unRob · October 20, 2017 04:27
diff --git a/amex-vs-uber.rb b/amex-vs-uber.rb
 require 'csv'
 require 'json'

 if ARGV.count == 0
  puts <<-USAGE
  Finds uber phantom charges on amex CSV's by comparing them to your manually-scraped uber trips

  \e[1mUsage\e[0m:
  ruby #{$0} CSV TXT

  CSV: is the one provided by AMEX, clean of non-uber transactions (search for "uber" transactions on amex's website, limit by date)
  TXT: is what you get when you paste your collapsed trip list from https://riders.uber.com/trips and remove all cancelled trips. lines look like

    30/02/42  DriverName  $37.93  uberY GhostTown •••• 0420

  \e[1mOutput\e[0m:
  Outputs the amount of matching transactions, the amount of phantom transactions and the charge ids of them.

  Then you go to uber's help website and search for \"My account has an unrnecognized charge\" and hope for the best.
  https://help.uber.com/h/fe547761-4384-42d4-8531-4cfb0e0e523e
 USAGE
  exit 99
 end

 class PrintableDate < Date
  def around? other
    (to_time - other.to_time).abs <= 864000
  end
 end

 Trip = Struct.new(:date, :amount, :id, :usd, :recognized?)

 known = open(ARGV[1]).read.split("\n").map { |d|
  # AA$0,000.00
  amount = d.match(/[A-Z]*\$[\d,.]+/)[0]
  # Converts poor date format choices to sane, parseable ones
  m,d,y = d.split(' ').first.split('/').map(&:to_i)
  [PrintableDate.new(2000+y, m, d), amount]
 }.sort {|a, b| a[0] <=> b[0]}

 trips = CSV.foreach(ARGV[0]).map do |row|
  next if row[2].include?('*EATS')

  desc = row[10]
  id = desc.split(' ', 2).first
  # There's many ways of fixing poor date formats, here's another one
  date_components = row[0][0,10].split('/', 3).map(&:to_i)
  date = PrintableDate.new(date_components.pop, *date_components)
  amount = "MX$"+desc.match(/Spend Amount:([\d.,]+)/)[1] rescue '$'+row[7].to_s

  # get the first known transaction to match
  needle = known.select { |(d,a)| amount == a && date.around?(d) }.first
  if needle
    date = needle[0]
    known.delete needle
  end

  Trip.new(date, amount, id, row[7].to_f, !needle.nil?)
 end.reject(&:nil?).sort { |a,b| a.date <=> b.date }

 unless known.empty?
  puts "Could not find charges for known trips:"
  puts known.map {|(date, amount)| "#{date} - #{amount}"}.join("\n")
  exit 3
 end

 if ENV['CSV']
  puts %w{date id usd ✓?}.join(',')
  puts trips
    .map {|t| [t.date, t.id, t.usd, t.recognized? ? 1 : 0].join(',') }
    .join("\n")
 else
  total_usd = -> (txns) { txns.reduce(0.0) { |sum, tx| sum + tx.usd }.round(2) }
  unrecognized = trips.reject(&:recognized?)
  puts "✓: $#{total_usd.(trips.select(&:recognized?))}"
  puts "𐄂: $#{total_usd.(unrecognized)}"
  puts "?: #{unrecognized.map(&:id).join(',')}"
  exit unrecognized.count
 end
	require 'csv'
	require 'json'

	if ARGV.count == 0
	puts <<-USAGE
	Finds uber phantom charges on amex CSV's by comparing them to your manually-scraped uber trips

	\e[1mUsage\e[0m:
	ruby #{$0} CSV TXT

	CSV: is the one provided by AMEX, clean of non-uber transactions (search for "uber" transactions on amex's website, limit by date)
	TXT: is what you get when you paste your collapsed trip list from https://riders.uber.com/trips and remove all cancelled trips. lines look like

	30/02/42 DriverName $37.93 uberY GhostTown •••• 0420

	\e[1mOutput\e[0m:
	Outputs the amount of matching transactions, the amount of phantom transactions and the charge ids of them.

	Then you go to uber's help website and search for \"My account has an unrnecognized charge\" and hope for the best.
	https://help.uber.com/h/fe547761-4384-42d4-8531-4cfb0e0e523e
	USAGE
	exit 99
	end

	class PrintableDate < Date
	def around? other
	(to_time - other.to_time).abs <= 864000
	end
	end

	Trip = Struct.new(:date, :amount, :id, :usd, :recognized?)

	known = open(ARGV[1]).read.split("\n").map { \|d\|
	# AA$0,000.00
	amount = d.match(/[A-Z]*\$[\d,.]+/)[0]
	# Converts poor date format choices to sane, parseable ones
	m,d,y = d.split(' ').first.split('/').map(&:to_i)
	[PrintableDate.new(2000+y, m, d), amount]
	}.sort {\|a, b\| a[0] <=> b[0]}

	trips = CSV.foreach(ARGV[0]).map do \|row\|
	next if row[2].include?('*EATS')

	desc = row[10]
	id = desc.split(' ', 2).first
	# There's many ways of fixing poor date formats, here's another one
	date_components = row[0][0,10].split('/', 3).map(&:to_i)
	date = PrintableDate.new(date_components.pop, *date_components)
	amount = "MX$"+desc.match(/Spend Amount:([\d.,]+)/)[1] rescue '$'+row[7].to_s

	# get the first known transaction to match
	needle = known.select { \|(d,a)\| amount == a && date.around?(d) }.first
	if needle
	date = needle[0]
	known.delete needle
	end

	Trip.new(date, amount, id, row[7].to_f, !needle.nil?)
	end.reject(&:nil?).sort { \|a,b\| a.date <=> b.date }

	unless known.empty?
	puts "Could not find charges for known trips:"
	puts known.map {\|(date, amount)\| "#{date} - #{amount}"}.join("\n")
	exit 3
	end

	if ENV['CSV']
	puts %w{date id usd ✓?}.join(',')
	puts trips
	.map {\|t\| [t.date, t.id, t.usd, t.recognized? ? 1 : 0].join(',') }
	.join("\n")
	else
	total_usd = -> (txns) { txns.reduce(0.0) { \|sum, tx\| sum + tx.usd }.round(2) }
	unrecognized = trips.reject(&:recognized?)
	puts "✓: $#{total_usd.(trips.select(&:recognized?))}"
	puts "𐄂: $#{total_usd.(unrecognized)}"
	puts "?: #{unrecognized.map(&:id).join(',')}"
	exit unrecognized.count
	end