Potassium

A Rails application generator from Platanus, inspired by Suspenders.

Installation

You have to install Potassium globally:

$ gem install potassium

Usage

Starting a new project from scratch

Use the potassium create command to create a new project:

$ potassium create <project-name>

It's important to note that it will perform a version check before running to ensure that you're using the latest potassium. Also, if you feel that it's too slow, you may need to update rubygems: gem update --system.

Adding recipes to an existing project

Use the potassium install command to add a recipe to a project:

$ potassium install devise

You can force an already installed recipe by passing the --force argument

$ potassium install devise --force

You can run the command on its own to view all the available recipes and select one:

$ potassium install

What's inside Potassium?

Potassium Rails apps includes the following gems and technologies:

Ruby Set the project ruby version based on http://ruby.platan.us/latest
dotenv load environmental variables in development
Yarn for frontend assets packages.
EditorConfig for keeping all our editor configurations the same.
pry and pry-byebug for a less painful debugging experience.
RSpec for unit and integration testing.
FactoryBot for test factories.
Guard for continuous testing and other watch-related tasks.
AWS-SDK for file uploads, sdks, etc and because we use AWS.
Puma to serve HTTP requests
Rack Timeout to abort requests that are taking too long
Tzinfo-Data for update timezone information
Faker for create development data

The following optional integrations are added too:

PostgreSQL or MySQL for the database.
Devise for authentication.
ActiveAdmin for admin interfaces.
ActiveAdminAddons for some help with ActiveAdmin.
Pundit for role-based authorization.
DelayedJob to execute longer tasks in the background.]
Sidekiq a simple, efficient background processing for Ruby.
Sidekiq-scheduler to run scheduled processes
Mailing configuration for AWS SES and Sendgrid with recipient interceptor support
Sentry to monitor exceptions and errors
Vue.js or Angular 2 for frontend development

A few more things are added to the project:

A low database connection pool limit
Setup continuous integration in CircleCI to run tests.
Create the github repository for the project (it used hub under the hood)
A bin/setup script to setup things on a newly cloned project
A bin/cibuild script to run continuous integration build on CI
A db:fake_data:load rake task to load fake data for development

API support

The optional API support includes:

Responders for dry-ing our api controllers.
Versionist for some flexible api versioning.
ActiveModel::Serializers for record serialization.
Simple Token Authentication for stateless API authentication.

Heroku

When you choose to deploy to heroku a few extra things are added for the project.

Adds the Rails Stdout Logging gem to configure the app to log to standard out, which is how Heroku's logging works.
Adds a Procfile to define the processes to run in heroku
Setup continuous integration using docker and herokuish to maintain better parity between testing and production environments
Adds a .buildpacks file with the default buildpacks to use. It use the following buildpacks:
Adds a bin/release file with the release phase script to run specific tasks before the app is deployed completely, for example rails db:migrate.

index	buildpack	description
1.	nodejs	to support javascript package management with `yarn` and `webpack` based asset compiling
2.	ruby-version	to support the use of `.ruby-version` file to instruct heroku which ruby version to use
3.	ruby	the base buildpack to run ruby applications

Also the heroku applications are created

Creates a staging and production applications
Creates a pipeline and assign the above application to the staging and production stages.
Setup initial configuration variables
Set the application buildpack to the multi-buildpack

You'll need to manually

Connect the pipeline with the github repository
Assign a branch to each stage for auto deployments
Enable deploy after CI pass

Continuous Integration

In order to CicleCI start building the project on each push you need tell circle ci. Go to https://circleci.com/add-projects, choose the repository from the list and hit Build Project

Hound CI

In order to Hound start checking your project's PRs you need enable that repository. Go to https://monkeyci.platan.us, choose the repository from the list and hit Activate

Development Tools

Creating a new Test project from scratch

This is useful when you are adding new recipes to Potassium, and you want to use the potassium create command, to check the new functionality without pain. To achieve this you need to run, In the Potassium's root path, the bin/potassium_test create instead of potassium create command. This command, will do the same as potassium create but first:

Will drop any existent database of the test app.
Will remove the directory containing a previous version of the test app.

It's important to remember that bin/potassium_test create:

Does not receive an app_path param. It always creates the test project inside /tmp/dummy_app
Can receive the same options as potassium create.
Runs with options with a default value. This is to avoid the "asking part" of the creation process. You need to enable what you want to test like this: $ bin/potassium_test create --clockwork

Contributing

If you want to add functionality please go to the contributing

Credits

Thank you contributors!

Platanus

potassium is maintained by platanus.