/Splifed_Drywall

This repo is a restructure of the drywall project.

Primary LanguageJavaScriptMIT LicenseMIT

Splifed_Drywall

This repo is a restructure of the drywall project. From the issue in the jedireza/drywall project named Structural proposition (https://github.com/jedireza/drywall/issues/212).

SO first my file hierarchy and with * are things i have changed : -configurator *
-controllers *
-layouts
-models *
-public
-util *
-views *
-GruntFile.js
-app.js *
-bower.json
-package.json

Now the overall concept:
-views and controllers have the same folders .Views only have jade files and controllers only have javascript files that implement the logic.
-models have the schemas but do not create the models themselves.
-util now has the passport.js and the http500.js
-configurator has the routes.js and models.js which are responsible for creating the routes and models.Important! routes.js does not work as it used to,meaning
that you do not have to write a long list of app.get app.post etc.(More on this later).
-controllers files only have one function and are named by them so for instance account/settings/ folder has ~20 files that implement the connect with Facebook,
Github,Google,update etc .And they follow a certain pattern in order to be defined as routes .
This pattern will be explained later.

Now on more details
1)In views i have removed every .js file, they are implemented in controllers now.
2)Schemas now exist in the models folder with a small change.Every schema file used to have a app.db.model(model.modelName, model.schema); statement .I have
removed this statement from all of them and i have added two new ,1 for the description of the schema and one for the model's name.The models creation is now
implemented in the configurator/models.js which sole purpose is to iterate through the files of the schemas folder and create all the models.It also keeps them in
an array and it is able to write them to the console for debugging purposes specifying the model's name and description.
3)As i said models and controllers have the exact same folders.
Controllers now are splitted in many files to define different operations.For example, in the views/account/settings used to be an index.js which defined all
operation over the account settings eg connect with fb,google,update account,identify etc.In my implementation all this logic is moved to controllers/account/
settings and every function is now a file eg i have a file named connectFacebook.js,connectGithub.js,identify.js,update.js. Those files though follow a certain
pattern which let the configurator/routes.js define the appropriate routes and handlers.This pattern is the following:

module.exports = {
    method : "GET",
    url : ["/account/settings/facebook/"],
    desc : "get : account settings facebook connectFacebook",
    passport : {   //this is null if no passport is required
        strategy : 'facebook',
        callbackURL : "/account/settings/facebook/callback/"
    },
    handler : function(req, res,next){
        //your code here
        }

This is from the controllers/account/settings/connectFacebook.js
As you see i export a javascript object which defines the method,the urls.the description,if passport authentication is required and the handler.

Now for the routes.
configurator/routes.js has an implementation of directory traversal of the controllers directory.
It takes each controller one by one and requires it .By that it creates the appropriate routes .
If project is on development mode it also stores them in a 'hashmap' and can output them for debugging purposes on the console and on the browser also.
It gives two extra endpoints for this.
-/someurl/dev
--Outputs a table describing the endpoint someurl
-/all/dev/
--Outputs all the endpoints with tables

What i think we get from this implementation:
1)Project is more readable and you can understand what is going pretty fast just by looking the files/folders names
2)People who create upon this do not need to define routes by themselves they just need to add their handlers under the controllers folder with the discussed
patern.
3)They can extend the database just by adding a schema to the schemas folder.
4)They get debugging sugars