ridepilot: A Ruby repository from mithkum

# RidePilot

RidePilot is a paratransit trip scheduling application.

RidePilot is a project of Ride Connection, a Portland-area community transportation provider. It is primarily written by hackers at OpenPlans.

## Dependencies (as of 2015-02-05)

* Postgresql 9.3
* PostGIS 2.1
* Ruby 2.1.4
* Rails 4.2
* Imagemagick

## Setting up a development environment

Because of the complexity involved with managing multiple versions of the same database, I have found the cleanest way to setup a development environment is to use Vagrant, Chef and VirtualBox for the database, and RVM for local Ruby and Gem management. This gives you an isolated environment that is easy to provision. If you don't want to go that route, you can setup all of the database dependencies yourself on your local machine. YMMV with either approach.

### Provisioning a Database VM with Vagrant

Using this method, you will set your Rails app to connect to your VM's database server. All of your development, testing, and other console commands can be run locally on the host.

1. Follow the "Using Vagrant for Database Management in Development" document in the Rideconnection Recipes repo.

2. Continue to "Common Setup Steps"

### Provisioning Locally (without Vagrant)

1. Install the required versions of Postgresql, PostGIS, and any other system packages required for your setup

2. To set up PostgreSQL for use with RidePilot, you will need the `fuzzystrmatch` library (included in the postgresql-contrib library in Ubuntu), and `postgis`. This adds support for `dmetaphone`, which is how we phonetically match names. Run `psql` and then the following commands:

-- Create a new database
CREATE DATABASE template_postgis;

-- Make it a template database
UPDATE pg_database SET datistemplate = TRUE WHERE datname = 'template_postgis';

-- Connect to new database and install the pl/pgsql language
\c template_postgis

CREATE LANGUAGE plpgsql;

-- Install PostGIS (your file paths may vary)
\i /opt/local/share/postgresql93/contrib/postgis-2.1/postgis.sql
\i /opt/local/share/postgresql93/contrib/postgis-2.1/spatial_ref_sys.sql
GRANT ALL ON geometry_columns TO PUBLIC;
GRANT ALL ON geography_columns TO PUBLIC;
GRANT ALL ON spatial_ref_sys TO PUBLIC;

-- vacuum freeze: it will guarantee that all rows in the database are
-- "frozen" and will not be subject to transaction ID wraparound
-- problems.
VACUUM FREEZE;

-- You'll also want to set up the fuzzy string matching functions
\i /opt/local/share/postgresql93/contrib/fuzzystrmatch.sql

3. Create the database user and development and test databases per the database config

4. Continue to "Common Setup Steps"

### Common Setup Steps

1. Install RVM

2. Install any additional libraries reported by `rvm requirements`

3. Install the required version of Ruby:

rvm list known
rvm install ruby-xxx.xxx

4. Install bundler

rvm gemset use ruby-xxx.xxx@global
gem install bundler

5. `cd` into (or out and back into) the project directory to pickup the RVM configuration files. Follow any instructions for creating the gemset.

6. Install the gem bundle. In the project dir, run:

bundle install

7. Make sure your `config/database.yml` contains the line in the section named `common: &common`:

template: template_postgis

8. Copy `config/app_config_template.yml` to `config/app_config.yml`. You can leave it as-is in most cases.

9. Setup the database:

bin/rake db:setup

Among other things, this will seed the database with some initial data. You can immediately log in using the email address "admin@rideconnection.org" and the password "password 1". *Be sure to change the initial password if this is a production server.*

10. Make sure you can start the app without error

bin/rails server

Then make sure you can browse the website at `http://localhost:3000` without error

11. Make sure all of the automated tests pass

bin/rake db:test:prepare
bin/rake test
bin/rake spec

## Deployment

This application uses capistrano for deployment. Check out `config/deploy.rb` and `config/deploy/*` for basic deployment recipes and configuration.

Deployment uses key-based authentication. To deploy, you'll need to add your public key on the staging/production servers so you can run commands as the "deployer" user.

To set this up, talk to another developer to get your public key on the machines. If you need to do system administration on the servers, you'll need your own user account set up as well.

Once you have SSH access as deployer, you can deploy:

cap [staging|production] deploy

Remember to push your changes to the main repository first, since the deploy process pulls from there. Database migrations are currently performed by hand after deploying.

### Deployment Gotchas

(TODO verify this is true) Ridepilot staging is on the production server. That means when you go into rails/ridepilot-staging, you actually have to set `RAILS_ENV=production` if you're going to run any commands. It will not affect the real production app if you're in the right directory.

mithkum/ridepilot