geocoder_batch_analyzer.rb

#!/usr/bin/env ruby

require 'rubygems'
require 'active_support/all'
require 'yaml'
require 'i18n'
require 'net/http'
require 'digest/sha1'

# from rails 3.2 framework
class Hash
  # Returns a hash that represents the difference between two hashes.
  #
  # Examples:
  #
  #   {1 => 2}.diff(1 => 2)         # => {}
  #   {1 => 2}.diff(1 => 3)         # => {1 => 2}
  #   {}.diff(1 => 2)               # => {1 => 2}
  #   {1 => 2, 3 => 4}.diff(1 => 2) # => {3 => 4}
  def diff(h2)
    dup.delete_if { |k, v| h2[k] == v }.merge!(h2.dup.delete_if { |k, v| has_key?(k) })
  end
end

require File.join(ENV["HOME"],'workspace', 'onlineantrag-rails', 'lib', 'behaviors', 'geocode_verification.rb')

all = []
hashed = []

# pre-filter data (path in hash and deep attributes)
while( entry = gets ) do
  if eval(entry)['tenancy_retail']
    all += eval(entry)['tenancy_retail']['lessee_attributes'].map do |i,k|
      k.slice('address_street','address_snumber', 'address_city', 'address_zip')
    end
  else
    all << eval(entry).slice('address_street', 'address_snumber', 'address_city', 'address_zip')
  end
end

module LoggingOverride

  def self.included(klass)
    klass.send :alias_method_chain, :logger, :hack
  end

  def logger_with_hack
    @logger ||= SimpleLogger
  end

  class SimpleLogger

    def self.method_missing(name, *args)
      puts args.join
    end

  end

end

Behaviors::GeocodeVerification::Geocoder.send(:include, LoggingOverride)

gc = Behaviors::GeocodeVerification::Geocoder.new({:enable_geocode_verification => true, :service_phone => "0", :debug => true})

good = []
bad_no_results = []
bad_ambigous = []
bad_wrong = []

# we have 5 cases:
# DUPLICATE => user tried multiple times
# GOOD      => all fine
# WRONG     => only one alternative location was found
# AMBIGOUS  => multiple alternative locations were found
# NO_RESULTS  => no alternatives at all found
all.each do |entry|
  current_digest = Digest::SHA1.hexdigest("#{entry}")
  if hashed.include?(current_digest)
    puts "DUPLICATE : #{entry}"
  else
    err = gc.getErrors(entry)
    if err.empty?
      good << entry
      puts "GOOD : #{entry}"
    else
      # ambigous is considered stronger than wrong
      if ( err.values.join(" ").match(/\.geocode_no_results/) )
        bad_no_results << entry
        puts "NO_RESULTS : #{entry}"
      elsif ( err.values.join(" ").match(/\.geocode_ambigous/) )
        bad_ambigous << entry
        puts "AMBIGOUS : #{entry}"
      else
        bad_wrong << entry
        puts "WRONG : #{entry}"
      end
    end
    hashed << current_digest
  end
end