Skip to content

Instantly share code, notes, and snippets.

@therebelrobot
Created January 25, 2017 01:53
Show Gist options
  • Save therebelrobot/c18d4bceff23317689d95845a1fb1404 to your computer and use it in GitHub Desktop.
Save therebelrobot/c18d4bceff23317689d95845a1fb1404 to your computer and use it in GitHub Desktop.
brigadehub pre-beta docs backup

Adding new features

This section includes advice on how to build new features for the project & what kind of process it includes.

  • This is how we like people to add new features:
    • Check in the issue tracker and make sure it's not already there
    • Open a new issue asking presenting the feature
    • Be aware that not all feature requests will be accepted, though we will do our best to accomodate
  • Here are some specifics on the coding style we prefer:
    • We use feross/standard for JS style
    • We use CSScomb for CSS styling (refer to .csscomb.json in the root of the project)
  • This is how you send your pull request:
    • If you are already part of the sfbrigade team:
      • Create a new feature/<feature-name> branch in the repo
      • When ready, open a pull request into edge
    • If you are not yet part of the team:
      • Fork the repo
      • make the desired changes, ensuring you fetch from upstream regularly
      • When ready, open a pull request into edge on upstream
    • Merge conflicts are the responsibility of the PR opener to resolve.
  • You should include the following tests:
    • Unit - granular tests making sure the individual functions of your addition work properly. This will include a *.tap.js file
  • These are we hope you include in your pull request description:
    • Make sure to list the feature additions, removals, bug fixes, etc.
    • Please include any relevant screenshots if effecting the view

Don’t get discouraged! We estimate that the response time from the maintainers is around: 24-48 hours

Branch Naming Format

We use Git Flow for branch naming, which follows the following format:

<type>/<branch name>

branch name should be something describing the actions being taken in the branch, not your username. The following types are used in this project:

  • release/: For official releases
  • hotfix/: for making fixes off a previous release
  • bugfix/: for making fixes for a current bug
  • feature/: for adding features to the system
  • docs/: for making changes/updates to documentation

These branch naming standards are being enforced on precommit via careful

Commit Message Format

This project adheres to the AngularJS Commit Message Conventions

Each commit message consists of a mandatory header, an optional body and an optional footer. The header has a special format that includes a type, a scope and a subject (full explanation):

<type>(<scope>): <subject>
<BLANK LINE>
<body>
<BLANK LINE>
<footer>

No worries if this format is new to you, try using commitizen to streamline the commit messaging. This project checks message format on commit using validate-commit-msg, so if your commit fails, please address the issues in the commit message and try committing again.

Patch Release

fix(pencil): stop graphite breaking when too much pressure applied

Minor Feature Release

feat(pencil): add 'graphiteWidth' option

Major Breaking Release

perf(pencil): remove graphiteWidth option

BREAKING CHANGE: The graphiteWidth option has been removed. The default graphite width of 10mm is always used for performance reason.

Bug triage

This section explains how bug triaging is done for brigadehub. Help beginners by including examples to good bug reports and providing them questions they should look to answer.

Documentation

Code needs explanation, and sometimes those who know the code will have trouble explaining it to someone just getting into it.

Community

If you are a brigade that wants to help out with brigadehub, you can create a synced #brigadehub channel in Slack by going to the following link: https://slackline.io/shared_channels/mmigV0dcH_WpIfLnMA567g

Most of our immediate conversations happen here. For more pressing matters, we try to set up Google Hangouts, and each week at SF Hack night, we set up a video Hangout for standup. Please let us know if you would like to join that.

Otherwise, feel free to open a discussion via Github Issues, and we can get back to you soon! Here are some other things you can do to help out:

  • Create an example of the project in real world by building something or showing what others have built.
  • Write about other people’s projects based on this project. Show how it’s used in daily life. Take screenshots and make videos!

Your first bugfix

This section should help a person get started with their very first bug fix and thinking through the problem.

  • If you have further questions, contact:

This contributing file template is from contribute.md.

Before starting

  • Have the application installed and running locally on your machine
  • Have Zenhub installed in your chrome browser so you can see both the Epics and the Boards we are using
  • Have looked through and understand the Brigadehub Roadmap and the feature list for the current release push.
  • Have joined the SFbrigade Github org (join by clicking here)

When you are ready to start

This application has a lot of moving parts, and we're working hard to make sure those parts are fitting together nicely. If you have any questions, please don't hesitate to ask.

Prerequisites

  • Command Line Tools
  •  Mac OS X:
    • Xcode (or OS X 10.9 Mavericks: xcode-select --install)
    • HomeBrew /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
  •  Windows: Visual Studio
  •  Ubuntu /  Linux Mint: sudo apt-get install build-essential
  •  Fedora: sudo dnf groupinstall "Development Tools"
  •  OpenSUSE: sudo zypper install --type pattern devel_basis
  • MongoDB
    •  Mac OS X: brew install mongodb
  • Node.js v4.x.x (Easiest install is via NVM)
    • Uninstall any previously installed Node versions (if you don't already have nvm installed)
    • curl -o- https://raw.githubusercontent.com/creationix/nvm/v0.31.0/install.sh | bash && echo 'export NVM_DIR="$HOME/.nvm"' >> $HOME/.bashrc && echo '[ -s "$NVM_DIR/nvm.sh" ] && . "$NVM_DIR/nvm.sh" # This loads nvm' >> $HOME/.bashrc && . $HOME/.bashrc && nvm install 4 && nvm alias default 4 && nvm use 4

Note: If you are new to Node or Express, I recommend to watch Node.js and Express 101 screencast by Alex Ford that teaches Node and Express from scratch. Alternatively, here is another great tutorial for complete beginners - Getting Started With Node.js, Express, MongoDB.

Getting Started

The easiest way to get started is to clone the repo:

# Get the latest snapshot
git clone https://github.com/sfbrigade/brigadehub.git

# Change directory
cd brigadehub

# Install NPM dependencies
npm install

# If needed, start mongodb in a separate tab
mongod

npm start

# or if starting for local development:

npm run develop

Obtaining API Keys

To use any of the included APIs or OAuth authentication methods, you will need to obtain appropriate credentials: Client ID, Client Secret, API Key, or Username & Password. You will need to go through each provider to generate new credentials.

I have included dummy keys and passwords for Github to get you up and running even faster. But don't forget to update them with your credentials when you are ready to deploy an app.

- Go to [Account Settings](https://github.com/settings/profile) - Select **Applications** from the sidebar - Then inside **Developer applications** click on **Register new application** - Enter *Application Name* and *Homepage URL* - For *Authorization Callback URL*: http://localhost:3000/auth/github/callback - Click **Register application** - Now copy and paste *Client ID* and *Client Secret* keys into `.env` file

Project Structure

Name Description
config/default-brigade.js Default brigade profile.
config/passport.js Passport Local and OAuth strategies, plus login middleware.
controllers/about.js Controller for about form.
controllers/blog.js Controller for editing/creating/displaying blog posts.
controllers/brigade.js Controller for brigade login and page management.
controllers/events.js Controller for events calendar and individual event pages.
controllers/home.js Controller for home page (index).
controllers/projects.js Controller for /projects route and all projects examples.
controllers/user.js Controller for user account management.
models/Brigade.js Mongoose schema and model for Brigade.
models/Content.js Mongoose schema and model for site content.
models/Events.js Mongoose schema and model for Events.
models/Posts.js Mongoose schema and model for blog Posts.
models/User.js Mongoose schema and model for User.
node_modules/ Storage for all node modules.
test/ Tests for routes and models.
themes/atl/public/ Static assets (fonts, css, js, img).
themes/atl/public/js/main.js Place your client-side JavaScript here.
themes/atl/public/css/main.css Main stylesheet for your app.
themes/atl/public/css/themes/default.scss Some Bootstrap overrides to make it look prettier.
themes/atl/views/about/ Templates for about page.
themes/atl/views/account/ Templates for login, password reset, signup, profile.
themes/atl/views/blog/ Templates for Blog pages.
themes/atl/views/events/ Templates for Event pages.
themes/atl/views/projects/ Templates for Project Examples.
themes/atl/views/partials/flash.jade Error, info and success flash notifications.
themes/atl/views/partials/header.jade Navbar partial template.
themes/atl/views/partials/footer.jade Footer partial template.
themes/atl/views/brigade.jade Brigade page template.
themes/atl/views/home.jade Home page template.
themes/atl/views/layout.jade Base template.
.travis.yml Travis CI integration.
.env.example Your API keys, tokens, passwords and database URI.
app.js Main application file.
setup.js Tool for removing authentication providers and other things.

Note: There is no preference how you name or structure your views. You could place all your templates in a top-level views directory without having a nested folder structure, if that makes things easier for you. Just don't forget to update extends ../layout and corresponding res.render() paths in controllers.

List of Packages

Package Description
async Utility library that provides asynchronous control flow.
bcrypt-nodejs Library for hashing and salting user passwords.
bitgo Multi-sig Bitcoin wallet API.
cheerio Scrape web pages using jQuery-style syntax.
clockwork Clockwork SMS API library.
connect-mongo MongoDB session store for Express.
dotenv Loads environment variables from .env file.
express Node.js web framework.
body-parser Express 4 middleware.
cookie-parser Express 4 middleware.
express-session Express 4 middleware.
morgan Express 4 middleware.
compression Express 4 middleware.
errorhandler Express 4 middleware.
method-override Express 4 middleware.
serve-favicon Express 4 middleware offering favicon serving and caching.
express-flash Provides flash messages for Express.
express-validator Easy form validation for Express.
fbgraph Facebook Graph API library.
github-api GitHub API library.
jade Template engine for Express.
lastfm Last.fm API library.
instagram-node Instagram API library.
lob Lob API library
lusca CSRF middleware.
mongoose MongoDB ODM.
node-foursquare Foursquare API library.
node-linkedin LinkedIn API library.
node-sass-middleware Sass middleware compiler.
nodemailer Node.js library for sending emails.
passport Simple and elegant authentication library for node.js
passport-facebook Sign-in with Facebook plugin.
passport-github Sign-in with GitHub plugin.
passport-google-oauth Sign-in with Google plugin.
passport-twitter Sign-in with Twitter plugin.
passport-instagram Sign-in with Instagram plugin.
passport-local Sign-in with Username and Password plugin.
passport-linkedin-oauth2 Sign-in with LinkedIn plugin.
passport-oauth Allows you to set up your own OAuth 1.0a and OAuth 2.0 strategies.
paypal-rest-sdk PayPal APIs library.
request Simplified HTTP request library.
stripe Offical Stripe API library.
tumblr.js Tumblr API library.
twilio Twilio API library.
twit Twitter API library.
lodash Handy JavaScript utlities library.
validator Used in conjunction with express-validator in controllers/api.js.
mocha Test framework.
chai BDD/TDD assertion library.
supertest HTTP assertion library.
multiline Multi-line strings for the generator.
blessed Interactive command line interface for the generator.
yui Used by the Yahoo API example.

Useful Tools and Resources

Recommended Design Resources

Recommended Node.js Libraries

  • Nodemon - Automatically restart Node.js server on code changes.
  • geoip-lite - Geolocation coordinates from IP address.
  • Filesize.js - Pretty file sizes, e.g. filesize(265318); // "265.32 kB".
  • Numeral.js - Library for formatting and manipulating numbers.
  • Node Inspector - Node.js debugger based on Chrome Developer Tools.
  • node-taglib - Library for reading the meta-data of several popular audio formats.
  • sharp - Node.js module for resizing JPEG, PNG, WebP and TIFF images.

Recommended Client-side Libraries

  • Framework7 - Full Featured HTML Framework For Building iOS7 Apps.
  • InstantClick - Makes your pages load instantly by pre-loading them on mouse hover.
  • NProgress.js - Slim progress bars like on YouTube and Medium.
  • Hover - Awesome CSS3 animations on mouse hover.
  • Magnific Popup - Responsive jQuery Lightbox Plugin.
  • jQuery Raty - Star Rating Plugin.
  • Headroom.js - Hide your header until you need it.
  • X-editable - Edit form elements inline.
  • Offline.js - Detect when user's internet connection goes offline.
  • Alertify.js - Sweet looking alerts and browser dialogs.
  • selectize.js - Styleable select elements and input tags.
  • drop.js - Powerful Javascript and CSS library for creating dropdowns and other floating displays.
  • scrollReveal.js - Declarative on-scroll reveal animations.

Pro Tips

  • When installing an NPM package, add a --save flag, and it will be automatically added to package.json as well. For example, npm install --save moment.
  • Use async.parallel() when you need to run multiple asynchronous tasks, and then render a page, but only when all tasks are completed. For example, you might want to scrape 3 different websites for some data and render the results in a template after all 3 websites have been scraped.
  • Need to find a specific object inside an Array? Use _.find function from Lodash. For example, this is how you would retrieve a Twitter token from database: var token = _.find(req.user.tokens, { kind: 'twitter' });, where 1st parameter is an array, and a 2nd parameter is an object to search for.

FAQ

I am getting MongoDB Connection Error, how do I fix it?

That's a custom error message defined in app.js to indicate that there was a problem connecting to MongoDB:

mongoose.connection.on('error', function() {
  console.error('MongoDB Connection Error. Please make sure MongoDB is running.');
});

You need to have a MongoDB server running before launching app.js. You can download MongoDB here, or install it via a package manager. Windows users, read Install MongoDB on Windows.

Tip: If you are always connected to the internet, you could just use MongoLab or Compose instead of downloading and installing MongoDB locally. You will only need to update database credentials in .env file.

I get an error when I deploy my app, why?

Chances are you haven't changed the Database URI in .env. If MONGODB/MONGOLAB_URI is set to localhost, it will only work on your machine as long as MongoDB is running. When you deploy to Heroku, OpenShift or some other provider, you will not have MongoDB running on localhost. You need to create an account with MongoLab or Compose, then create a free tier database. See Deployment for more information on how to setup an account and a new database step-by-step with MongoLab.

Why Jade instead of Handlebars?

When I first started this project I didn't have any experience with Handlebars. Since then I have worked on Ember.js apps and got myself familiar with the Handlebars syntax. While it is true Handlebars is easier, because it looks like good old HTML, I have no regrets picking Jade over Handlebars. First off, it's the default template engine in Express, so someone who has built Express apps in the past already knows it. Secondly, I find extends and block to be indispensable, which as far as I know, Handlebars does not have out of the box. And lastly, subjectively speaking, Jade looks much cleaner and shorter than Handlebars, or any non-HAML style for that matter.

Why do you have all routes defined in app.js?

For the sake of simplicity. While there might be a better approach, such as passing app context to each controller as outlined in this blog, I find such style to be confusing for beginners. It took me a long time to grasp the concept of exports and module.exports, let alone having a global app reference in other files. That to me is a backward thinking. The app.js is the "heart of the app", it should be the one referencing models, routes, controllers, etc. When working solo on small projects I actually prefer to have everything inside app.js as is the case with this REST API server.

I don't need a sticky footer, can I delete it?

Absolutely. But unlike a regular footer there is a bit more work involved. First, delete #wrap and #footer ID selectors and html, body { height: 100%; } from main.less. Next, delete #wrap and #footer lines from layout.jade (By the way, if no element is specified before class or id, Jade assumes it is a div element). Don't forget to indent everything under #wrap to the left once, since this project uses two spaces per block indentation.

How It Works (mini guides)

This section is intended for giving you a detailed explanation about how a particular functionality works. Maybe you are just curious about how it works, or maybe you are lost and confused while reading the code, I hope it provides some guidance to you.

###Custom HTML and CSS Design 101

HTML5 UP has many beautiful templates that you can download for free.

When you download the ZIP file, it will come with index.html, images, css and js folders. So, how do you integrate it with Hackathon Starter? Hackathon Starter uses Bootstrap CSS framework, but these templates do not. Trying to use both CSS files at the same time will likely result in undesired effects.

Note: Using the custom templates approach, you should understand that you cannot reuse any of the views I have created: layout, home page, api browser, login, signup, account management, contact. Those views were built using Bootstrap grid and styles. You will have to manually update the grid using a different syntax provided in the template. Having said that, you can mix and match if you want to do so: Use Bootstrap for main app interface, and a custom template for a landing page.

Let's start from the beginning. For this example I will use Escape Velocity template: Alt

Note: For the sake of simplicity I will only consider index.html, and skip left-sidebar.html, no-sidebar.html, right-sidebar.html.

Move all JavaScript files from html5up-escape-velocity/js to public/js. Then move all CSS files from html5up-escape-velocity/css to public/css. And finally, move all images from html5up-escape-velocity/images to public/images. You could move it to the existing img folder, but that would require manually changing every img reference. Grab the contents of index.html and paste it into HTML To Jade.

Note: Do not forget to update all the CSS and JS paths accordingly.

Create a new file escape-velocity.jade and paste the Jade markup in views folder. Whenever you see the code res.render('account/login') - that means it will search for views/account/login.jade file.

Let's see how it looks. Create a new controller escapeVelocity inside controllers/home.js:

exports.escapeVelocity = function(req, res) {
  res.render('escape-velocity', {
    title: 'Landing Page'
  });
};

And then create a route in app.js. I placed it right after the index controller:

app.get('/escape-velocity', homeController.escapeVelocity);

Restart the server (if you are not using nodemon), then you should see the new template at http://localhost:3000/escape-velocity.

I will stop right here, but if you would like to use this template as more than just a single page, take a look at how these Jade templates work: layout.jade - base template, index.jade - home page, partials/header.jade - Bootstrap navbar, partials/footer.jade - sticky footer. You will have to manually break it apart into smaller pieces. Figure out which part of the template you want to keep the same on all pages - that's your new layout.jade. Then, each page that changes, be it index.jade, about.jade, contact.jade will be embedded in your new layout.jade via block content. Use existing templates as a reference.

This is a rather lengthy process, and templates you get from elsewhere, might have yet another grid system. That's why I chose Bootstrap for the Hackathon Starter. Many people are already familiar with Bootstrap, plus it's easy to get started with it if you have never used Bootstrap. You can also buy many beautifully designed Bootstrap themes at Themeforest, and use them as a drop-in replacement for Hackathon Starter. However, if you would like to go with a completely custom HTML/CSS design, this should help you to get started!


How do flash messages work in this project?

Flash messages allow you to display a message at the end of the request and access it on next request and only next request. For instance, on a failed login attempt, you would display an alert with some error message, but as soon as you refresh that page or visit a different page and come back to the login page, that error message will be gone. It is only displayed once. This project uses express-flash module for flash messages. And that module is built on top of connect-flash, which is what I used in this project initially. With express-flash you don't have to explicity send a flash message to every view inside res.render(). All flash messages are available in your views via messages object by default, thanks to express-flash.

Flash messages have a two-step process. You use req.flash('errors', { msg: 'Error messages goes here' } to create a flash message in your controllers, and then display them in your views:

if messages.errors
  .alert.alert-danger.fade.in
    for error in messages.errors
      div= error.msg

In the first step, 'errors' is the name of a flash message, which should match the name of the property on messages object in your views. You place alert messages inside if message.errors because you don't want to show them flash messages are actually present. The reason why you pass an error like { msg: 'Error messages goes here' } instead of just a string - 'Error messages goes here', is for the sake of consistency. To clarify that, express-validator module which is used for validating and sanitizing user's input, returns all errors as an array of objects, where each object has a msg property with a message why an error has occurred. Here is a more general example of what express-validator returns when there are errors present:

[
  { param: "name", msg: "Name is required", value: "<received input>" },
  { param: "email", msg: "A valid email is required", value: "<received input>" }
]

To keep consistent with that style, you should pass all flash messages as { msg: 'My flash message' } instead of a string. Otherwise you will just see an alert box without an error message. That is because, in partials/flash.jade template it will try to output error.msg (i.e. "My flash message".msg), in other words it will try to call a msg method on a String object, which will return undefined. Everything I just mentioned about errors, also applies to "info" and "success" flash messages, and you could even create a new one yourself, such as:

Data Usage Controller (Example)

req.flash('warning', { msg: 'You have exceeded 90% of your data usage' });

User Account Page (Example)

if messages.warning
  .alert.alert-warning.fade.in
    for warning in messages.warning
      div= warning.msg

partials/flash.jade is a partial template that contains how flash messages are formatted. Previously, flash messages were scattered throughout each view that used flash messages (contact, login, signup, profile), but now, thankfully it is uses a DRY approach.

The flash messages partial template is included in the layout.jade, along with footer and navigation.

body
  #wrap
    include partials/navigation
    .container
      include partials/flash
      block content
  include partials/footer

If you have any further questions about flash messages, please feel free to open an issue and I will update this mini-guide accordingly, or send a pull request if you would like to include something that I missed.


How do I create a new page?

A more correct way to be to say "How do I create a new route". The main file app.js contains all the routes. Each route has a callback function associated with it. Sometimes you will see 3 or more arguments to routes. In cases like that, the first argument is still a URL string, while middle arguments are what's called middleware. Think of middleware as a door. If this door prevents you from continuing forward, you won't get to your callback function. One such example is a route that requires authentication.

app.get('/account', passportConf.isAuthenticated, userController.getAccount);

It always goes from left to right. A user visits /account page. Then isAuthenticated middleware checks if you are authenticated:

exports.isAuthenticated = function(req, res, next) {
  if (req.isAuthenticated()) {
    return next();
  }
  res.redirect('/login');
};

If you are authenticated, you let this visitor pass through your "door" by calling return next();. It then proceeds to the next middleware until it reaches the last argument, which is a callback function that typically renders a template on GET requests or redirects on POST requests. In this case, if you are authenticated, you will be redirected to Account Management page, otherwise you will be redirected to Login page.

exports.getAccount = function(req, res) {
  res.render('account/profile', {
    title: 'Account Management'
  });
};

Express.js has app.get, app.post, app.put, app.delete, but for the most part you will only use the first two HTTP verbs, unless you are building a RESTful API. If you just want to display a page, then use GET, if you are submitting a form, sending a file then use POST.

Here is a typical workflow for adding new routes to your application. Let's say we are building a page that lists all books from database.

Step 1. Start by defining a route.

app.get('/books', bookController.getBooks);

Note: As of Express 4.x you can define you routes like so:

app.route('/books')
  .get(bookController.getBooks)
  .post(bookController.createBooks)
  .put(bookController.updateBooks)
  .delete(bookController.deleteBooks)

And here is how a route would look if it required an authentication and an authorization middleware:

app.route('/api/twitter')
  .all(passportConf.isAuthenticated)
  .all(passportConf.isAuthorized)
  .get(apiController.getTwitter);
  .post(apiController.postTwitter)

Use whichever style that makes sense to you. Either one is acceptable. I really think that chaining HTTP verbs on app.route is very clean and elegant approach, but on the other hand I can no longer see all my routes at a glance when you have one route per line.

Step 2. Create a new schema and a model Book.js inside the models directory.

var bookSchema = new mongoose.Schema({
  name: String
});

var Book = mongoose.model('Book', bookSchema);
module.exports = Book;

Step 3. Create a new controller file called book.js inside the controllers directory.

/**
 * GET /books
 * List all books.
 */
var Book = require('../models/Book.js');

exports.getBooks = function(req, res) {
  Book.find(function(err, docs) {
    res.render('books', { books: docs });
  });
};

Step 4. Import that controller in app.js.

var bookController = require('./controllers/book');

Step 5. Create books.jade template.

extends layout

block content
  .page-header
    h3 All Books

  ul
    for book in books
      li= book.name

That's it! I will say that you could have combined Step 1, 2, 3 as following:

app.get('/books', function(req, res) {
  Book.find(function(err, docs) {
    res.render('books', { books: docs });
  });
});

Sure, it's simpler, but as soon as you pass 1000 lines of code in app.js it becomes a little difficult to navigate the file. I mean, the whole point of this boilerplate project was to separate concerns, so you could work with your teammates without running into MERGE CONFLICTS. Imagine you have 4 developers working on a single app.js, I promise you it won't be fun resolving merge conflicts all the time. If you are the only developer then it's fine. But as I said, once it gets up to a certain LoC size, it becomes difficult to maintain everything in a single file.

That's all there is to it. Express.js is super simple to use. Most of the time you will be dealing with other APIs to do the real work: Mongoose for querying database, socket.io for sending and receiving messages over websockets, sending emails via Nodemailer, form validation using express-validator library, parsing websites using Cheerio, and etc.


How do I use Socket.io with this?

Dan Stroot submitted an excellent pull request that adds a real-time dashboard with socket.io. And as much as I'd like to add it to the project, I think it violates one of the main principles of the Hackathon Starter:

When I started this project, my primary focus was on simplicity and ease of use. I also tried to make it as generic and reusable as possible to cover most use cases of hackathon web apps, without being too specific.

When I need to use socket.io, I really need it, but most of the time - I don't. But more importantly, websockets support is still experimental on most hosting providers. As of October 2013, Heroku supports websockets, but not until you opt-in by running this command:

heroku labs:enable websockets -a myapp

And what if you are deploying to OpenShift? They do support websockets, but it is currently in a preview state. So, for OpenShift you would need to change the socket.io connect URI to the following:

var socket = io.connect('http://yoursite-namespace.rhcloud.com:8000');

Wait, why is it on port 8000? Who knows, and if I didn't run across this blog post I wouldn't even know I had to use port 8000.

I am really glad that Heroku and OpenShift at least have a websockets support, because many other PaaS providers still do not support it. Due to the aforementioned issues with websockets, I cannot include socket.io as part of the Hackathon Starter. For now... If you need to use socket.io in your app, please continue reading.

First you need to install socket.io:

npm install socket.io --save

Replace var app = express(); with the following code:

var app = express();
var server = require('http').Server(app);
var io = require('socket.io')(server);

I like to have the following code organization in app.js (from top to bottom): module dependencies, import controllers, import configs, connect to database, express configuration, routes, start the server, socket.io stuff. That way I always know where to look for things.

Add the following code at the end of app.js:

io.on('connection', function(socket) {
  socket.emit('greet', { hello: 'Hey there browser!' });
  socket.on('respond', function(data) {
    console.log(data);
  });
  socket.on('disconnect', function() {
    console.log('Socket disconnected');
  });
});

One last thing left to change:

app.listen(app.get('port'), function() {

to

server.listen(app.get('port'), function() {

At this point we are done with the back-end.

You now have a choice - to include your JavaScript code in Jade templates or have all your client-side JavaScript in a separate file - in main.js. I will admit, when I first started out with Node.js and JavaScript in general, I placed all JavaScript code inside templates because I have access to template variables passed in from Express right then and there. It's the easiest thing you can do, but also the least efficient and harder to maintain. Since then I almost never include inline JavaScript inside templates anymore.

But it's also understandable if you want take the easier road. Most of the time you don't even care about performance during hackathons, you just want to "get shit done" before the time runs out. Well, either way, use whichever approach makes more sense to you. At the end of the day, it's what you build that matters, not how you build it.

If you want to stick all your JavaScript inside templates, then in layout.jade - your main template file, add this to head block.

script(src='/socket.io/socket.io.js')
script.
    var socket = io.connect(window.location.href);
    socket.on('greet', function (data) {
      console.log(data);
      socket.emit('respond', { message: 'Hey there, server!' });
    });

Note: Notice the path of the socket.io.js, you don't actually have to have socket.io.js file anywhere in your project; it will be generated automatically at runtime.

If you want to have JavaScript code separate from templates, move that inline script code into main.js, inside the $(document).ready() function:

$(document).ready(function() {

  // Place JavaScript code here...
  var socket = io.connect(window.location.href);
  socket.on('greet', function (data) {
    console.log(data);
    socket.emit('respond', { message: 'Hello to you too, Mr.Server!' });
  });

});

And we are done!

Mongoose Cheatsheet

Find all users:

User.find(function(err, users) {
  console.log(users);
});

Find a user by email:

var userEmail = '[email protected]';
User.findOne({ email: userEmail }, function(err, user) {
  console.log(user);
});

Find 5 most recent user accounts:

User
  .find()
  .sort({ _id: -1 })
  .limit(5)
  .exec(function(err, users) {
    console.log(users);
  });

Get total count of a field from all documents:

Let's suppose that each user has a votes field and you would like to count the total number of votes in your database across all users. One very inefficient way would be to loop through each document and manually accumulate the count. Or you could use MongoDB Aggregation Framework instead:

User.aggregate({ $group: { _id: null, total: { $sum: '$votes' } } }, function(err, votesCount) {
  console.log(votesCount.total);
});

Deployment

Once you are ready to deploy your app, you will need to create an account with a cloud platform to host it. These are not the only choices, but they are my top picks. From my experience, Heroku is the easiest to get started with, it will automatically restart your Node.js process when it crashes, zero-downtime deployments and custom domain support on free accounts. Additionally, you can create an account with MongoLab and then pick one of the 4 providers below. Again, there are plenty of other choices and you are not limited to just the ones listed below.

1-Step Deployment with Heroku

- Download and install [Heroku Toolbelt](https://toolbelt.heroku.com/) - In terminal, run `heroku login` and enter your Heroku credentials - From *your app* directory run `heroku create` - Run `heroku addons:create mongolab`. This will set up the MongoLab add-on and configure the `MONGOLAB_URI` environment variable in your Heroku app for you. - Lastly, do `git push heroku master`. Done!

Note: To install Heroku add-ons your account must be verified.


- Open [mongolab.com](https://mongolab.com) website - Click the yellow **Sign up** button - Fill in your user information then hit **Create account** - From the dashboard, click on **:zap:Create new** button - Select **any** cloud provider (I usually go with AWS) - Under *Plan* click on **Single-node (development)** tab and select **Sandbox** (it's free) - *Leave MongoDB version as is - `2.4.x`* - Enter *Database name** for your web app - Then click on **:zap:Create new MongoDB deployment** button - Now, to access your database you need to create a DB user - Click to the recently created database - You should see the following message: - *A database user is required to connect to this database.* **Click here** *to create a new one.* - Click the link and fill in **DB Username** and **DB Password** fields - Finally, in `.env` instead of `mongodb://localhost:27017/test`, use the following URI with your credentials: - `db: 'mongodb://USERNAME:[email protected]:27479/DATABASE_NAME'`

Note: As an alternative to MongoLab, there is also Compose.

- First, install this Ruby gem: `sudo gem install rhc` 💎 - Run `rhc login` and enter your OpenShift credentials - From your app directory run `rhc app create MyApp nodejs-0.10` - **Note:** *MyApp* is the name your app (no spaces) - Once that is done, you will be provided with **URL**, **SSH** and **Git Remote** links - Visit provided **URL** and you should see the *Welcome to your Node.js application on OpenShift* page - Copy and and paste **Git Remote** into `git remote add openshift YOUR_GIT_REMOTE` - Before you push your app, you need to do a few modifications to your code

Add these two lines to app.js, just place them anywhere before app.listen():

var IP_ADDRESS = process.env.OPENSHIFT_NODEJS_IP || '127.0.0.1';
var PORT = process.env.OPENSHIFT_NODEJS_PORT || 8080;

Then change app.listen() to:

app.listen(PORT, IP_ADDRESS, function() {
  console.log("Express server listening on port %d in %s mode", PORT, app.settings.env);
});

Add this to package.json, after name and version. This is necessary because, by default, OpenShift looks for server.js file. And by specifying supervisor app.js it will automatically restart the server when node.js process crashes.

"main": "app.js",
"scripts": {
  "start": "supervisor app.js"
},
  • Finally, you can now push your code to OpenShift by running git push -f openshift master
  • Note: The first time you run this command, you have to pass -f (force) flag because OpenShift creates a dummy server with the welcome page when you create a new Node.js app. Passing -f flag will override everything with your Hackathon Starter project repository. Do not run git pull as it will create unnecessary merge conflicts.
  • And you are done!

  • Login to Windows Azure Management Portal
  • Click the + NEW button on the bottom left of the portal
  • Click COMPUTE, then WEB APP, then QUICK CREATE
  • Enter a name for URL and select the datacenter REGION for your web site
  • Click on CREATE WEB APP button
  • Once the web site status changes to Running, click on the name of the web site to access the Dashboard
  • At the bottom right of the Quickstart page, select Set up a deployment from source control
  • Select Local Git repository from the list, and then click the arrow
  • To enable Git publishing, Azure will ask you to create a user name and password
  • Once the Git repository is ready, you will be presented with a GIT URL
  • Inside your Hackathon Starter directory, run git remote add azure [Azure Git URL]
  • To push your changes simply run git push azure master
  • Note: You will be prompted for the password you created earlier
  • On Deployments tab of your Windows Azure Web App, you will see the deployment history

  • Go to Codename: Bluemix to signup for the free trial, or login with your IBM id
  • Install Cloud Foundry CLI
  • Navigate to your hackathon-starter directory and then run cf push [your-app-name] -m 512m command to deploy the application
  • Note: You must specify a unique application name in place of [your-app-name]
  • Run cf create-service mongodb 100 [your-service-name] to create a MongoDB service
  • Run cf bind-service [your-app-name] [your-service-name] to associate your application with a service created above
  • Run cf files [your-app-name] logs/env.log to see the *environment variables created for MongoDB.
  • Copy the MongoDB URI that should look something like the following: mongodb://68638358-a3c6-42a1-bae9-645b607d55e8:[email protected]:10123/db
  • Then set it as an environment variable for your application by running cf set-env [your-app-name] MONGODB [your-mongodb-uri]
  • Run cf restart [your-app-name] for the changes to take effect.
  • Visit your starter app at http://[your-app-name].ng.bluemix.net
  • Done!

Note: Alternative directions, including how to setup the project with a DevOps pipeline are available at http://ibm.biz/hackstart. A longer version of these instructions with screenshots is available at http://ibm.biz/hackstart2. Also, be sure to check out the Jump-start your hackathon efforts with DevOps Services and Bluemix video.

The data model is comprised of all the Mongoose schemas in the models directory. In order to ensure stability, there are a few rules for modifying the data model. Please find the appropriate heading and follow the recommended steps.

Adding A Schema

Adding an object is always safe. New schemas may have required properties without default values because no data for this Schema exists in production yet. Adding default values is recommended for all properties that are not required.

Adding A Property

All new properties must have a default value. Without a default value, each brigade will need to run a migration on their MongoDb instance. This is a high cost that we can easily avoid. If a default value does not make sense, consider adding a new schema and performing a query instead.

Deleting A Property

Remove the property and all code that references the property in the same commit. This ensures that each commit can be safely and atomically pushed to production.

Deleting A Schema

Remove the schema and all code that references the property in the same commit. This ensures that each commit can be safely and atomically pushed to production.

Refer to #41 for context + discussion

pre-alpha versioning: 0.x.x

Alpha #51

1:1 feature parity with the current CfSF website. We build it as a replica, using the tools we have and implementing the appropriate content-management systems to maintain it as it is, but using the new auth and adding an admin backend. This will take the load off the W&T team for updating the content of the website, and at the very least, will be a good standalone application in its own right. Architecture should be built with additional features in mind, but focused on current website functionality. During this time, we should also be reaching out to other brigades, and seeing what it would take to entice them to make the move to brigadehub, and start adding those features to the next iteration enhancements.

The scope/goal for this release would be to replace the current CfSF website fulltime. Timeline currently is April 21st (refer milestones)

versioning: 1.x.x-alpha, 1.x.x-alpha.x, etc. (following semver)

List of features
  • A working landing page + CMS updating #54
  • A working about page + CMS updating #55
  • A working events page + CMS updating #56
  • A working blogging platform + CMS updating #57
  • A working projects list + CMS updating #58

Beta

We expand out features to include user onboarding, project matching, and any other additional functionality not currently present in website. This would include 1-click deployment to heroku, docker imaging, and any other devops-easing functionality as well. Theming should be included in this release. We should also be reaching out to other brigades during this time to see if anyone is willing to start playing with the application, and getting quality beta feedback for usage/features.

The scope/goal of this release would be to replace between 2-5 brigades websites, or at least have them install it to play with it. Timeline June 9th 2016

versioning: 1.x.x-beta, 1.x.x-beta.x, etc. (following semver)

Live

We expand our outreaching efforts, and try to get brigades across the globe to adopt the platform, actively talking with other leadership teams, and making sure to support/help when issues arise. We should be expanding out the theme library, with up to 5-6 separate themes for brigades to choose from, and perfect the deployment pipeline for the application. Additional discovery and feature expansion can be made after this point.

The scope/goal of this release, while lofty, would be to replace the websites of all brigades, or as many as we can get to adopt it. Timeline Aug 16th 2016.

versioning: 1.x.x-rc.1, 1.x.x, etc. (following semver)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment