All Projects → VoliJS → Nestedtypes

VoliJS / Nestedtypes

Licence: mit
BackboneJS compatibility layer for Type-R data framework.

Projects that are alternatives of or similar to Nestedtypes

ng-observe
Angular reactivity streamlined...
Stars: ✭ 65 (-30.85%)
Mutual labels:  observable, reactive-programming
observe-component
A library for accessing React component events as reactive observables
Stars: ✭ 36 (-61.7%)
Mutual labels:  observable, reactive-programming
ObservableComputations
Cross-platform .NET library for computations whose arguments and results are objects that implement INotifyPropertyChanged and INotifyCollectionChanged (ObservableCollection) interfaces.
Stars: ✭ 94 (+0%)
Mutual labels:  observable, reactive-programming
Od Virtualscroll
🚀 Observable-based virtual scroll implementation in Angular
Stars: ✭ 133 (+41.49%)
Mutual labels:  reactive-programming, observable
Recycle
Convert functional/reactive object description using RxJS into React component
Stars: ✭ 374 (+297.87%)
Mutual labels:  reactive-programming, observable
Pharmacist
Builds observables from events.
Stars: ✭ 221 (+135.11%)
Mutual labels:  observable, reactive-programming
rxjs-proxify
Turns a Stream of Objects into an Object of Streams
Stars: ✭ 34 (-63.83%)
Mutual labels:  observable, reactive-programming
Redux Most
Most.js based middleware for Redux. Handle async actions with monadic streams & reactive programming.
Stars: ✭ 137 (+45.74%)
Mutual labels:  reactive-programming, observable
Swiftrex
Swift + Redux + (Combine|RxSwift|ReactiveSwift) -> SwiftRex
Stars: ✭ 267 (+184.04%)
Mutual labels:  reactive-programming, observable
observable-playground
Know your Observables before deploying to production
Stars: ✭ 96 (+2.13%)
Mutual labels:  observable, reactive-programming
Lightweightobservable
📬 A lightweight implementation of an observable sequence that you can subscribe to.
Stars: ✭ 114 (+21.28%)
Mutual labels:  reactive-programming, observable
Evolui
A tiny reactive user interface library, built on top of RxJs.
Stars: ✭ 43 (-54.26%)
Mutual labels:  reactive-programming, observable
realar
5 kB Advanced state manager for React
Stars: ✭ 41 (-56.38%)
Mutual labels:  observable, reactive-programming
react-mobx-router5
React components for routing solution using router5 and mobx
Stars: ✭ 58 (-38.3%)
Mutual labels:  observable, reactive-programming
Observer Util
Transparent reactivity with 100% language coverage. Made with ❤️ and ES6 Proxies.
Stars: ✭ 905 (+862.77%)
Mutual labels:  reactive-programming, observable
Rocket.jl
Functional reactive programming extensions library for Julia
Stars: ✭ 69 (-26.6%)
Mutual labels:  reactive-programming, observable
Coq Serapi
Coq Protocol Playground with Se(xp)rialization of Internal Structures.
Stars: ✭ 87 (-7.45%)
Mutual labels:  serialization
Protobuf Nim
Protobuf implementation in pure Nim that leverages the power of the macro system to not depend on any external tools
Stars: ✭ 90 (-4.26%)
Mutual labels:  serialization
Spring 5 Examples
This repository is contains spring-boot 2 / spring framework 5 project examples. Using reactive programming model / paradigm and Kotlin
Stars: ✭ 87 (-7.45%)
Mutual labels:  reactive-programming
Binary
Generic and fast binary serializer for Go
Stars: ✭ 86 (-8.51%)
Mutual labels:  serialization

NestedTypes 2.0 is BackboneJS compatibility layer for the Type-R data framework. Type-R is Model/Collection core written from the scratch with the TypeScript and has no side dependencies.

NestedTypes adds support for REST (standard BackboneJS API), Underscore methods, and Backbone 1.1 classes.

If you're upgrading from the version 1.3, there are compatibility issues. Mostly due to the fact that the Type-R and NestedTypes 2.0 is built around the concept of aggregation trees. NestedTypes 1.3 code won't work without refactoring.

Important Notice

Staring with v2.0 Type-R includes generic I/O abstraction which is far superior to the legacy BackboneJS I/O. NestedTypes & NestedReact will be maintained as the BackboneJS compatibility layer as long as Verizon/Volicon systems have legacy Backbone code. Therefore:

  • NestedTypes docs won't be updated. Use Type-R documentation as you primary source of documentation.
  • Functional-wise, there's no reason to prefer NestedTypes over the Type-R any more. If you don't need BackboneJS backward compatibility, move to the Type-R which doesn't have any legacy dependencies like jQuery and underscore.

Features

Post-backbone data framework. 10 times faster, first-class support for nested models and collections and relations by id.

  • ES6 classes support.
  • Deeply observable changes.
  • First-class support for aggregation and relations by id.
  • Attribute type annotations and dynamic type safety.
  • More than 10 times faster than BackboneJS and 2-4 times faster than NestedTypes 1.3 in all browsers.

Installation & Requirements

All modern JS engines are supported (IE10+, Safari, Firefox, Edge, Chrome, nodejs). May work in IE9 but not tested.

npm install nestedtypes

underscore and jquery are hard dependencies.

For lighter framework version without dependencies and Backbone compatibility shim check out Type-R.

Quick API Reference

Central concept in NestedTypes is Record type, which is the JS class with following capabilities:

  • Class members are deeply observable.
  • It is serializable to JSON by default.
  • Class members are typed and their changes are guardered with run-time type checks.

Model is the Record subclass representing REST API endpoint. Models, records, and their collections are used as building blocks to describe both application's UI state and its data layer.

Record definition looks like normal ES6 class definition, but it's mandatory to declare attributes. It looks like this:

import { define, Record } from 'nestedtypes'

@define // <- decorator to perform class transformation
class User extends Model {
    urlRoot : '/api/users',

    static attributes = { // <- attributes declaration
        name : '', // <- can be either default value
        email : String, // <- or any JS type constructor
        isActive : true,
        lastLogin : Date.value( null ) // <- or both
    }
}

const user = new User({ id : 5 }); // <- constructor takes optional attributes hash as an argument
user.fetch().done( () => { // GET /api/users/5
    user.name = 'John';
    user.save(); // PUT /api/users/5
});

Record's Attributes Type Annotations Basics

All record's attributes must be declared with static attributes = { [ attrName ] : TypeAnnotation } member. Type annotation can be one of the following:

  • attrName : Constructor. Such as attrName : Date.
  • attrName : Constructor.value( defaultValue ). Such as lastLogin : Date.value( null ).
  • attrName : defaultValue. In this case, attribute type will be inferred from the value, so isActive : true has the same effect as isActive : Boolean.value( true ).

Record attributes can be accessed directly, like user.name = x. When attribute is assigned, the type of the the value is checked and being converted to the declared type with its constructor invocation if it's necessary.

For the assignments like user.isActive = x, where isActive is declared as Boolean

  • it is assigned as is if x is null or boolean.
  • for primitive types, it's converted with plain constructor invokation like Boolean( x ).
  • For non-primitives convertion will invoke constructor with new, like new Date( x ).

If it's impossible to convert the value it may be assigned with NaN or Invalid Date (or depending on the type update will be rejected), and there will be an error in the console.

Therefore, it's guaranteed that Record attributes always have declared type.

Collections

Every model has corresponding Collection type declared implicitly. Collections implements Backbone Collection API.

var users = new User.Collection();

users.fetch().done( () => {
    console.log( users.length );
});

When Record is extended, its collection is extended too. The creation of implicit Colleciton type is equivalent to this:

@define
class Users extends Record.Collection {} 

@define
class User extends Record {
    static Collection = Users;
    // ...
}

You can use this pattern when you need to add custom members to the record's collection.

Nested Records and Collections

Nested records and collections are declared with mentioning constructor in attribute type annotation. All changes in nested objects are deeply observable; any change in children will cause change events in a parent.

Records and collections emiting the standard set of Backbone events, with following differences:

  • Collections does not bubble change:[attribute] event from the model by default (change event is bubbled); event bubbling needs to be enabled for every particular event with static itemEvents = { 'change:attr1' : true, ... } declaration.
  • Collections have changes event which is semantically similar to the model's change.

Transactions

Record's change event (and collection's changes event) are transactional. Whatever some changes are made as the reaction on any of change event, it won't cause additional change event for the owner.

Also, you can explicitly group the sequence of changes to the single transaction:

    some.record.transaction( record => {
        record.a = 1;
        record.b = 2;
        ...
    }); // some.record will emit single 'change' event if there was any changes.

    // Execute collection.each in the scope of transaction.
    todoCollection.updateEach( item => item.done = true ); // One 'changes' event will be emitted. 

Aggregation

Record can aggregate other records and collections in its attributes.

@define
class Team extends Record {
    static attributes = {
        members : User.Collection
        leader : User
    }
}

const team = new Team();
team.members.add( new User({ name : 'John' }) );

Aggregated members are:

  • serialized as nested JSON.
  • following operations recursively when the operation happens to its owner.

Aggregated members forms the tree of exclusive ownership. The same record or collection instance cannot be aggregated in two places at the same time, and this rule is checked and enforced.

Shared nested objects

Records may have nested members which belongs to different ownership trees. Special type annotations are required to mark attribute as shared.

  • RecordType.shared or CollectionType.shared. Reference to collection or record which may be aggregated somewhere else. null by default.
  • CollectionType.Refs constructor. Collection of records which may be aggregated somewhere else. Defaults to empty collection.

CollectionType.Refs is an equivalent to CollectionType.shared.value( [] ) when used as attribute type annotation.

Shared types are:

  • not a part of record's ownership tree.
  • not serialized.
  • Not a subject of recursive operations. They are empty by default, not cloned, not validated, and not disposed when the corresponding operation applied to the parent.

In all other aspects, they are indistinguishable from aggregated records and collections.

@define
class Team extends Record {
    static attributes = {
        members : User.Collection.Refs
        leader : User.shared
    }
}

Relationship by id

It's possible to create serializable reference to the shared object which is represented in JSON as a record's id (or array of ids). Special type annotation is required to point out the master collection which will be used to resolve ids to the records.

  • RecordType.from( masterCollection ) represents an id reference to the model.
  • CollectionType.subsetOf( masterCollection ) represents the collection of models id references.

id-reference types behaves as shared types, but:

  • they are serializable as an object id (or array of ids for collections).
  • they are not observable (internal changes do not trigger change events on the record).

masterCollection reference may be either:

  • direct reference to the globally available collection.
  • function returning the reference to the collection.
  • string, which is the symbolic reference to collection (dot-separated path to the collection taken relative to the record's this).
class Team extends Record {
    static attributes = {
        members : User.Collection,
        leader : User.from( 'members' ) // <- leader is serializable reference to the record from members collection.  
    }
}

Owner-references

^ symbol in symbolic reference represents getOwner() call and returns the record owner. Collections are skipped.

Following example expects that Team record will be aggregated (alone or in a collection) together with users collection.

class Team extends Record {
    static attributes = {
        members : User.Collection.subsetOf( '^users' ),
        leader : User.from( 'members' )  
    }
}

Tilda-References and Stores

Symbolic reference staring with ~ is resolved relative to the record called store, which is located with record.getStore() method. For instance, reference ~users will be resolved as this.getStore().users.

getStore() uses following store location algorithm:

  1. It traverse an ownership tree upwards and return the first Store model it has found.
  2. If none of the record's owners is the Store, it returns global store from Nested.store.

Store is the subclass of the Record and behaves as a regular Record. Therefore, resolution of id references depends on the context and you may have as many stores as you like.

Following example expects that there's users collection in some upper record which is inherited from Store, or (if there are none) in the global store:

class Team extends Record {
    static attributes = {
        members : User.Collection.subsetOf( '~users' ),
        leader : User.from( 'members' )  
    }
}

Attribute has-annotations

It's possible to control different aspects of record's attribute behavior through additional metadata. All of them starts with a keyword .has added to the constructor type.

Object describing an attribute is called metatype. Operations on metatypes are immutable (returns new metatype), and can be chained.

// Declare Month metatype.
const Month = Number.value( 1 ).has.check( x => x > 0 && x <= 12 );

attribute : Type.has.toJSON( false | ( x, name ) => json )

Override default serializer for the attribute. false option will exclude attribute from serialization.

attribute : Type.has.parse( ( json, name ) => data )

Override default JSON parser for the attribute.

attribute : Type.has.get( ( value, name ) => value )

Get hook which may transform attribute value on read. Get hooks can be chained.

attribute : Type.has.set( ( value, name ) => value )

Set hook which may transform attribute value before it's assigned. Set hooks can be chained.

attribute : RecordOrCollectionType.has.changeEvents( false )

When nested attribute is changed, don't mark the owner as changed.

attribute : Type.has.events({ [ event ] : handler | handlerName })

Listen to the specified events from the attribute. handler can be either function or the name of the record's method.

attribute : Type.has.check( x => boolean, errorMsg? : any )

Attach check to the attribute. Checks can be chained. Attribute is valid whenever check function returns truthy value.

errorMessage is optional.

attribute : Type.isRequired

Similar to Type.has.check( x => x, 'Required' ).

Validation

Validation is performed recursively on ownership tree. Record and collection shares the same validation API.

record.validate()

Override it to add custom record-level validation. Method shoudl return truthy value in case of validation error.

For attribute level checks see Type.has.check annotation.

record.isValid() : boolean

Checks whenever record is valid.

record.validationError

Return validation error object or null if there are no errors.

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].