic state
- Action watching (eg count how many times a particular action was run)
- Based on itermediary results (eg 2 reducers sharing the same action & 1 needs the result of the other)
- Async actions without thunk/promise middleware. Just write your logic directly in your reducer
- Visual, dare I say graphiQL, API to make debugging a breeze
Where can I see it working?
- https://github.com/mattkrick/redux-operations-counter-example
- https://github.com/slorber/scalable-frontend-with-elm-or-redux/tree/master/redux-operations-DrorT
- Your app here!
How's it different from redux-saga?
risking oversimplification, redux-saga uses generators and puts business logic in the middleware. It also lets you cancel promises.
redux-operations keeps all logic in the reducer by adding info to action.meta.operations
behind the scenes.
This allows it to solve a few extra hard problems like dynamic state.
It also plays nicely with vanilla redux so you only need to use it for the tricky parts of your app.
But, if you like a visual API for debugging, you might as well use it for your whole app.
See scalable-frontend-with-elm-or-redux to see how the two solve the same problem and decide which fits your use case.
##Usage
###Create the store enhancer
Just like redux-devtools
or applyMiddleware
, redux-operations is a storeEnhancer
.
API
const storeEnhancer = reduxOperations();
Example
import {createStore} from 'redux'
import {reduxOperations} from 'redux-operations';
return createStore(rootReducer, initialState, reduxOperations());
If you use this in conjunction with redux-devtools, you'll see an enhanced state there:
state = {
api: <YOUR REDUX-OPERATIONS API HERE>
userState: <YOUR STATE HERE>
})
There's no need to adjust any of your code, your application only sees what's inside userState
.
###Write your reducer
When your store is created, redux-operations ignores your regular reducers and only uses reducers designed for it.
These are easily created by using a reducer factory that takes in an operationName
, initialState
and an object full of "operations".
An operation is an action that is specific to the reducer. In other words, one action type has 1 or many operations.
This already occurs in the wild, but the execution order is arbitrary and intermediary results are not passed through.
The operationName is the same name that you use in your combineReducers
. By making you repeat that name here, we allow for perfect compatibility with standard redux.
API
operationReducerFactory(operationName, initialState, reducerObject);
Example
import {operationReducerFactory} from 'redux-operations';
const initialState = 0;
export const counter = operationReducerFactory('counter', initialState, {
INCREMENT_COUNTER: {
resolve: (state, action)=> state + 1
},
INCREMENT_ASYNC: {
priority: 1, // if this action type is used in another reducer, this determines which runs first
resolve: (state, action)=> {
setTimeout(()=> {
const {dispatch, locationInState} = action.meta.operations;
// yes, that counter variable below is a circular reference to the reducer object
const inc = bindOperationToActionCreators(locationInState, counter, increment);
dispatch(inc());
}, 1000);
return state;
}
},
SET_COUNTER: {
resolve: (state, action) => action.payload.newValue, // set the state to the variable passed in
arguments: {
newValue: {type: Number, description: 'The new value for the counter'} // show this in the API
}
}
});
Notice that all the logic occurs in the resolve
method, even async actions.
In plain redux, this logic is split between the action creator and the resolve function, which subjectively makes the flow harder to follow.
###Integrate into the model-view layer (eg your redux container)
redux-operations works with all frontend frameworks, but we'll show an example of it working in react.
For the example, imagine you have 2 counters that share the same reducer.
You need to know where in the state tree to find each (called locationInState
)
and the reducerObject
so it can initialize the state at runtime (for dynamically generated states)
First, we need to get the possibly-dynamic state from the state tree.
API
walkState(locationInState, state, reducerObject);
Example
import {counterReducer} from './counterReducer';
import {walkState} from 'redux-operations';
const mapStateToProps = (state, props) => {
return {
// `locationInState` is static here, but is usually passed in via props.
counter: walkState(['counters', 'top'], state, counterReducer);
}
};
Next, we need to make sure that our action creators attach this info to the new actions.
This is done by attaching locationInState
and the operationName
to the action.meta.operations
property.
Since the operationName
is stored in the reducerObject
, we just pass that in.
To make it easy, redux-operations offers a HOF to do the work for you.
It takes in a single function or an object of functions, similar to redux's bindDispatchToActions
.
API
bindOperationtToActionCreators(locationInState, reducerObject, actionCreators);
Example
import {bindOperationToActionCreators} from 'redux-operations';
import {counterReducer} from './counterReducer';
import * from './actionCreators';
import {connect} from 'react-redux';
@connect(mapStateToProps)
export default class Counter extends Component {
render() {
const {incrementAsync} = bindOperationToActionCreators(['counters', 'top'], counterReducer, actionCreators);
return (
<div>
<button onClick={() => dispatch(increment())}>+</button>
</div>
)
}
}
More advanced cases
For more advanced use cases, see the Counter example or read the blog post.