Simple CSV import to Elasticsearch with Tire

generate.rb

# Usage:
#
#     $ ruby generate.rb > people.csv

require 'time'
require 'faker'
require 'oj'

COUNT = (ENV['COUNT'] || 1000).to_i

people = []

1.upto(COUNT) do |i|
  people << {
    'first_name' => Faker::Name.first_name,
    'last_name'  => Faker::Name.last_name,
    'phone'      => Faker::PhoneNumber.cell_phone,
    'department' => ['Marketing', 'Development', 'Production'].sample,
    'birthday'   => Time.at(0.0 + rand * ((Time.now-631152000).to_f)).iso8601
  }
end

STDOUT.puts 'FirstName,LastName,Phone,Department,Birthday'

people.each_with_index do |person, i|
  STDOUT.puts [
    person['first_name'],
    person['last_name'],
    person['phone'],
    person['department'],
    person['birthday']
  ].join(',')
end

import.rb

# Usage:
#
#     $ ruby import  /path/to/file.csv

require 'tire'
require 'csv'

path       = ARGV[0]
batch_size = 100
buffer     = []
index_name = ENV['INDEX'] || File.basename(path, '.*')

unless path
  puts "Usage: #{__FILE__} /path/to/csv"
  exit 1
end

CSV.foreach path, headers: true do |row|

  # Add line as JSON into buffer.
  #
  buffer << row.to_hash

  # When we hit the batch boundary...
  if buffer.size % batch_size == 0

    # ... load batch into Elasticsearch ...
    Tire.index index_name do
      STDERR.puts import(buffer), '-'*80
    end

    # ... and empty the buffer.
    buffer = []
  end

end

# Import any rest.
#
Tire.index index_name do
  STDERR.puts import(buffer), '-'*80
end unless buffer.empty?