Share
Embed
Using MongoDB with Ruby on Rails - Ryan Fischer, 20spokes
This talk will go through using MongoDB and Ruby on Rails to build a web application. It will introduce using Mongoid an Object Document Mapper(ODM)... that provides a similar API to Active Record while utilizing MongoDB's schemaless and performant document-based design. The talk will cover model design and relations with MongoDB, querying using Mongoid's API, and testing the application with MongoDB and Mongoid.
mongodb
November 04, 2011
Transcript
30 slides
-
Using MongoDB with Ruby on Rails presented by Ryan Fischer,
Founder of 20spokes -
What is covered • Why we use MongoDB and Ruby
on Rails • Choices made for Object Mapper • Simple to get started! -
Using MongoDB with Ruby on Rails • Fast in-place updates
with atomic modifiers • Ruby on Rails is productive! • Object Mappers available for MongoDB • Mongo Ruby Driver -
Mobile / Web • Ruby on Rails makes REST easy,
making APIs easier. • Render responses as text, JSON, or XML • Geospatial indexing for location based queries. • Location-centric websites and mobile applications. -
Why we chose MongoDB • Cowrite - collaborative writing web
application • Versioning needed • Originally using GridFS • Travel720 - Gift registry web site • Self contained relations can take advantage of embedding documents -
Mongo Object Mappers for Ruby • MongoMapper • Mongoid •
Mongo ODM • MongoModel -
Why we chose Mongoid • Excellent documentation • Active community
• Compatibility with other projects/gems • Similar API to ActiveRecord • Uses ActiveValidation • Mongoid Extras: Caching, Paranoid Documents, Versioning, Timestamping, Composite Keys -
Compatible Gems with Mongoid • Devise - Authentication solution for
Rails based on Warden. Supports Mongoid out of box. • Carrierwave - simple and flexible way to upload files from Ruby Applications. Supports grid_fs. • Geocoder - complete geocoding solution for Rails. Adds geocoding by street or IP address, reverse geocoding, and distance queries. • Mongoid-rspec - RSpec matchers and macros for Mongoid. -
Getting Started gem "mongoid", "~> 2.3" gem "bson_ext", "~> 1.4"
Include in Gem file Run the install for Mongoid rails generate mongoid:config That’s it! -
mongoid.yml development: host: localhost database: geekbusters_development test: host: localhost database:
geekbusters_test production: host: <%= ENV['MONGOID_HOST'] %> port: <%= ENV['MONGOID_PORT'] %> username: <%= ENV['MONGOID_USERNAME'] %> password: <%= ENV['MONGOID_PASSWORD'] %> database: <%= ENV['MONGOID_DATABASE'] %> # slaves: # - host: slave1.local # port: 27018 # - host: slave2.local # port: 27019 -
Developing with MongoDB/Mongoid • Generating models is the same using
the console as with ActiveRecord. • rails generate model Team name:string city:string location:array • No migrations needed! • Take advantage of embedded documents in models where applicable for increased performance. • Store large files using GridFS -
Fields available for Mongoid • Array • BigDecimal (Stored as
a String) • Boolean • Date • DateTime • Float • Hash • Integer • Range • String • Symbol • Time -
Mongoid Document class Team include Mongoid::Document include Mongoid::Timestamps field :name,
type: String field :city, type: String field :location, :type => Array validates :name, :city, :presence => true end -
Persisting in the Controller def create Team.create(:city => "New York",
:name => "Giants") end def update @team = Team.find(params[:id]) @team = Team.update_attributes(params[:team]) end -
Indexing class Team include Mongoid::Document field :name, type: String index
:name, unique: true end index( [ [ :name, MONGO::ASCENDING ] [ :city, MONGO::ASCENDING ] ] ) Indexing on multiple fields - -
Indexing To create indexes in the database use the rake
task rake db:mongoid:create_indexes Or configure to autocreate in mongoid.yml (not recommended) defaults: &defaults autocreate_indexes: true -
Relations in Models • Associations between models can be embedded
or referenced • NO JOINS! • Objects that are referenced involve a separate query • Embedded documents can be very efficient with reducing queries to one while managing the size of the document • One to One, One to Many, and Many to Many relations available -
Relations in Mongoid - Embedded • Embedded Relations - stored
inside other documents in the database. class Blog include Mongoid::Document embeds_many :posts end class Post include Mongoid::Document embedded_in :blog end -
Embedded Posts { "_id" : ObjectId("4e9ba47fcdffba523f000004"), "name" : "Geekbusters Blog",
"posts" : [ {"content" : "Slimer is right behind you.", _id" : ObjectId("4e9ba47fcdffba523f000005")} ] } -
Polymorphic Behavior class Image include Mongoid::Document embeds_many :comments, as: :commentable
end class Document include Mongoid::Document embeds_many :comments, as: :commentable end class Comment include Mongoid::Document embedded_in :commentable, polymorphic: true end -
Relations in Mongoid - Referenced class Blog include Mongoid::Document has_many
:posts, dependent: :delete end class Post include Mongoid::Document belongs_to :blog end Referenced Relations - stores reference to a document in another collection, typically an id -
Many to Many Relationship class Tag include Mongoid::Document has_and_belongs_to_many :posts
end class Post include Mongoid::Document has_and_belongs_to_many :tags end -
Querying • Queries are of type Criteria, which is a
chainable and lazily evaluated wrapper to a MongoDB dynamic query. • Chainable queries include: all_in all_of also_in and any_of asc desc distinct excludes includes limit near not_in only order_by skip where without -
Query Examples Team.near(location: [20.70, 38.15]).limit(20) Team.any_in(names: ["Giants", "Bears", "Lions"]) Team.find_or_create_by(name:
"Giants") Team.where(name: "Cowboys").destroy_all -
Versioning with Mongoid • Embeds a version of the object
on each save. • Can skip versioning and also set the max versions. • Add to model - include Mongoid::Versioning -
Versioning Example {"_id" : ObjectId("4e9ba7fdcdffba52d6000004"), "content" : "Impossible. I am
too loud to fall asleep to.", "created_at" : ISODate("2011-10-17T03:58:53Z"), "title" : "Impossible to sleep to - this guy.", "updated_at" : ISODate("2011-10-17T03:58:53Z"), "version" : 3, "versions" : [ {"title" : "Who is asleep?", "content" : "Wake up the guy next to you if they are asleep. Thanks.", "version" : 1, "created_at" : ISODate("2011-10-17T03:58:53Z")}, {"title" : "Who is asleep?", "content" : "Impossible. I am too loud to fall asleep to.", "created_at" : ISODate("2011-10-17T03:58:53Z"), "version" : 2 } ] } -
And More • Identity Map • Callbacks • Scopes •
Dirty Tracking -
Testing Rails with MongoDB • RSpec-Mongoid provides test matchers •
RSpec does not refresh the database with each test run. • Database Cleaner to the rescue. It be setup to truncate a database before running tests. Add below after installing the DatabaseCleaner Gem. config.before(:suite) do DatabaseCleaner.strategy = :truncation DatabaseCleaner.orm = "mongoid" end -
Hosting Options • Heroku is a widely used cloud hosting
for Ruby on Rails • MongoHQ and MongoLab both have add on options • Options and pricing are similar