4
\$\begingroup\$

I've been using Ruby for about a year now, mostly for small scripts with uses ranging from data munging to DevOps tasks. However, I am not intimately familiar with OOP (I appreciate the functional programming approach I'm often able to take with Ruby) and so I think my code is harder to read and debug than it could be.

I do not think my code is idiomatically Ruby, and I think I'm using a procedural style. I've included a recently-written Ruby script as an example, but at what point should one refactor into a OOP style? Is it always helpful to begin with a OOP structure? And I mean helpful in two senses: the sense that it will be easier for other team members to read the code; and also in the sense that it will make the program more robust and easier to debug.

The script below was intended as a small auditing tool. It takes a CSV (in two possible formats) and compares it to a database, but it has to do a lot of munging and matching because the results are not easily directly comparable.

#!/usr/bin/env ruby

require 'csv'
require 'date'
require 'json'
require 'mysql2'
require 'pp'

# Command line argument handling
if ARGV.length < 2
  puts "Usage: <CSV path or filename> <User's screenname>"
  puts "Outputs JSON of results we do not have in our database"
  exit
end

ARGV.each do |arg|
  # If the arg is a file that exists...
  File.file?(arg) ? @file = arg : @user = arg
end


def redacted?(filename)
  /redacted/.match(filename)
end

def make_redacted(csv)
  header = csv.slice!(0).map {|x| x.split.join.downcase.gsub(/\(\$\)/, '').to_sym}
  fd_csv = csv.map {|row| Hash[header.zip(row)]}
  fd_csv.map do |row|
    row[:name] = row[:title]
    row[:hashtag] = row[:sport]
    row[:opponent_score] = row[:oppscore]
    row[:place_finished] = row[:position].to_i
    row[:prize] = row[:winnings].to_f
    row[:date] = Date.parse(row[:date]).strftime('%Y-%m-%d')
    #row[:opponent_name] = row[:opponent]
    row[:score] = row[:score].to_f
    row[:entrants] = row[:entries].to_i
    row[:buy_in] = row[:entry].to_f
    row.delete_if { |k,v| [:title, :entries, :sport, :oppscore, :position,
     :opponent].include? k }
   end
 end

 def make_other_redacted(csv)
  ## Same sort of deal, except mapping other fields
  return dd_csv
end

# Read the database into memory too
def our_results
  # Set site ID to magic int
  site_id = redacted?(@file) ? 2 : 20
  db_hash = []
  client = Mysql2::Client.new(:host     => "hostname",
    :username => "etc",
    :password => "etc",
    :port     => etc,
    :database => "etc"
    )
  sql = "long sql query with some #{string_interpolation}"

  # Had to turn off casting due to some issues in the DB results
  db_results = client.query(sql, :symbolize_keys => true, :cast => false).each do |row|
    # Here I'm mapping hash names again
    row[:score] = row[:score].to_f
    # etc...
    db_hash << row
  end
  db_hash
end

# Takes a single object and an array
def match_redacted?(source, target)
  target.detect{|x| source[:prize] == x[:prize] &&
    # Match based on a list of hash values matching 
  }
end

def match_other_redacted?(source, target)
  target.detect{|x| source[:prize] == x[:prize] &&
    # Same thing, different matching criteria
  }
end 

def get_remainder(source, target)
  if redacted?(@file)
    source.delete_if {|s| match_redacted?(s, target)}
  else
    source.delete_if {|s| match_other?(s, target)}
  end
end


db_results = our_results()
# Read the csv into memory, fix it up
csv = CSV.read(@file)
source = fanduel?(@file) ? make_redacted(csv) : make_other_redacted(csv)
# Delete the results from the source that we match in the DB
remainder = get_remainder(source, db_results)
unless remainder.empty?
  puts JSON.generate(remainder)
end
\$\endgroup\$
  • 1
    \$\begingroup\$ The desire to improve code is implied for all questions on this site. Question titles should reflect the purpose of the code, not how you wish to have it reworked. See How to Ask. \$\endgroup\$ – Jamal Jun 26 '15 at 18:00
  • \$\begingroup\$ Thanks for the link. It doesn't seem like this question is able to be asked on this site. \$\endgroup\$ – Ryan Roberts Jun 26 '15 at 20:15
  • 1
    \$\begingroup\$ Don't misunderstand: The question's fine for this site in the sense that there is reviewable code there. However, if you only want to know "how do I make this object-oriented?" then it's a straight-up "how do I...?"-question and not really a request for a review. A review can touch on any aspect of the code, so a valid review could just look at the code you have and recommend improvements without ever discussing object-orientation - or even advise against it. Or it could do the other thing, and even provide new code. \$\endgroup\$ – Flambino Jun 26 '15 at 21:39
  • \$\begingroup\$ You do not need OOP. \$\endgroup\$ – Nakilon Jun 27 '15 at 19:37
3
\$\begingroup\$

Never use OOP just because people say it's cool. If you do not know how to apply it, that means you just do not need it in this task. That's it. Forget about OOP. Involving it without much need only makes code worse and unreadable.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.