geocoding.rb

require 'csv'
require 'httparty'
require 'json'

header = File.open('input.tsv', &:readline) # get header from input file
header = header.to_s.gsub("\n","") # remove newline characters
header << "\t\t\tlatitude\t\t\tlongitude\n" # format header
File.open("output.tsv", 'w') {|file| file.write(header) } # create output file and append header 

base_url = "http://maps.googleapis.com/maps/api/geocode/json?address="

CSV.foreach("input.tsv", :headers => true, :col_sep => "\t", :encoding => 'windows-1251:utf-8') do |row|

	url = base_url

	# each row contains 5 arrays which contain a column name and the corresponding attribute from the row that is in that column
	# [street_address, street_address_from_row] [unit, unit_from row] [city, city_from_row] [stat, state_from_row] [zip, zip_from_row]

	row.each do |address_element_array| 
		address_element = address_element_array[1] # 0th element in array is the header information, 1st element is value of interest
		if address_element != nil # make sure address element isn't nil (skips header)
			address_element = address_element.split(' ').join('+') # replace spaces with '+' for URL
			url += "+" + address_element # build the URL
		end
	end

	url += "&sensor=false" # ending required by Geocode API

	response = HTTParty.get(url)
	json = JSON.parse(response.body)

	if json['results'].length == 0 # define lat and lng if no results
		lat = ""
		lng = ""
	else
		# the parsed json is a hash that contains an array that contains nested hashes
		lat = json['results'][0]['geometry']['location']['lat'].to_s
		lng = json['results'][0]['geometry']['location']['lng'].to_s
	end

	row = row.to_s.gsub(",","\t").gsub("\n","") # format output file
	row << lat + "\t\t\t" + lng + "\n" # append latitude and longitude to row

	File.open("output.tsv", 'a') do |file|
		file.write(row) 
	end
end