I'm developing an application in Marionette, but the topic refers to raw Backbone as well.
In my app, I've got many collections. Among them, there are 3 important ones, that are used all over the application (e.g. user data, such as names, which is displayed in most of views). The main question is: what is the best way (a good pattern to follow) to keep references to collections in Backbone/Marionette? I came up with few ideas below:
should I attach them as attributes of the Application object? If so, I'd have to pass the reference to Application object to all views, then views to their subviews, subviews to subsubviews, etc. But this seems a lame and ugly design.
instead, I could pass each collection separately, but this seems even worse solution, because I can't predict which (sub)view will ever need a collection. Keeping those nested references in order would be much more difficult than passing the Application object which I can pass once, always.
the is also a possibility to import the Application as a singleton. I'm using requireJS, most of the modules return constructors now (constructors of views, models and collections). Now the app module returns Application constructor, but instead, it could return the Application object. Then if a view requires to display some data from the collections, it could just require the app module and that's all.
finally, basing on the previous point, I thought I could create a plain map of collections and make it a singleton just as before. This is only to disable all views to have access to Application object, which still seems a bad pattern.
Please, suggest what you think is the best (commentson points above are welcome as well). I just need a good design pattern here. As far as I know, Marionette docs doesn't suggest anything here.
I follow the suggestions made in David Sulc's book Backbone.Marionette.js: A Gentle Introduction. (also the next book on goes into how to then sturcutre the same project with require https://leanpub.com/structuring-backbone-with-requirejs-and-marionette)The code examples he uses are available on github so you could look at the final example which to get an idea of what he does if you didn;t want to buy the book but i really recommend it as it really helped me with how to structure my projects.
To start I have setup an Entities module. The structure of the files also follows this I have an Entities folder which has separate entities.
Each Entity file concerns all actions and stucture of that particular entity. I like this approach as I when i want to edit an entities strcuture or method of getting data from the server i only need to go to one place to make this change.
Interactions with entity are handled through marionettes req/res system. In this way you can expose a handler to the rest of your app but they don;t need to be concerned with how that req gets handled as long as it brings back the required response.
Here is an example of one of my entities to show what i mean - My app needs a collection called positions at various stages so this is something that is loaded early in the app then is available through-out it's life-cycle.
//i am using require so i define my app to use
define(["app", ], function(MyApp) {
//All of this will be added to the Entities module so if i want direct access
//i can go Myapp.Entities.whateverEntityIhaveDeclared
MyApp.module("Entities", function(Entities, MyApp, Backbone, Marionette, $, _) {
//model (nothing new here)
Entities.Position = Backbone.Model.extend({
urlRoot: "api/positions",
defaults: {
name: "",
}
});
//collection again nothing new here
Entities.PositionCollection = Backbone.Collection.extend({
url: "api/positions",
model: Entities.Position,
comparator: "name"
});
//an init function to attach a position collection onto Entities module so it can be available throughout my app
var initializePositions = function() {
if (Entities.positions === undefined) {
Entities.positions = new Entities.PositionCollection();
}
};
//
var API = {
//returns a jquery deferred promise so that this request can easily be handled async
loadPositionEntitiesRemote: function() {
//init positions make's sure i have the collectoin avaliable if it
//has not yet been defined
initializePositions();
//setup defer object
var defer = $.Deferred();
//I actually use a custom sever object here when dealing
//with ajax requests
//but because my app always go through this API i can
//easily swap out how the collection is retrieved.
// Here is an example using backbones fetch
Entities.positions.fetch({
success: function() {
defer.resolve();
},
error: function(data) {
defer.reject(data);
}
});
//setup promise to return
var promise = defer.promise();
return promise;
},
//get the positions collection from here i could
//directly access the attributes or add to the collection
refrencePositionEntities: function() {
initializePositions();
return Entities.positions;
},
//get a position from the collection based on id
//
getPositionEntity: function(positionId) {
initializePositions();
return Entities.positions.get(positionId);
}
};
//setup handlers for the app to use
MyApp.reqres.setHandler("position:entities:remote", function() {
return API.loadPositionEntitiesRemote();
});
MyApp.reqres.setHandler("position:entities:refrence", function() {
return API.refrencePositionEntities();
});
MyApp.reqres.setHandler("position:entity", function(id) {
return API.getPositionEntity(id);
});
MyApp.reqres.setHandler("position:entity:new", function(position) {
return new Entities.Position(position);
});
});
return;
});
now to use this in my app here is an example of how it can now be used
someFunction: function(){
//so early in the app i will get the positions
var positionPromise = MyApp.request("position:entities:remote");
$.when(positionPromise).done(function() {
//do what ever as data has been loaded
}).fail(function(data){
//something failed so handle here might through up an error page but up to you
}).always(function(){
//something to always do no matter if fail or sucess
});
}
anotherFunction: function(){
//later in the app in another controller i might to get the collection
// I could also get it through MyApp.Entities.positions but i rather use the
// API set up so if i ever decided i want to add so checks or something in
// when retrieving the collection its super easy
var positionsCollection = MyApp.request("position:entities:refrence");
}
Not sure if i've done a great job explaining this but if you are looking for a ideas on good design in marionette check out the book as it explains this a lot better than i have just done
Related
I have two pages. One of them is dashboard with a lot of functionality. The second page is shared dashboard - the simple version of the first page.
The dashboard contains the view of the database (it can contain much other info, but the problem with this one). You can click on the filter button and modal window will be opened. So, simple version of the dashboard doesn't have this possibility. I'd like to add it, but I don't want to copy+past code from the full version of the dashboard because the code of this part is about two thousand lines. I'll add some primitive code example:
DashboardView = SomeAnotherView.extend({
initialize: function() {...},
events: {...} // huge objects of jQuery events,
render: function () {...},
... // 2k lines of functions for events
});
How can I use this View on another page? I tried to call a function from this view:
DashboardView.prototype.filterClicked(event);
But in this case event.curentTarget is null (it is necessary for this function), I also tried to send "this" to get the context, but it was failed.
Is there a possibility in Backbone.js to use one View for 2+ pages without any huge copy/past code?
Ideally if you have a simple version and full version of a view, you should have a "base view" (simple one) and the full version should extend the base view.
It'll look something like:
var SimpleDashbard = Backbone.view.extend({});
var Dashboard = SimpleDashbard.extend({});
In this way Dashboard will have access to the methods from SimpleDashbard.
Your situation sounds like you need to use a method from extended view in base view. Which is not a good idea. Ideally if it's shared you should move it to the base view/extract it into a utility method or service, and of course this involve re-writing this method to be reusable
If you have views that share a large amount of functionality, you could consider using the same View type, but opening it up to some configuration when instancing. For example:
var DashboardView = Backbone.View.extend({
initialize: function(options) {
this.allowFunctionX = (options && options.allowFunctionX);
this.allowFunctionY = (options && options.allowFunctionY);
},
// etc
functionX: function() {
if (!this.allowFunctionX) { return; }
// do the function...
},
functionY: function() {
if (!this.allowFunctionY) { return; }
// do the function...
},
});
Then on one page:
var firstDashView = new DashboardView({allowFunctionX: true});
and on another page:
var secondDashView = new DashboardView({allowFunctionY: true});
This may become not worth it if the functionality diverges too much (and there are likely better ways to configure than passing in a long list of booleans!). If your requirements are significantly different on your two pages, I feel like duplicating the code they both need is not a major sin.
TL; DR;
I'm looking for trite example of DDD node.js application.
Hi,
I'm going to create node application. I wonder that I can not find any example of application with business logic separated in domain.
OK, there are some examples like:
https://github.com/adrai/node-cqrs-domain - but this is whole CQRS with event sourcing implementation.
My idea is to do that like that:
//domain/book.js
function Book(title, author)
{
this._title = title;
this._author = author;
}
// domain methods ...
//infrastructure/persistance/repository/book-repository.js
function BookRepository()
{}
BookRepository.prototype.save(book)
{
var bookModel = mappers.mapToOrm(book);
return bookModel.save();
}
// [...] get, getAll, getNextId
//infrastructure/persistance/orm/book.js
//using http://bookshelfjs.org/
var Book = bookshelf.Model.extend({
tableName: 'books'
});
//infrastructure/mappers/book-mapper.js
function mapToOrm(book) {
//mapping [...]
return new persistance.Book();
}
function mapToDomain(domain) {
//mapping [...]
return new domain.Book();
}
but on the other hand I've never seen any similar solution (with domain model, orm model, repository and mappers). Am I thinking in the right way? Maybe there is no reason to separate business logic in domain in node.js applications. If so, why? If not, can you send me an example of DDD implementation or improve my code?
[2017/01/13]
I've created sample application in TypeScript. For now without repositories and not much services. Issues and pull requests are welcome.
https://github.com/dawiddominiak/ddd-typescript-bin-packing-problem-solution
I'm very new to Node.js world.
But I believe if you do your work using TypeScript with Node you can force most of DDD principles to be used.
The problem "and advantage in the same time" in node.js that there aren't so restrictions like we have in OOP languages like C# or Java. and this freedom "and messy" of JavaScript making create robust complex DomainModel and Business logic very hard
I'm looking to do the same thing at the moment, and I'm coming from the Ruby world. So, let me do 2 things:
Point you to the best Ruby implementation I've seen of Domain-Driven Design I have found, Hanami: http://hanamirb.org/guides/models/overview/ which you could use as a reference.
Discuss what I'm finding (literally right now, as I type) to attempt to find the analogs in Node.
I've found this page: https://github.com/sindresorhus/awesome-nodejs
which catalogs a ton of high-quality / high-popularity Node packages.
The first thing is, we need something that is going to do validation and schema construction for our Domain Models. Just looking at the first entry in the data validation section, Joi seems to be a decent choice for that:
https://github.com/hapijs/joi
For Persistence of the domain objects, I'm just setting up a stub object, borrowing from Hanami's interface:
var repo = {
find: function(entity_name, id) {
// - Fetch an entity from the collection by its ID
},
create: function(entity_name, data) {
// – Create a record for the given data and return an entity
},
update: function(entity_name, id, data) {
// – Update the record corresponding to the id and return the updated entity
},
delete: function(entity_name, id) {
// – Delete the record corresponding to the given entity
},
all: function(entity_name) {
// - Fetch all the entities from the collection
},
query: function(entity_name, query_object) {
},
first: function(entity_name) {
// - Fetch the first entity from the collection
},
last: function(entity_name) {
// - Fetch the last entity from the collection
},
clear: function(entity_name) {
// - Delete all the records from the collection
}
}
module.exports = repo
whether you choose to use Bookshelf, Sequelize, or even the LoopBack framework, you can code an object that is going to fit the above interface that then does the dirty work of integrating with those frameworks.
If I were to try different ORM's, I would create a different repo object for each of the above. Notice that as I've written it, the repo is a singleton that is aware of different entities and how to persist them. In many cases, this will no doubt delegate to different repository objects on a per-entity basis. However, that might not always be true. a simple in-memory repo, could just have an array of objects for each entity.
That leaves Services/Interactors - The functions/classes that actually do work. These are easy - they are the ones that take a domain object, perform some business logic, and in the CRUD cases, make a call out to the Repository. A probably-syntactically-wrong example:
const repository = require('./myFileRepository')
function createBook(bookEntity) {
if(bookEntity.valid?) {
repository.create('book', bookEntity)
return true
}
else {
return { error: 'book not valid' }
}
}
module.exports = createBook
for the Service functions, I just today learned about Node Machines, and they seem like a really smart idea: http://node-machine.org
they seem to be a JS-attempt at monads + documentation. so I'm considering writing them up like that.
anyway, given it's been a year since your post, you've probably moved on. hope this helps you / the community!
Many would argue that JavaScript is NOT well suited to modeling a complex problem into a domain model and then into code. That's especially the case if the domain is in business, industry and commerce, rather than in computer or data science.
I'm not saying one can't create a domain model in JavaScript. Just like one could create one in C. But does that mean one should?
The example you give uses some terminology in domain-driven design, but misses the whole purpose and ethos of applying it.
I have a collection called "Materials" defined on the client and server. In one template, I can insert and update correctly to Materials. On another template, I can insert and update but when I refresh the browser, the changes are lost.
The two views are called
view_materials (inserting/updating works)
view_orders (doesn't work)
Both templates have the materials collection binded to them like so:
//Bind viewOrders to Materials collection
Template.view_order.materials = function () {
return Materials.find();
};
// Bind materialsTemplate to Materials collection
Template.view_materials.materials = function () {
return Materials.find();
};
and both are using the function below to update.
var docid = Materials.findOne({material_number: newMaterial.material_number});
console.log(docid._id);
Materials.update({_id:docid._id},{$set: {material_qty: total}});
Please note that the ID printed out to the console matches perfectly. Like I mentioned, on view_orders it updates for a moment on the client but not in another browser window nor does it persist after being reloaded from the server. On view_materials it works perfectly. Any ideas?
I also deployed an example of the bug here:
http://upexmple.meteor.com/
and added the source to github:
https://github.com/stereochromatic/update_example
The relevant code can be found in client/views/view_materials/view_materials.js
and
client/views/view_orders/view_orders.js
To duplicate the error:
Click on inventory and under Raw Materials, type A for material number and -50 for quantity. You will see it get updated correctly. Now click on create release and under raw material type, select A and -50 for quantity. You will see the correct info get printed to the console and you may also see the changes on inventory but upon refresh those changes are gone.
- show quoted text -
You need to define your Meteor Collections outside of the client directory. I usually put mine into a /lib folder, but defining collections on just the client side can cause issues.
So just move
Materials = new Meteor.Collection("materials");
and
MaterialsLog = new Meteor.Collection("materialsLog");
into a folder outside of the client/server folders.
I can't reproduce the error. I've done as you said, but the changes do not disappear for me upon refresh. I can even close the tab and reopen it, and the changes are still there.
Finally figured this out...
I removed the autopublish package with
meteor remove autopublish
and then defined my permissions (and subscriptions) in /lib/Materials.js like so:
// Declare client materials collection
Materials = new Meteor.Collection("materials");
materials = Materials
if (Meteor.isClient) {
Meteor.subscribe('materials');
}
if (Meteor.isServer) {
Meteor.publish('materials', function() {
return materials.find();
});
materials.allow({
insert: function (document) {
return true;
},
update: function () {
return true;
},
remove: function () {
return false;
}
});
}
Thanks to everyone who helped out. I realize that for a production app this is best practice anyways so this needed really to be done at some point. I originally was going to wait until the end to remove autopublish, insecure and wrap collection in publish/subscribe, allow/deny rules but this issue helped expedite that work :)
Using your code above my with file structure:
Please take note that I remove the materials = Materials
client folder:
subscriber.js
Meteor.subscribe('materials');
server folder
publisher.js
Meteor.publish('materials', function() {
return Materials.find();
});
allow.js
Materials.allow({
insert: function (document) {
return true;
},
update: function () {
return true;
},
remove: function () {
return false;
}
});
}
collection folder. this is outside from the client and the server folders
collections.js
Materials = new Meteor.Collection("materials");
let me know if its works and though this very late at least people can see it
These days I find myself putting a lot of my code in the $(document).ready() which does not seem clean to me. For example, if I am creating something that will query my database via ajax and return it and append it to my list i would do something like this:
$(function(){
//Initialize my DOM elements
$MyList = $("#MyList"),
$MyButton = $("#MyButton");
//Add my Click event
$MyButton.click(function(){
$.ajax({
type: 'POST',
url: "/lists/getmylist",
contentType: "application/json",
success: function(results){
//Parse my results and append them using my favorite templating helper
var jsonResults = $.parseJSON(result);
$MyList.mustache("my-template", jsonResults);
}
});
})
});
Now I know this is a small example but it starts to get really big and messy when I have multiple click events, ajax requests etc. It all ends up going in my document ready. I know that I can possibly put all my ajax requests in an external javascript file to help make it cleaner, but is this architecture in general ok? just seems like its really messy. I have seen others use plugin architectures or init functions. I usually have this document ready at the bottom of all my pages and just throw in whatever is necessary to make my page work correctly. Is this a good way to structure my js?
I think the addition of some Model objects and general object oriented programming principals might go a long way here. If you break your your data fetching and storing out into model classes it should help a lot.
Here are some links that should get you started thinking about OO with Javascript.
Writing Object-Oriented JavaScript
Javascript Design Patterns
Javascript: prototypal inheritance
Javascript: prototypal inheritance 2
Another thing that might help out would be to break the Javascript into multiple files. One for global scripts that might be included via a header that attaches to all your pages and a script for each of your pages that requires it.
Perhaps Backbone.js ( or one of the other frameworks ) could be part of the rescue you are looking for.
I found Backbone immensely helpful organising some inherited spaghetti. Your starting point might be to transition your massive document ready into a backbone view (or multiples of)
Organise your scripts by separating out the views, collections, models into individual files then bundle and minify them together into a single file so the browser only needs to make one request instead of many.
ASP.NET MVC4 can do the bundling for you, it also works similarly on MVC3
This is just a example of simple starting point, there are more advanced techniques (eg. AMD, require.js) to reduce the script size per page, but with caching and gzip I find that the single everything script bundle is fine for a lot of cases.
As for your example, here's a possible backbone implementation
Remember to namespace out your code...
var app = app || {};
$(function ($) {
// depending on your server setup you might be able to just override the url
// and get away with what you want. Otherwise you might look into overriding
// the save/fetch or sync methods of the collection
app.MyListCollection = Backbone.Collection.extend({
url: '/lists/getmylist'
});
app.MyListView = Backbone.View.extend({
//bind the view to the existing element in the HTML.
el: '#MyList',
// your mustache template
template:$('#list-template').html(),
// hook up your event handlers to page components
//(within the context of your el)
events: {
'click #MyButton': 'onMyButtonClick'
},
//called on object creation.
initialize: function () {
//you could create a collection here or pass it into the init function
this.collection = new app.MyListCollection();
//when the collection is changes, call the render method
this.listenTo(this.collection, 'reset', this.render);
},
// render is named by convention, but its where you would "redraw" your view and apply your template
render: function () {
this.$el.html(
Mustache.render(
this.template(this.collection.toJSON())));
return this;
},
//your click handler
onMyButtonClick: function(e){
this.collection.fetch();
}
});
});
use your doc ready to spin up whatever backbone functionality you need
and use it bootstrap your javascript with any server side data that you may have.
$(function () {
// create an instance of your view
new app.MyListView();
//example bootstrap using razor
app.title = #Model.Title;
});
tl;dr: How does one implement MVC in JavaScript in a clean way?
I'm trying to implement MVC in JavaScript. I have googled and reorganized with my code countless times but have not found a suitable solution. (The code just doesn't "feel right".)
Here's how I'm going about it right now. It's incredibly complicated and is a pain to work with (but still better than the pile of code I had before). It has ugly workarounds that sort of defeat the purpose of MVC.
And behold, the mess, if you're really brave:
// Create a "main model"
var main = Model0();
function Model0() {
// Create an associated view and store its methods in "view"
var view = View0();
// Create a submodel and pass it a function
// that will "subviewify" the submodel's view
var model1 = Model1(function (subview) {
view.subviewify(subview);
});
// Return model methods that can be used by
// the controller (the onchange handlers)
return {
'updateModel1': function (newValue) {
model1.update(newValue);
}
};
}
function Model1(makeSubView) {
var info = '';
// Make an associated view and attach the view
// to the parent view using the passed function
var view = View1();
makeSubView(view.__view); // Dirty dirty
// Return model methods that can be used by
// the parent model (and so the controller)
return {
'update': function (newValue) {
info = newValue;
// Notify the view of the new information
view.events.value(info);
}
};
}
function View0() {
var thing = document.getElementById('theDiv');
var input = document.getElementById('theInput');
// This is the "controller", bear with me
input.onchange = function () {
// Ugly, uses a global to contact the model
main.updateModel1(this.value);
};
return {
'events': {},
// Adds a subview to this view.
'subviewify': function (subview) {
thing.appendChild(subview);
}
};
}
// This is a subview.
function View1() {
var element = document.createElement('div');
return {
'events': {
// When the value changes this is
// called so the view can be updated
'value': function (newValue) {
element.innerHTML = newValue;
}
},
// ..Expose the DOM representation of the subview
// so it can be attached to a parent view
'__view': element
};
}
How does one implement MVC in JavaScript in a cleaner way? How can I improve this system? Or is this the completely wrong way to go, should I follow another pattern?
There are at least a couple of established and usable MVC frameworks for JavaScript JavaScriptMVC and pureMVC. There are probably more. I've used JavaScriptMVC for browser based and Air apps and keep coming back to it - it has its problems but I've found it to be quite useful.
There are other solutions too, have a look at Sammy, a new thing I've heard good things about. I haven't used myself but intend to try soon. I don't know enough about it to describe it properly, but to me it seems like a front controller which works on routes, a templating system and ReSTful data stores. I'm not sure if it is MVC but has similar ingredients.
I have to disagree with mway's answer. MVC may be a bit diferent to implement in JavaScript but its benefits are very important to organising this mess. The design patterns usually associated with OO languages don't go out the window just because js isn't class based.
I would say that MVC is more suitable for JavaScript apps than for request based (server side) applications. Those objects can hang around for a while in a one page JavaScript app - minutes if not hours - and having a well organised way of organising their interaction will make your code much more robust and easy to deal with. There are books on the subject.
A couple of other points regarding the code you posted.
The view objects have responsibility for applying event listeners to DOM elements. This is the controller's job. The view just renders the HTML - the controller listens for the events and acts accordingly.
Your models seem to know your views. The model layer should have minimal knowledge of the view layer (perhaps being registered as observers). Keep your model clean and to the point, I mean the business point - business logic. In js apps you may just be proxying for a sever side model layer but it is important for your sanity to keep your model to the business logic and nothing else. Application logic is the controllers job
To be honest, MVC isn't well-suited for Javascript. It can support the basic fundamentals of the design, sure - you can create pseudoclasses to act as controllers or models, support basic inheritance, and you can have it manipulate or create any number of DOM elements, but there's a price that you pay for that - in overhead, accessibility, and usability.
In my opinion, I consider Javascript more of an augmentation - the KISS mentality exists for a good reason. If you're interested in better ways to organize your code, there's always the option of packaging related functionality into modules (sic) and abstracting out portions as appropriate. For example, creating a factory to do more complex AJAX request management, or a pseudoclass to handle processing of similar types of data. Using a standard base function for controllers, another for models, etc, as prototypes for new instances of those objects can accomplish similar functionality... but again, it's sort of going against the grain of Javascript.
However, if you're stuck on the MVC idea just for the sake of structure, consider something like the following:
;(function(window, $) {
/**
* Event Object
* A quick description goes here.
**/
var Events = window.Events = {
'bindTrackables': function() {
$('a.trackable').live('click', function() {
if(!_gaq)
_gaq = [];
_gaq.push(['_trackPageview', '/ajax/foobar']);
});
},
'bindSomeEvent': function() {
// etc
}
};
/**
* Data Cache
* I'll need to remember stuff later, so I store it here
**/
var Cache = window.Cache = {
'data': {},
'store': function(key, value) {
Cache.data[key] = value;
},
'fetch': function(key) {
return Cache.data[key];
}
};
/**
* Request Object
* Stores native AJAX requests for later use
**/
var Request = window.Request = {
'current_requests': [],
'send': function(url, type, data, callback) {
Request.current_requests.push($.ajax({
'url': url,
'type': type,
'data': data,
'callback': callback
}));
},
}
// add some private logic here
})(window, jQuery);
It's extremely basic, but you get the idea. Modular code is key... in JS, this is more important than forcing your application (or the language) to fit a certain style.