Record-Conjurer

Record-Conjurer is an object-relational mapper inspired by Ruby on Rails' Active Record. This ORM allows Ruby classes to connect to SQLite3 relational database tables. A sample database is provided for experimentation in Pry.

Table of contents

• Setup
• Usage
• Things to Try
• Demo Classes
• Sample Database Schema
• Future Implementations

Setup (for sample DB)

1. Clone the repo
2. Run bundle install in the root directory
3. Run bundle exec pry in the root directory
4. Have fun poking about!

Usage

All classes inherit from RecordConjurer, thus will have the following interfaces

• ::all returns all of the records in the associated table as an array
• ::table_name returns the name of the table use in the db
• ::new(field_hash) will create a new instance and will set fields to value provided in the optional hash
• ::find(id) returns a record with that id, or nil if not found
• #save will add a new record to the db with the values provided. Returns the new id.
• #field_name returns the value of the field for that record
• #field_name= sets the value of the field for that record
• #update saves that changes made to a record
• #attributes returns a hash with field:value pairs
• #attribute_values returns an array of all the values

Things to Try

  Artist.all #=> Array of artists
  Track.all #=> Array of tracks
  accept_band = Artist.find(2) #=> #<Artist:0x007fcbc01d7e80 @attributes={:id=>2, :Name=>"Accept"}>
  accept_band.albums #=> Array of albums belonging to accept_band
  aerosmith = Artist.where(name: "Aerosmith").first #=> #<Artist:0x007fcbbfaa5c20
                                                    #      @attributes={:id=>3, :Name=>"Aerosmith"}>
  aerosmith_album = aerosmith.albums.first  #=> #<Album:0x007fcbc11a5768
                                            #     @attributes={
                                            #       :id=>5,
                                            #       :Title=>"Big Ones",
                                            #       :ArtistId=>3}>
  aerosmith_album.tracks #=> Array of tracks
  song_where = {name: "Dude (Looks Like A Lady)"}
  song = Track.where(song_where).first  #=> #<Track:0x007fcbc144b210
                                        #      @attributes=
                                        #       {:id=>27,
                                        #        :Name=>"Dude (Looks Like A Lady)",
                                        #        :AlbumId=>5,
                                        #        :MediaTypeId=>1,
                                        #        :GenreId=>1,
                                        #        :Composer=>"Steven Tyler, Joe Perry, Desmond Child",
                                        #        :Milliseconds=>264855,
                                        #        :Bytes=>8679940,
                                        #        :UnitPrice=>0.99}>
  song.artist #=> #<Artist:0x007fcbc02c7f98 @attributes={:id=>3, :Name=>"Aerosmith"}>
  aerosmith.Name = 'not aerosmith' #=> "not aerosmith"
  aerosmith.update #=> 0
  Artist.find(3) #=> #<Artist:0x007f921bbae2d0 @attributes={:id=>3, :Name=>"not aerosmith"}>
  new_artist = Artist.new(Name: 'Sonic') #=> #<Artist:0x007f921bb3d5d0 @attributes={:Name=>"Sonic"}>
  new_artist.save #=> 276
  Artist.all.last #=> #<Artist:0x007f921b94f840 @attributes={:id=>276, :Name=>"Sonic"}>

Demo Classes

These are set up in demo.rb and autoloaded via .pryrc when pry is ran

• Albums
• Artists
• Generes
• MediaTypes
• Playlists
• PlaylistTracks
• Tracks

Sample Database Schema

Database is a subset of the chinook database

Albums

column name	data type	details
id	integer	not null, primary key
Title	nvarchar(160)	not null
ArtistId	integer	not null, foreign key

• index on ArtistId
• ArtistID references Artists

Artists

column name	data type	details
id	integer	not null, primary key
Name	nvarchar(120)	not null, indexed, unique

Genres

column name	data type	details
id	integer	not null, primary key
Name	nvarchar(120)

media_types

column name	data type	details
id	integer	not null, primary key
Name	nvarchar(120)

Playlists

column name	data type	details
id	integer	not null, primary key
Name	nvarchar(120)

playlist_tracks

column name	data type	details
PlayListId	integer	not null, foreign key, primary key
TrackId	integer	not null, foreign key

• index on [:PlayListId, :TrackId], unique: true
• PlayListId references Playlists
• TrackId references Tracks

Tracks

column name	data type	details
id	integer	not null, primary key
name	nvarchar(200)	not null, indexed, unique
AlbumId	integer	foreign key
MediaTypeId	integer	not null, foreign key
GenreId	integer	foreign key
Composer	nvarchar(220)
Milliseconds	integer	not null
Bytes	integer
UnitPrice	numeric(10,2)	not null

• AlbumId references Albums
• MediaTypeId references MediaTypes
• GenreId references Genres

Future Implementations

Forthcoming features include:

• The ability to select specific fields vs *
• The ability to include relations to prevent n+1 queries
• The abiltity to represetn has_many_through relations
• The ability to see the generated SQL statements