Seven Languages in Seven Weeks: Ruby, Day 3

This is my 3rd day of Ruby in the Seven Languages in Seven Weeks series of posts. You can find the previous day here.

Ruby, Day 3: Thoughts

The third day combines metaprogramming techniques (define_method, method_missing, and mixins) with what what we learned in the previous chapters (flexible syntax, blocks, yield) to work some magic. Whereas day 1 and 2 showed how Ruby could be more concise and expressive than other languages, this chapter shows some of the capabilities available in Ruby, such as beautiful DSLs and composable designs, that are nearly impossible in stricter languages.

I saw small of examples of this when I was working on the Resume Builder: the profile data I was fetching from the LinkedIn APIs came back as JSON. I wanted to have a nice Ruby class to wrap the JSON data and was able to do this cleanly and concisely using some very simple metaprogramming:

	class LinkedInProfile
	SIMPLE_PROFILE_FIELDS = %w[id summary headline honors interests specialties industry first_name last_name public_profile_url picture_url associations]
	SIMPLE_PROFILE_FIELDS.each do |field|
	define_method(field.to_sym) do
	@json[field]
	end
	end
	def initialize(json)
	@json = json
	end
	end

view raw LinkedInProfile.rb hosted with ❤ by GitHub

Instead of defining dozens of getters and setters as in the LinkedIn API Java Library, I just declared the fields in an array (SIMPLE_PROFILE_FIELDS), looped over them, and used define_method to create the appropriate methods. To be fair, this is kids stuff; if you really want to see metaprogramming shine, take a gander over at ActiveRecord.

Of course, with great power comes great big bullet wounds in the foot. Metaprogramming must be used with more a bit more caution than other programming techniques, as chasing down errors in dynamic methods and trying to discern "magic" can be painful.

Ruby, Day 3: Problems


CSV application

There was only one problem to solve on this day: modify the CSV application (see the original code here and the original output here) to return a CsvRow object. Use method_missing on that CsvRow to return the value for the column given a heading.

	module ActsAsCsv
	def self.included(base)
	base.extend ClassMethods
	end
	module ClassMethods
	def acts_as_csv
	include InstanceMethods
	include Enumerable
	end
	end
	module InstanceMethods
	attr_accessor :headers, :csv_contents
	def initialize
	read
	end
	def read
	@csv_contents = []
	filename = self.class.to_s.downcase + '.csv'
	file = File.new(filename)
	@headers = parse_row file.gets
	file.each do |row|
	@csv_contents << CsvRow.new(@headers, parse_row(row))
	end
	end
	def parse_row(row)
	row.chomp.split(', ')
	end
	def each
	@csv_contents.each { |row| yield row }
	end
	class CsvRow
	def initialize(headers, row)
	@headers = headers
	@row = row
	end
	def respond_to?(sym)
	@headers.index(name.to_s) || super(sym)
	end
	def method_missing name, *args, &block
	index = @headers.index(name.to_s)
	if index
	@row[index]
	else
	super
	end
	end
	end
	end
	end
	class RubyCsv
	include ActsAsCsv
	acts_as_csv
	end
	csv = RubyCsv.new
	puts csv.headers.inspect
	puts csv.csv_contents.inspect
	csv.each { |row| puts "#{row.name}, #{row.age}" }

view raw csv_new.rb hosted with ❤ by GitHub

Using this sample file:

	name	location	age
	Jim	Menlo Park	27
	Bob	Palo Alto	37
	Steve	NYC	28

view raw rubycsv.csv hosted with ❤ by GitHub

The code above will produce the following output:

	$ ruby csv_new.rb
	["name", "location", "age"]
	[#<ActsAsCsv::InstanceMethods::CsvRow:0x25814 @row=["Jim", "Menlo Park", "27"], @headers=["name", "location", "age"]>, #<ActsAsCsv::InstanceMethods::CsvRow:0x25738 @row=["Bob", "Palo Alto", "37"], @headers=["name", "location", "age"]>, #<ActsAsCsv::InstanceMethods::CsvRow:0x2565c @row=["Steve", "NYC", "28"], @headers=["name", "location", "age"]>]
	Jim, 27
	Bob, 37
	Steve, 28

view raw csv_new_output.txt hosted with ❤ by GitHub

Moving on

This was the final day in the Ruby chapter. Join me next time as I work my way through a totally new language: Io.