WATCHR

Finally, an app that lets you keep track of your favorite movies.

Quick Start

Simply run the included Docker compose script and navigate to localhost:3000

docker-compose up -d --build

Once there, you'll be able to create an account, log in, and begin using the application.

API

Watchr includes a JSON API with the following endpoints:
GET /movies.json?search={search} - search for movies by title
GET /movies/:id.json - get details about a specific movie
GET /favorites.json - see your favorites
POST /favorites/:id.json - add a favorite

Anyone can access the movie database, but favorites are specific to your account. In order to view them, you'll first need to authenticate by sending a POST request to /authenticate?email=your_email&password=your_password

You'll receive a response with {"auth_token":"a_long_string_of_letters_and_numbers"}

You can now use this auth token by including it in the Authorize header of your requests. Without this header, any API calls will be rejected with an authorization error.

The Movie Database

You may have noticed we have a pretty amazing movie selection, truly unrivaled elsewhere. Well, it's not ours. It comes from OMDb.

If you'd like to make adjustments to the types of movies or how many of them are available, you can simply edit lib/tasks/db_seed.rake.

By default, Watchr includes a generic, free OMDb account. This only allows 1000 movies to be queried per day, though, so it has to be used sparingly. In order to add more movies to your own Watchr application, sign up for an OMDb account, update the api_key variable, and increase the number of pages that are fetched.

Search functionality

In order to provide robust, scalable search with a wide variety of options, Watchr uses Elasticsearch. This is included in docker-compose.yml, and any cluster settings that you'd like to adjust can be changed there.

Running without Docker

You shouldn't do this. But you can. And maybe you have reasons.

Run Postgres:

docker run -d --name pg -e POSTGRES_PASSWORD=postgres -p 5432:5432 postgres

Run Elasticsearch:

docker run --name es \
  --publish 9200:9200 \
  --publish 9300:9300 \
  --env "discovery.type=single-node" \
  --env "cluster.name=movies" \
  --env "cluster.routing.allocation.disk.threshold_enabled=false" \
  --env "ES_JAVA_OPTS=-Xms512m -Xmx512m" \
  docker.elastic.co/elasticsearch/elasticsearch:7.17.7

Note that Elasticsearch versioning is complex, to say the least. We are running the 7.2.1 version of the Rails gem, but the version we need to match is the version of elasticsearch-ruby, which is an implicit dependency that you have to check the Gemfile.lock file for the version of.

Since you're not running Docker Compose, you'll need to manually setup the database and the elasticsearch indices. Typically, this is handled by entrypoint.sh, which you can either make executable and run in its entirety or simply execute each of the commands from inside of it:

bin/rails db:create
bin/rails db:migrate
bin/rake db:seed_movies
bin/rails runner "Movie.__elasticsearch__.create_index!"
bin/rails runner "Movie.import"

Next, you'll need to remove the Docker networking for Postgres and Elasticsearch in config/database.yml and config/initializers/elasticsearch.rb, respectively. For ease of development, these changes are already included and only have to be uncommented to take effect.

With all of this in place, you should now be able to start Rails with ./bin/dev as usual, visit localhost:3000, and access the application.

Development

Several development tools are included, namely pry-byebug for debugging and rspec for testing. While most of the existing functionality depends on Rails core or its associated gem libraries and is outside of the bounds of most tests, some guards are in place around authorization and favoriting logic. You can run them at anytime with the rspec command. To run a specific file and line number, for example the Favorites Request spec to test adding favorites, you can use the following syntax:

rspec ./spec/requests/favorites_spec.rb:11

Note that while Elasticsearch isn't needed for tests, you will need a Postgres database available at localhost:5432. You'll also need to update config/database.yml as above to run the app without containerization.

What's next

There are many fundamental ways that Watchr could improve. Here's our current roadmap:

1. Escape the clutches of the Hollywood elite and add support for
   importing your own movies.
2. Get more information for the movies that we do have. Wouldn't a little
   cover art be nice?
3. Add different levels of ranking since, after all, nothing is as simple
   as "favorite" in this wild and diverse world.

kxcrl/watchr