From 61b2e53ae32fff0536d9f79c1261f55c270f27a8 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 20 Jan 2013 10:16:18 -0500
Subject: [PATCH 01/22] Update README.md

add gem version badge for funsies
---
 README.md | 1 +
 1 file changed, 1 insertion(+)

diff --git a/README.md b/README.md
index 604db04..a6c5b25 100644
--- a/README.md
+++ b/README.md
@@ -5,6 +5,7 @@ and constants handling code) from [Geocommons](http://geocommons.com/)'
 [Geocoder::US 2.0](https://github.com/geocommons/geocoder) gem.
 
 [![Build Status](https://secure.travis-ci.org/daveworth/Indirizzo.png)](http://travis-ci.org/daveworth/Indirizzo)
+[![Gem Version](https://badge.fury.io/rb/Indirizzo.png)](http://badge.fury.io/rb/Indirizzo)
 
 ## Background
 

From 10b06a15a57dda66df7a3a73e513cbf96ea1286c Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:07:17 -0400
Subject: [PATCH 02/22] update rubygems in Gemfile

---
 Gemfile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Gemfile b/Gemfile
index a6ea715..01fb6c2 100644
--- a/Gemfile
+++ b/Gemfile
@@ -1,4 +1,4 @@
-source :rubygems
+source 'https://rubygems.org'
 
 gem 'rake'
 

From 713d984cc9afe734dc81f5013e4c92133f3a00ae Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:07:33 -0400
Subject: [PATCH 03/22] s/cover_me/simplecov/

---
 Gemfile             |  2 +-
 Rakefile            | 12 ------------
 test/test_helper.rb |  3 ++-
 3 files changed, 3 insertions(+), 14 deletions(-)

diff --git a/Gemfile b/Gemfile
index 01fb6c2..a16bb93 100644
--- a/Gemfile
+++ b/Gemfile
@@ -3,7 +3,7 @@ source 'https://rubygems.org'
 gem 'rake'
 
 group :test, :development do
-  gem 'cover_me'
+  gem 'simplecov', require: false
   gem 'awesome_print'
 end
 
diff --git a/Rakefile b/Rakefile
index e134955..925d9de 100644
--- a/Rakefile
+++ b/Rakefile
@@ -17,15 +17,3 @@ Rake::TestTask.new(:test) do |test|
 end
 
 task :default => :test
-
-namespace :cover_me do
-  desc "Generates and opens code coverage report."
-  task :report do
-    require 'cover_me'
-    CoverMe.complete!
-  end
-end
-
-task :test do
-  Rake::Task['cover_me:report'].invoke
-end
diff --git a/test/test_helper.rb b/test/test_helper.rb
index 683618e..8f40974 100644
--- a/test/test_helper.rb
+++ b/test/test_helper.rb
@@ -1,2 +1,3 @@
-require 'cover_me'
+require 'simplecov'
+SimpleCov.start
 require 'test/unit'

From 1d2fa70b824ef904d746e5b4a5134c79d699de3b Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:07:38 -0400
Subject: [PATCH 04/22] add pry

---
 Gemfile | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Gemfile b/Gemfile
index a16bb93..2e70ee3 100644
--- a/Gemfile
+++ b/Gemfile
@@ -5,6 +5,7 @@ gem 'rake'
 group :test, :development do
   gem 'simplecov', require: false
   gem 'awesome_print'
+  gem 'pry'
 end
 
 group :test do

From 1352dafc0b1d045cbc5e99a264c0014b1f3c3df0 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:09:21 -0400
Subject: [PATCH 05/22] extract NumberHelper

---
 lib/indirizzo/address.rb       | 24 ++----------------------
 lib/indirizzo/number_helper.rb | 29 +++++++++++++++++++++++++++++
 test/test_address.rb           |  8 --------
 test/test_number_helper.rb     | 10 ++++++++++
 4 files changed, 41 insertions(+), 30 deletions(-)
 create mode 100644 lib/indirizzo/number_helper.rb
 create mode 100644 test/test_number_helper.rb

diff --git a/lib/indirizzo/address.rb b/lib/indirizzo/address.rb
index 9889f5d..77874a6 100644
--- a/lib/indirizzo/address.rb
+++ b/lib/indirizzo/address.rb
@@ -1,4 +1,5 @@
 require 'indirizzo/constants'
+require 'indirizzo/number_helper'
 
 module Indirizzo
   # Defines the matching of parsed address tokens.
@@ -101,29 +102,8 @@ def assign_text_to_address(text)
       end
     end
 
-    # Expands a token into a list of possible strings based on
-    # the Geocoder::US::Name_Abbr constant, and expands numerals and
-    # number words into their possible equivalents.
     def expand_numbers (string)
-      if /\b\d+(?:st|nd|rd|th)?\b/o.match string
-        match = $&
-        num = $&.to_i
-      elsif Ordinals.regexp.match string
-        num = Ordinals[$&]
-        match = $&
-      elsif Cardinals.regexp.match string
-        num = Cardinals[$&]
-        match = $&
-      end
-      strings = []
-      if num and num < 100
-        [num.to_s, Ordinals[num], Cardinals[num]].each {|replace|
-          strings << string.sub(match, replace)
-        }
-      else
-        strings << string
-      end
-      strings
+      NumberHelper.expand_numbers(string)
     end
 
     def parse_state(regex_match, text)
diff --git a/lib/indirizzo/number_helper.rb b/lib/indirizzo/number_helper.rb
new file mode 100644
index 0000000..dc61886
--- /dev/null
+++ b/lib/indirizzo/number_helper.rb
@@ -0,0 +1,29 @@
+module Indirizzo
+  class NumberHelper
+    # Expands a token into a list of possible strings based on
+    # the Geocoder::US::Name_Abbr constant, and expands numerals and
+    # number words into their possible equivalents.
+    def self.expand_numbers (string)
+      if /\b\d+(?:st|nd|rd|th)?\b/o.match string
+        match = $&
+          num = $&.to_i
+      elsif Ordinals.regexp.match string
+        num = Ordinals[$&]
+        match = $&
+      elsif Cardinals.regexp.match string
+        num = Cardinals[$&]
+        match = $&
+      end
+      strings = []
+      if num and num < 100
+        [num.to_s, Ordinals[num], Cardinals[num]].each {|replace|
+          strings << string.sub(match, replace)
+        }
+      else
+        strings << string
+      end
+      strings
+    end
+  end
+end
+
diff --git a/test/test_address.rb b/test/test_address.rb
index ccef9e3..0b771cd 100644
--- a/test/test_address.rb
+++ b/test/test_address.rb
@@ -16,14 +16,6 @@ def test_new
     assert_equal "1600 Pennsylvania Av, Washington DC", addr.text
   end
 
-  def test_expand_numbers
-    num_list = ["5", "fifth", "five"]
-    num_list.each {|n|
-      addr = Address.new(n)
-      assert_equal num_list, addr.expand_numbers(n).to_a.sort
-    }
-  end
-
   def test_expand_street
     addr = Address.new("1 First St, Atlanta GA, 12345")
     expected_streets = ["1 st", "first st", "one st"]
diff --git a/test/test_number_helper.rb b/test/test_number_helper.rb
new file mode 100644
index 0000000..e2fb268
--- /dev/null
+++ b/test/test_number_helper.rb
@@ -0,0 +1,10 @@
+require 'test_helper'
+include Indirizzo
+class TestNumberHelper < Test::Unit::TestCase
+  def test_expand_numbers
+    num_list = ["5", "fifth", "five"]
+    num_list.each {|n|
+      assert_equal num_list, NumberHelper.expand_numbers(n).to_a.sort
+    }
+  end
+end

From 3636c08e8ab359e712c550a071de8e73b2c75354 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:13:23 -0400
Subject: [PATCH 06/22] extract city class for #city_parts

---
 lib/indirizzo/address.rb | 11 +----------
 lib/indirizzo/city.rb    | 17 +++++++++++++++++
 2 files changed, 18 insertions(+), 10 deletions(-)
 create mode 100644 lib/indirizzo/city.rb

diff --git a/lib/indirizzo/address.rb b/lib/indirizzo/address.rb
index 77874a6..30fb016 100644
--- a/lib/indirizzo/address.rb
+++ b/lib/indirizzo/address.rb
@@ -241,16 +241,7 @@ def remove_noise_words(strings)
     end
 
     def city_parts
-      strings = []
-      @city.map do |string|
-        tokens = string.split(" ")
-        strings |= (0...tokens.length).to_a.reverse.map {|i|
-                   (i...tokens.length).map {|j| tokens[i..j].join(" ")}}.flatten
-      end
-      # Don't return strings that consist solely of abbreviations.
-      # NOTE: Is this a micro-optimization that has edge cases that will break?
-      # Answer: Yes, it breaks on "Prairie"
-      strings.reject { |s| Std_Abbr.key?(s) }.uniq
+      City.city_parts(@city)
     end
 
     def city= (strings)
diff --git a/lib/indirizzo/city.rb b/lib/indirizzo/city.rb
new file mode 100644
index 0000000..a70717f
--- /dev/null
+++ b/lib/indirizzo/city.rb
@@ -0,0 +1,17 @@
+module Indirizzo
+  class City
+    def self.city_parts(city)
+      strings = []
+      city.map do |string|
+        tokens = string.split(" ")
+        strings |= (0...tokens.length).to_a.reverse.map do |i|
+                   (i...tokens.length).map {|j| tokens[i..j].join(" ")}
+        end.flatten
+      end
+      # Don't return strings that consist solely of abbreviations.
+      # NOTE: Is this a micro-optimization that has edge cases that will break?
+      # Answer: Yes, it breaks on "Prairie"
+      strings.reject { |s| Std_Abbr.key?(s) }.uniq
+    end
+  end
+end

From 8061535fa401df4cc21287d1ec84783dacfee17f Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:15:50 -0400
Subject: [PATCH 07/22] extract Match regexps

---
 lib/indirizzo/address.rb | 13 +------------
 lib/indirizzo/match.rb   | 13 +++++++++++++
 2 files changed, 14 insertions(+), 12 deletions(-)
 create mode 100644 lib/indirizzo/match.rb

diff --git a/lib/indirizzo/address.rb b/lib/indirizzo/address.rb
index 30fb016..4c4cb73 100644
--- a/lib/indirizzo/address.rb
+++ b/lib/indirizzo/address.rb
@@ -1,19 +1,8 @@
 require 'indirizzo/constants'
 require 'indirizzo/number_helper'
+require 'indirizzo/match'
 
 module Indirizzo
-  # Defines the matching of parsed address tokens.
-  Match = {
-    # FIXME: shouldn't have to anchor :number and :zip at start/end
-    :number   => /^(\d+\W|[a-z]+)?(\d+)([a-z]?)\b/io,
-    :street   => /(?:\b(?:\d+\w*|[a-z'-]+)\s*)+/io,
-    :city     => /(?:\b[a-z][a-z'-]+\s*)+/io,
-    :state    => State.regexp,
-    :zip      => /\b(\d{5})(?:-(\d{4}))?\b/o,
-    :at       => /\s(at|@|and|&)\s/io,
-    :po_box => /\b[P|p]*(OST|ost)*\.*\s*[O|o|0]*(ffice|FFICE)*\.*\s*[B|b][O|o|0][X|x]\b/
-  }
-
   # The Address class takes a US street address or place name and
   # constructs a list of possible structured parses of the address
   # string.
diff --git a/lib/indirizzo/match.rb b/lib/indirizzo/match.rb
new file mode 100644
index 0000000..fc11c2b
--- /dev/null
+++ b/lib/indirizzo/match.rb
@@ -0,0 +1,13 @@
+module Indirizzo
+  # Defines the matching of parsed address tokens.
+  Match = {
+    # FIXME: shouldn't have to anchor :number and :zip at start/end
+    :number   => /^(\d+\W|[a-z]+)?(\d+)([a-z]?)\b/io,
+    :street   => /(?:\b(?:\d+\w*|[a-z'-]+)\s*)+/io,
+    :city     => /(?:\b[a-z][a-z'-]+\s*)+/io,
+    :state    => State.regexp,
+    :zip      => /\b(\d{5})(?:-(\d{4}))?\b/o,
+    :at       => /\s(at|@|and|&)\s/io,
+    :po_box => /\b[P|p]*(OST|ost)*\.*\s*[O|o|0]*(ffice|FFICE)*\.*\s*[B|b][O|o|0][X|x]\b/
+  }
+end

From 6f61158b719de0bee46ce4175b14b810ec5b2355 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:23:47 -0400
Subject: [PATCH 08/22] extract Street class and associated methods

---
 lib/indirizzo/address.rb | 50 ++++------------------------------------
 lib/indirizzo/helper.rb  | 24 +++++++++++++++++++
 lib/indirizzo/street.rb  | 40 ++++++++++++++++++++++++++++++++
 3 files changed, 69 insertions(+), 45 deletions(-)
 create mode 100644 lib/indirizzo/helper.rb
 create mode 100644 lib/indirizzo/street.rb

diff --git a/lib/indirizzo/address.rb b/lib/indirizzo/address.rb
index 4c4cb73..49f73c4 100644
--- a/lib/indirizzo/address.rb
+++ b/lib/indirizzo/address.rb
@@ -1,6 +1,7 @@
 require 'indirizzo/constants'
-require 'indirizzo/number_helper'
 require 'indirizzo/match'
+require 'indirizzo/city'
+require 'indirizzo/street'
 
 module Indirizzo
   # The Address class takes a US street address or place name and
@@ -177,56 +178,15 @@ def parse
     end
 
     def expand_streets(street)
-      if !street.empty? && !street[0].nil?
-        street.map! {|s|s.strip}
-        add = street.map {|item| item.gsub(Name_Abbr.regexp) {|m| Name_Abbr[m]}}
-        street |= add
-        add = street.map {|item| item.gsub(Std_Abbr.regexp) {|m| Std_Abbr[m]}}
-        street |= add
-        street.map! {|item| expand_numbers(item)}
-        street.flatten!
-        street.map! {|s| s.downcase}
-        street.uniq!
-      else
-        street = []
-      end
-      street
+      Street.expand(street)
     end
 
     def street_parts
-      strings = []
-      # Get all the substrings delimited by whitespace
-      @street.each {|string|
-        tokens = string.split(" ")
-        strings |= (0...tokens.length).map {|i|
-                   (i...tokens.length).map {|j| tokens[i..j].join(" ")}}.flatten
-      }
-      strings = remove_noise_words(strings)
-
-      # Try a simpler case of adding the @number in case everything is an abbr.
-      strings += [@number] if strings.all? {|s| Std_Abbr.key? s or Name_Abbr.key? s}
-      strings.uniq
+      Street.parts(@street, @number)
     end
 
     def remove_noise_words(strings)
-      # Don't return strings that consist solely of abbreviations.
-      # NOTE: Is this a micro-optimization that has edge cases that will break?
-      # Answer: Yes, it breaks on simple things like "Prairie St" or "Front St"
-      prefix = Regexp.new("^" + Prefix_Type.regexp.source + "\s*", Regexp::IGNORECASE)
-      suffix = Regexp.new("\s*" + Suffix_Type.regexp.source + "$", Regexp::IGNORECASE)
-      predxn = Regexp.new("^" + Directional.regexp.source + "\s*", Regexp::IGNORECASE)
-      sufdxn = Regexp.new("\s*" + Directional.regexp.source + "$", Regexp::IGNORECASE)
-      good_strings = strings.map {|s|
-        s = s.clone
-        s.gsub!(predxn, "")
-        s.gsub!(sufdxn, "")
-        s.gsub!(prefix, "")
-        s.gsub!(suffix, "")
-        s
-      }
-      good_strings.reject! {|s| s.empty?}
-      strings = good_strings if !good_strings.empty? {|s| not Std_Abbr.key?(s) and not Name_Abbr.key?(s)}
-      strings
+      Helper.remove_noise_words(strings)
     end
 
     def city_parts
diff --git a/lib/indirizzo/helper.rb b/lib/indirizzo/helper.rb
new file mode 100644
index 0000000..a5b276d
--- /dev/null
+++ b/lib/indirizzo/helper.rb
@@ -0,0 +1,24 @@
+module Indirizzo
+  class Helper
+    def self.remove_noise_words(strings)
+      # Don't return strings that consist solely of abbreviations.
+      # NOTE: Is this a micro-optimization that has edge cases that will break?
+      # Answer: Yes, it breaks on simple things like "Prairie St" or "Front St"
+      prefix = Regexp.new("^" + Prefix_Type.regexp.source + "\s*", Regexp::IGNORECASE)
+      suffix = Regexp.new("\s*" + Suffix_Type.regexp.source + "$", Regexp::IGNORECASE)
+      predxn = Regexp.new("^" + Directional.regexp.source + "\s*", Regexp::IGNORECASE)
+      sufdxn = Regexp.new("\s*" + Directional.regexp.source + "$", Regexp::IGNORECASE)
+      good_strings = strings.map {|s|
+        s = s.clone
+        s.gsub!(predxn, "")
+        s.gsub!(sufdxn, "")
+        s.gsub!(prefix, "")
+        s.gsub!(suffix, "")
+        s
+      }
+      good_strings.reject! {|s| s.empty?}
+      strings = good_strings if !good_strings.empty? {|s| not Std_Abbr.key?(s) and not Name_Abbr.key?(s)}
+      strings
+    end
+  end
+end
diff --git a/lib/indirizzo/street.rb b/lib/indirizzo/street.rb
new file mode 100644
index 0000000..c1b98e9
--- /dev/null
+++ b/lib/indirizzo/street.rb
@@ -0,0 +1,40 @@
+require 'indirizzo/constants'
+require 'indirizzo/number_helper'
+require 'indirizzo/helper'
+
+module Indirizzo
+  class Street
+    def self.expand(street)
+      if !street.empty? && !street[0].nil?
+        street.map! {|s|s.strip}
+        add = street.map {|item| item.gsub(Name_Abbr.regexp) {|m| Name_Abbr[m]}}
+        street |= add
+        add = street.map {|item| item.gsub(Std_Abbr.regexp) {|m| Std_Abbr[m]}}
+        street |= add
+        street.map! {|item| NumberHelper.expand_numbers(item)}
+        street.flatten!
+        street.map! {|s| s.downcase}
+        street.uniq!
+      else
+        street = []
+      end
+      street
+    end
+
+    def self.parts(street, number)
+      strings = []
+      # Get all the substrings delimited by whitespace
+      street.each do |string|
+        tokens = string.split(" ")
+        strings |= (0...tokens.length).map do |i|
+                   (i...tokens.length).map {|j| tokens[i..j].join(" ")}
+        end.flatten
+      end
+      strings = Helper.remove_noise_words(strings)
+
+      # Try a simpler case of adding the @number in case everything is an abbr.
+      strings += [number] if strings.all? {|s| Std_Abbr.key?(s) || Name_Abbr.key?(s)}
+      strings.uniq
+    end
+  end
+end

From 6899077187d386d5edb15755f6e39461cb0ca1e3 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:46:43 -0400
Subject: [PATCH 09/22] extract crazy parser into a class

---
 lib/indirizzo/address.rb | 80 +--------------------------------
 lib/indirizzo/parser.rb  | 96 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 98 insertions(+), 78 deletions(-)
 create mode 100644 lib/indirizzo/parser.rb

diff --git a/lib/indirizzo/address.rb b/lib/indirizzo/address.rb
index 49f73c4..a5092a0 100644
--- a/lib/indirizzo/address.rb
+++ b/lib/indirizzo/address.rb
@@ -1,4 +1,5 @@
 require 'indirizzo/constants'
+require 'indirizzo/parser'
 require 'indirizzo/match'
 require 'indirizzo/city'
 require 'indirizzo/street'
@@ -96,85 +97,8 @@ def expand_numbers (string)
       NumberHelper.expand_numbers(string)
     end
 
-    def parse_state(regex_match, text)
-      idx = text.rindex(regex_match)
-      @full_state = @state[0].strip # special case: New York
-      @state = State[@full_state]
-      @city = "Washington" if @state == "DC" && text[idx...idx+regex_match.length] =~ /washington\s+d\.?c\.?/i
-      text
-    end
-
     def parse
-      text = @text.clone.downcase
-
-      @zip = text.scan(Match[:zip]).last
-      if @zip
-        last_match = $&
-        zip_index = text.rindex(last_match)
-        zip_end_index = zip_index + last_match.length - 1
-        @zip, @plus4 = @zip.map {|s| s and s.strip }
-      else
-        @zip = @plus4 = ""
-        zip_index = text.length
-        zip_end_index = -1
-      end
-
-      @country = @text[zip_end_index+1..-1].sub(/^\s*,\s*/, '').strip
-      @country = nil if @country == text
-
-      @state = text.scan(Match[:state]).last
-      if @state
-        last_match = $&
-        state_index = text.rindex(last_match)
-        text = parse_state(last_match, text)
-      else
-        @full_state = ""
-        @state = ""
-      end
-
-      @number = text.scan(Match[:number]).first
-      # FIXME: 230 Fish And Game Rd, Hudson NY 12534
-      if @number # and not intersection?
-        last_match = $&
-        number_index = text.index(last_match)
-        number_end_index = number_index + last_match.length - 1
-        @prenum, @number, @sufnum = @number.map {|s| s and s.strip}
-      else
-        number_end_index = -1
-        @prenum = @number = @sufnum = ""
-      end
-
-      # FIXME: special case: Name_Abbr gets a bit aggressive
-      # about replacing St with Saint. exceptional case:
-      # Sault Ste. Marie
-
-      # FIXME: PO Box should geocode to ZIP
-      street_search_end_index = [state_index,zip_index,text.length].reject(&:nil?).min-1
-      @street = text[number_end_index+1..street_search_end_index].scan(Match[:street]).map { |s| s and s.strip }
-
-      @street = expand_streets(@street) if @options[:expand_streets]
-      # SPECIAL CASE: 1600 Pennsylvania 20050
-      @street << @full_state if @street.empty? and @state.downcase != @full_state.downcase
-
-      street_end_index = @street.map { |s| text.rindex(s) }.reject(&:nil?).min||0
-
-      if @city.nil? || @city.empty?
-        @city = text[street_end_index..street_search_end_index+1].scan(Match[:city])
-        if !@city.empty?
-          #@city = [@city[-1].strip]
-          @city = [@city.last.strip]
-          add = @city.map {|item| item.gsub(Name_Abbr.regexp) {|m| Name_Abbr[m]}}
-          @city |= add
-          @city.map! {|s| s.downcase}
-          @city.uniq!
-        else
-          @city = []
-        end
-
-        # SPECIAL CASE: no city, but a state with the same name. e.g. "New York"
-        @city << @full_state if @state.downcase != @full_state.downcase
-      end
-
+      @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country = Parser.new(@text, @options).parse
     end
 
     def expand_streets(street)
diff --git a/lib/indirizzo/parser.rb b/lib/indirizzo/parser.rb
new file mode 100644
index 0000000..33e328d
--- /dev/null
+++ b/lib/indirizzo/parser.rb
@@ -0,0 +1,96 @@
+require 'indirizzo/match'
+require 'indirizzo/street'
+require 'indirizzo/constants'
+
+module Indirizzo
+  class Parser
+    def initialize(text, options={})
+      @text = text
+      @options = options
+    end
+
+    def parse
+      text = @text.clone.downcase
+
+      @zip = text.scan(Match[:zip]).last
+      if @zip
+        last_match = $&
+          zip_index = text.rindex(last_match)
+        zip_end_index = zip_index + last_match.length - 1
+        @zip, @plus4 = @zip.map {|s| s and s.strip }
+      else
+        @zip = @plus4 = ""
+        zip_index = text.length
+        zip_end_index = -1
+      end
+
+      @country = @text[zip_end_index+1..-1].sub(/^\s*,\s*/, '').strip
+      @country = nil if @country == text
+
+      @state = text.scan(Match[:state]).last
+      if @state
+        last_match = $&
+          state_index = text.rindex(last_match)
+        text = parse_state(last_match, text)
+      else
+        @full_state = ""
+        @state = ""
+      end
+
+      @number = text.scan(Match[:number]).first
+      # FIXME: 230 Fish And Game Rd, Hudson NY 12534
+      if @number # and not intersection?
+        last_match = $&
+          number_index = text.index(last_match)
+        number_end_index = number_index + last_match.length - 1
+        @prenum, @number, @sufnum = @number.map {|s| s and s.strip}
+      else
+        number_end_index = -1
+        @prenum = @number = @sufnum = ""
+      end
+
+      # FIXME: special case: Name_Abbr gets a bit aggressive
+      # about replacing St with Saint. exceptional case:
+      # Sault Ste. Marie
+
+      # FIXME: PO Box should geocode to ZIP
+      street_search_end_index = [state_index,zip_index,text.length].reject(&:nil?).min-1
+      @street = text[number_end_index+1..street_search_end_index].scan(Match[:street]).map { |s| s and s.strip }
+
+      @street = Street.expand(@street) if @options[:expand_streets]
+      # SPECIAL CASE: 1600 Pennsylvania 20050
+      @street << @full_state if @street.empty? and @state.downcase != @full_state.downcase
+
+      street_end_index = @street.map { |s| text.rindex(s) }.reject(&:nil?).min||0
+
+      if @city.nil? || @city.empty?
+        @city = text[street_end_index..street_search_end_index+1].scan(Match[:city])
+        if !@city.empty?
+          #@city = [@city[-1].strip]
+          @city = [@city.last.strip]
+          add = @city.map {|item| item.gsub(Name_Abbr.regexp) {|m| Name_Abbr[m]}}
+          @city |= add
+          @city.map! {|s| s.downcase}
+          @city.uniq!
+        else
+          @city = []
+        end
+
+        # SPECIAL CASE: no city, but a state with the same name. e.g. "New York"
+        @city << @full_state if @state.downcase != @full_state.downcase
+      end
+
+      return @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
+    end
+
+    private
+
+    def parse_state(regex_match, text)
+      idx = text.rindex(regex_match)
+      @full_state = @state[0].strip # special case: New York
+      @state = State[@full_state]
+      @city = "Washington" if @state == "DC" && text[idx...idx+regex_match.length] =~ /washington\s+d\.?c\.?/i
+      text
+    end
+  end
+end

From 7a32a6c9967562a5b883b0fbc541818fc6ed9a7b Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:49:14 -0400
Subject: [PATCH 10/22] extract #clean to Helper.clean

---
 lib/indirizzo/address.rb | 5 ++---
 lib/indirizzo/helper.rb  | 6 ++++++
 2 files changed, 8 insertions(+), 3 deletions(-)

diff --git a/lib/indirizzo/address.rb b/lib/indirizzo/address.rb
index a5092a0..704d744 100644
--- a/lib/indirizzo/address.rb
+++ b/lib/indirizzo/address.rb
@@ -3,6 +3,7 @@
 require 'indirizzo/match'
 require 'indirizzo/city'
 require 'indirizzo/street'
+require 'indirizzo/helper'
 
 module Indirizzo
   # The Address class takes a US street address or place name and
@@ -34,9 +35,7 @@ def initialize (text, options={})
 
     # Removes any characters that aren't strictly part of an address string.
     def clean (value)
-      value.strip \
-           .gsub(/[^a-z0-9 ,'&@\/-]+/io, "") \
-           .gsub(/\s+/o, " ")
+      Helper.clean(value)
     end
 
     def assign_text_to_address(text)
diff --git a/lib/indirizzo/helper.rb b/lib/indirizzo/helper.rb
index a5b276d..7c2a447 100644
--- a/lib/indirizzo/helper.rb
+++ b/lib/indirizzo/helper.rb
@@ -20,5 +20,11 @@ def self.remove_noise_words(strings)
       strings = good_strings if !good_strings.empty? {|s| not Std_Abbr.key?(s) and not Name_Abbr.key?(s)}
       strings
     end
+
+    def self.clean(value)
+      value.strip \
+           .gsub(/[^a-z0-9 ,'&@\/-]+/io, "") \
+           .gsub(/\s+/o, " ")
+    end
   end
 end

From 05f8e903d9b49ed29e69e69b0cd2488b3b028a2f Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sat, 4 May 2013 23:56:39 -0400
Subject: [PATCH 11/22] extract #assign_text_to_address to Parser

---
 lib/indirizzo/address.rb | 52 +----------------------------------
 lib/indirizzo/parser.rb  | 58 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 59 insertions(+), 51 deletions(-)

diff --git a/lib/indirizzo/address.rb b/lib/indirizzo/address.rb
index 704d744..2fe5f83 100644
--- a/lib/indirizzo/address.rb
+++ b/lib/indirizzo/address.rb
@@ -39,57 +39,7 @@ def clean (value)
     end
 
     def assign_text_to_address(text)
-      if !text[:address].nil?
-        @text = clean text[:address]
-        parse
-      else
-        @street = []
-        @prenum = text[:prenum]
-        @sufnum = text[:sufnum]
-        if !text[:street].nil?
-          @street = text[:street].scan(Match[:street])
-        end
-        @number = ""
-        if !@street.nil?
-          if text[:number].nil?
-            @street.map! { |single_street|
-              single_street.downcase!
-              @number = single_street.scan(Match[:number])[0].reject{|n| n.nil? || n.empty?}.first.to_s
-              single_street.sub! @number, ""
-              single_street.sub! /^\s*,?\s*/o, ""
-            }
-          else
-            @number = text[:number].to_s
-          end
-          @street = expand_streets(@street) if @options[:expand_streets]
-          street_parts
-        end
-        @city = []
-        if !text[:city].nil?
-          @city.push(text[:city])
-          @text = text[:city].to_s
-        else
-          @city.push("")
-        end
-        if !text[:region].nil?
-          # @state = []
-          @state = text[:region]
-          if @state.length > 2
-            # full_state = @state.strip # special case: New York
-            @state = State[@state]
-          end
-        elsif !text[:state].nil?
-          @state = text[:state]
-        elsif !text[:country].nil?
-          @state = text[:country]
-        end
-
-        @zip = text[:postal_code]
-        @plus4 = text[:plus4]
-        if !@zip
-          @zip = @plus4 = ""
-        end
-      end
+      @text, @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country = Parser.extract_data_from_hash(text, @options)
     end
 
     def expand_numbers (string)
diff --git a/lib/indirizzo/parser.rb b/lib/indirizzo/parser.rb
index 33e328d..97970ef 100644
--- a/lib/indirizzo/parser.rb
+++ b/lib/indirizzo/parser.rb
@@ -1,4 +1,5 @@
 require 'indirizzo/match'
+require 'indirizzo/helper'
 require 'indirizzo/street'
 require 'indirizzo/constants'
 
@@ -83,6 +84,63 @@ def parse
       return @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
     end
 
+    def self.extract_data_from_hash(address_hash, options)
+      text = address_hash
+      if !text[:address].nil?
+        @text = Helper.clean text[:address]
+        return self.new(@text, options).parse
+      else
+        @street = []
+        @prenum = text[:prenum]
+        @sufnum = text[:sufnum]
+        if !text[:street].nil?
+          @street = text[:street].scan(Match[:street])
+        end
+        @number = ""
+        if !@street.nil?
+          if text[:number].nil?
+            @street.map! { |single_street|
+              single_street.downcase!
+              @number = single_street.scan(Match[:number])[0].reject{|n| n.nil? || n.empty?}.first.to_s
+              single_street.sub! @number, ""
+              single_street.sub! /^\s*,?\s*/o, ""
+            }
+          else
+            @number = text[:number].to_s
+          end
+          @street = Street.expand(@street) if options[:expand_streets]
+          #Street.parts
+        end
+        @city = []
+        if !text[:city].nil?
+          @city.push(text[:city])
+          @text = text[:city].to_s
+        else
+          @city.push("")
+        end
+        if !text[:region].nil?
+          # @state = []
+          @state = text[:region]
+          if @state.length > 2
+            # full_state = @state.strip # special case: New York
+            @state = State[@state]
+          end
+        elsif !text[:state].nil?
+          @state = text[:state]
+        elsif !text[:country].nil?
+          @state = text[:country]
+        end
+
+        @zip = text[:postal_code]
+        @plus4 = text[:plus4]
+        if !@zip
+          @zip = @plus4 = ""
+        end
+      end
+
+      return @text, @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
+    end
+
     private
 
     def parse_state(regex_match, text)

From 69782f6b1a25290c24ca83443893ef99d9cf4a13 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:01:10 -0400
Subject: [PATCH 12/22] extract terrible method to its own class instead

---
 lib/indirizzo/address.rb                |  3 +-
 lib/indirizzo/address_hash_extractor.rb | 61 +++++++++++++++++++++++++
 lib/indirizzo/parser.rb                 | 57 -----------------------
 3 files changed, 63 insertions(+), 58 deletions(-)
 create mode 100644 lib/indirizzo/address_hash_extractor.rb

diff --git a/lib/indirizzo/address.rb b/lib/indirizzo/address.rb
index 2fe5f83..89d7996 100644
--- a/lib/indirizzo/address.rb
+++ b/lib/indirizzo/address.rb
@@ -1,5 +1,6 @@
 require 'indirizzo/constants'
 require 'indirizzo/parser'
+require 'indirizzo/address_hash_extractor'
 require 'indirizzo/match'
 require 'indirizzo/city'
 require 'indirizzo/street'
@@ -39,7 +40,7 @@ def clean (value)
     end
 
     def assign_text_to_address(text)
-      @text, @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country = Parser.extract_data_from_hash(text, @options)
+      @text, @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country = AddressHashExtractor.extract(text, @options)
     end
 
     def expand_numbers (string)
diff --git a/lib/indirizzo/address_hash_extractor.rb b/lib/indirizzo/address_hash_extractor.rb
new file mode 100644
index 0000000..3bb3e7a
--- /dev/null
+++ b/lib/indirizzo/address_hash_extractor.rb
@@ -0,0 +1,61 @@
+module Indirizzo
+  class AddressHashExtractor
+    def self.extract(address_hash, options)
+      text = address_hash
+      if !text[:address].nil?
+        @text = Helper.clean text[:address]
+        return Parser.new(@text, options).parse
+      else
+        @street = []
+        @prenum = text[:prenum]
+        @sufnum = text[:sufnum]
+        if !text[:street].nil?
+          @street = text[:street].scan(Match[:street])
+        end
+        @number = ""
+        if !@street.nil?
+          if text[:number].nil?
+            @street.map! { |single_street|
+              single_street.downcase!
+              @number = single_street.scan(Match[:number])[0].reject{|n| n.nil? || n.empty?}.first.to_s
+              single_street.sub! @number, ""
+              single_street.sub! /^\s*,?\s*/o, ""
+            }
+          else
+            @number = text[:number].to_s
+          end
+          @street = Street.expand(@street) if options[:expand_streets]
+          #Street.parts
+        end
+        @city = []
+        if !text[:city].nil?
+          @city.push(text[:city])
+          @text = text[:city].to_s
+        else
+          @city.push("")
+        end
+        if !text[:region].nil?
+          # @state = []
+          @state = text[:region]
+          if @state.length > 2
+            # full_state = @state.strip # special case: New York
+            @state = State[@state]
+          end
+        elsif !text[:state].nil?
+          @state = text[:state]
+        elsif !text[:country].nil?
+          @state = text[:country]
+        end
+
+        @zip = text[:postal_code]
+        @plus4 = text[:plus4]
+        if !@zip
+          @zip = @plus4 = ""
+        end
+      end
+
+      return @text, @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
+    end
+
+  end
+end
diff --git a/lib/indirizzo/parser.rb b/lib/indirizzo/parser.rb
index 97970ef..74656ee 100644
--- a/lib/indirizzo/parser.rb
+++ b/lib/indirizzo/parser.rb
@@ -84,63 +84,6 @@ def parse
       return @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
     end
 
-    def self.extract_data_from_hash(address_hash, options)
-      text = address_hash
-      if !text[:address].nil?
-        @text = Helper.clean text[:address]
-        return self.new(@text, options).parse
-      else
-        @street = []
-        @prenum = text[:prenum]
-        @sufnum = text[:sufnum]
-        if !text[:street].nil?
-          @street = text[:street].scan(Match[:street])
-        end
-        @number = ""
-        if !@street.nil?
-          if text[:number].nil?
-            @street.map! { |single_street|
-              single_street.downcase!
-              @number = single_street.scan(Match[:number])[0].reject{|n| n.nil? || n.empty?}.first.to_s
-              single_street.sub! @number, ""
-              single_street.sub! /^\s*,?\s*/o, ""
-            }
-          else
-            @number = text[:number].to_s
-          end
-          @street = Street.expand(@street) if options[:expand_streets]
-          #Street.parts
-        end
-        @city = []
-        if !text[:city].nil?
-          @city.push(text[:city])
-          @text = text[:city].to_s
-        else
-          @city.push("")
-        end
-        if !text[:region].nil?
-          # @state = []
-          @state = text[:region]
-          if @state.length > 2
-            # full_state = @state.strip # special case: New York
-            @state = State[@state]
-          end
-        elsif !text[:state].nil?
-          @state = text[:state]
-        elsif !text[:country].nil?
-          @state = text[:country]
-        end
-
-        @zip = text[:postal_code]
-        @plus4 = text[:plus4]
-        if !@zip
-          @zip = @plus4 = ""
-        end
-      end
-
-      return @text, @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
-    end
-
     private
 
     def parse_state(regex_match, text)

From 969a88428ece201e1b344b095ffd261f8fc2c56a Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:06:32 -0400
Subject: [PATCH 13/22] extract zip-parsing into method

* this is hinky thanks to the weird indices but at least this is a start
---
 lib/indirizzo/parser.rb | 27 ++++++++++++++++-----------
 1 file changed, 16 insertions(+), 11 deletions(-)

diff --git a/lib/indirizzo/parser.rb b/lib/indirizzo/parser.rb
index 74656ee..c2c1ea4 100644
--- a/lib/indirizzo/parser.rb
+++ b/lib/indirizzo/parser.rb
@@ -13,17 +13,7 @@ def initialize(text, options={})
     def parse
       text = @text.clone.downcase
 
-      @zip = text.scan(Match[:zip]).last
-      if @zip
-        last_match = $&
-          zip_index = text.rindex(last_match)
-        zip_end_index = zip_index + last_match.length - 1
-        @zip, @plus4 = @zip.map {|s| s and s.strip }
-      else
-        @zip = @plus4 = ""
-        zip_index = text.length
-        zip_end_index = -1
-      end
+      @zip, @plus4, zip_index, zip_end_index = extract_zip_from_text(text)
 
       @country = @text[zip_end_index+1..-1].sub(/^\s*,\s*/, '').strip
       @country = nil if @country == text
@@ -93,5 +83,20 @@ def parse_state(regex_match, text)
       @city = "Washington" if @state == "DC" && text[idx...idx+regex_match.length] =~ /washington\s+d\.?c\.?/i
       text
     end
+
+    def extract_zip_from_text(text)
+      zip = text.scan(Match[:zip]).last
+      if zip
+        last_match = $&
+          zip_index = text.rindex(last_match)
+        zip_end_index = zip_index + last_match.length - 1
+        zip, plus4 = zip.map {|s| s and s.strip }
+      else
+        zip = plus4 = ""
+        zip_index = text.length
+        zip_end_index = -1
+      end
+      return zip, plus4, zip_index, zip_end_index
+    end
   end
 end

From c2c7a59c6cf67fd2a73a345a5e3544b368f037c0 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:12:58 -0400
Subject: [PATCH 14/22] extract state fetching into method

* inline parse_state to avoid side-effects that were weird
---
 lib/indirizzo/parser.rb | 34 +++++++++++++++++-----------------
 1 file changed, 17 insertions(+), 17 deletions(-)

diff --git a/lib/indirizzo/parser.rb b/lib/indirizzo/parser.rb
index c2c1ea4..0ea7e9d 100644
--- a/lib/indirizzo/parser.rb
+++ b/lib/indirizzo/parser.rb
@@ -18,15 +18,7 @@ def parse
       @country = @text[zip_end_index+1..-1].sub(/^\s*,\s*/, '').strip
       @country = nil if @country == text
 
-      @state = text.scan(Match[:state]).last
-      if @state
-        last_match = $&
-          state_index = text.rindex(last_match)
-        text = parse_state(last_match, text)
-      else
-        @full_state = ""
-        @state = ""
-      end
+      @state, @full_state, @city, state_index = extract_state_from_text(text)
 
       @number = text.scan(Match[:number]).first
       # FIXME: 230 Fish And Game Rd, Hudson NY 12534
@@ -76,14 +68,6 @@ def parse
 
     private
 
-    def parse_state(regex_match, text)
-      idx = text.rindex(regex_match)
-      @full_state = @state[0].strip # special case: New York
-      @state = State[@full_state]
-      @city = "Washington" if @state == "DC" && text[idx...idx+regex_match.length] =~ /washington\s+d\.?c\.?/i
-      text
-    end
-
     def extract_zip_from_text(text)
       zip = text.scan(Match[:zip]).last
       if zip
@@ -98,5 +82,21 @@ def extract_zip_from_text(text)
       end
       return zip, plus4, zip_index, zip_end_index
     end
+
+    def extract_state_from_text(text)
+      state = text.scan(Match[:state]).last
+      if state
+        last_match = $&
+          state_index = text.rindex(last_match)
+        idx = text.rindex(last_match)
+        full_state = state[0].strip # special case: New York
+        state = State[full_state]
+        city = "Washington" if state == "DC" && text[idx...idx+last_match.length] =~ /washington\s+d\.?c\.?/i
+      else
+        full_state = ""
+        state = ""
+      end
+      return state, full_state, city, state_index
+    end
   end
 end

From 2ff67e282fe39de11d3f03eabd68aaf0ff1ff17e Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:17:23 -0400
Subject: [PATCH 15/22] extract city processing into method

---
 lib/indirizzo/parser.rb | 36 ++++++++++++++++++++----------------
 1 file changed, 20 insertions(+), 16 deletions(-)

diff --git a/lib/indirizzo/parser.rb b/lib/indirizzo/parser.rb
index 0ea7e9d..328c03a 100644
--- a/lib/indirizzo/parser.rb
+++ b/lib/indirizzo/parser.rb
@@ -46,22 +46,7 @@ def parse
 
       street_end_index = @street.map { |s| text.rindex(s) }.reject(&:nil?).min||0
 
-      if @city.nil? || @city.empty?
-        @city = text[street_end_index..street_search_end_index+1].scan(Match[:city])
-        if !@city.empty?
-          #@city = [@city[-1].strip]
-          @city = [@city.last.strip]
-          add = @city.map {|item| item.gsub(Name_Abbr.regexp) {|m| Name_Abbr[m]}}
-          @city |= add
-          @city.map! {|s| s.downcase}
-          @city.uniq!
-        else
-          @city = []
-        end
-
-        # SPECIAL CASE: no city, but a state with the same name. e.g. "New York"
-        @city << @full_state if @state.downcase != @full_state.downcase
-      end
+      process_city(text, street_end_index, street_search_end_index)
 
       return @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
     end
@@ -98,5 +83,24 @@ def extract_state_from_text(text)
       end
       return state, full_state, city, state_index
     end
+
+    def process_city(text, street_end_index, street_search_end_index)
+      if @city.nil? || @city.empty?
+        @city = text[street_end_index..street_search_end_index+1].scan(Match[:city])
+        if !@city.empty?
+          #@city = [@city[-1].strip]
+          @city = [@city.last.strip]
+          add = @city.map {|item| item.gsub(Name_Abbr.regexp) {|m| Name_Abbr[m]}}
+          @city |= add
+          @city.map! {|s| s.downcase}
+          @city.uniq!
+        else
+          @city = []
+        end
+
+        # SPECIAL CASE: no city, but a state with the same name. e.g. "New York"
+        @city << @full_state if @state.downcase != @full_state.downcase
+      end
+    end
   end
 end

From d1f77ddc5048e9d0b863ad2af2b1f17976802bb2 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:21:54 -0400
Subject: [PATCH 16/22] extract number processing into method

---
 lib/indirizzo/parser.rb | 27 ++++++++++++++++-----------
 1 file changed, 16 insertions(+), 11 deletions(-)

diff --git a/lib/indirizzo/parser.rb b/lib/indirizzo/parser.rb
index 328c03a..3430184 100644
--- a/lib/indirizzo/parser.rb
+++ b/lib/indirizzo/parser.rb
@@ -20,17 +20,7 @@ def parse
 
       @state, @full_state, @city, state_index = extract_state_from_text(text)
 
-      @number = text.scan(Match[:number]).first
-      # FIXME: 230 Fish And Game Rd, Hudson NY 12534
-      if @number # and not intersection?
-        last_match = $&
-          number_index = text.index(last_match)
-        number_end_index = number_index + last_match.length - 1
-        @prenum, @number, @sufnum = @number.map {|s| s and s.strip}
-      else
-        number_end_index = -1
-        @prenum = @number = @sufnum = ""
-      end
+      @prenum, @number, @sufnum, number_end_index = process_number(text)
 
       # FIXME: special case: Name_Abbr gets a bit aggressive
       # about replacing St with Saint. exceptional case:
@@ -102,5 +92,20 @@ def process_city(text, street_end_index, street_search_end_index)
         @city << @full_state if @state.downcase != @full_state.downcase
       end
     end
+
+    def process_number(text)
+      number = text.scan(Match[:number]).first
+      # FIXME: 230 Fish And Game Rd, Hudson NY 12534
+      if number # and not intersection?
+        last_match = $&
+          number_index = text.index(last_match)
+        number_end_index = number_index + last_match.length - 1
+        prenum, number, sufnum = number.map {|s| s and s.strip}
+      else
+        number_end_index = -1
+        prenum = number = sufnum = ""
+      end
+      return prenum, number, sufnum, number_end_index
+    end
   end
 end

From 4ce9b1f96c51c303788bdc3c103f77b423605765 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:28:18 -0400
Subject: [PATCH 17/22] basic extraction refactor driven by tests

---
 lib/indirizzo/address_hash_extractor.rb | 111 +++++++++++++-----------
 1 file changed, 62 insertions(+), 49 deletions(-)

diff --git a/lib/indirizzo/address_hash_extractor.rb b/lib/indirizzo/address_hash_extractor.rb
index 3bb3e7a..c571cc9 100644
--- a/lib/indirizzo/address_hash_extractor.rb
+++ b/lib/indirizzo/address_hash_extractor.rb
@@ -1,61 +1,74 @@
 module Indirizzo
   class AddressHashExtractor
     def self.extract(address_hash, options)
-      text = address_hash
-      if !text[:address].nil?
-        @text = Helper.clean text[:address]
-        return Parser.new(@text, options).parse
+      AddressHashExtractor.new(address_hash, options).extract
+    end
+
+    def initialize(address_hash, options={})
+      @address_hash = address_hash
+      @options = options
+    end
+
+    def extract
+      if !@address_hash[:address].nil?
+        @text = Helper.clean @address_hash[:address]
+        return Parser.new(@text, @options).parse
       else
-        @street = []
-        @prenum = text[:prenum]
-        @sufnum = text[:sufnum]
-        if !text[:street].nil?
-          @street = text[:street].scan(Match[:street])
-        end
-        @number = ""
-        if !@street.nil?
-          if text[:number].nil?
-            @street.map! { |single_street|
-              single_street.downcase!
-              @number = single_street.scan(Match[:number])[0].reject{|n| n.nil? || n.empty?}.first.to_s
-              single_street.sub! @number, ""
-              single_street.sub! /^\s*,?\s*/o, ""
-            }
-          else
-            @number = text[:number].to_s
-          end
-          @street = Street.expand(@street) if options[:expand_streets]
-          #Street.parts
-        end
-        @city = []
-        if !text[:city].nil?
-          @city.push(text[:city])
-          @text = text[:city].to_s
+        handle_hash
+      end
+
+      return @text, @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
+    end
+
+    private
+    def handle_hash
+      text = @address_hash
+      @street = []
+      @prenum = text[:prenum]
+      @sufnum = text[:sufnum]
+      if !text[:street].nil?
+        @street = text[:street].scan(Match[:street])
+      end
+      @number = ""
+      if !@street.nil?
+        if text[:number].nil?
+          @street.map! { |single_street|
+            single_street.downcase!
+            @number = single_street.scan(Match[:number])[0].reject{|n| n.nil? || n.empty?}.first.to_s
+            single_street.sub! @number, ""
+            single_street.sub! /^\s*,?\s*/o, ""
+          }
         else
-          @city.push("")
-        end
-        if !text[:region].nil?
-          # @state = []
-          @state = text[:region]
-          if @state.length > 2
-            # full_state = @state.strip # special case: New York
-            @state = State[@state]
-          end
-        elsif !text[:state].nil?
-          @state = text[:state]
-        elsif !text[:country].nil?
-          @state = text[:country]
+          @number = text[:number].to_s
         end
-
-        @zip = text[:postal_code]
-        @plus4 = text[:plus4]
-        if !@zip
-          @zip = @plus4 = ""
+        @street = Street.expand(@street) if @options[:expand_streets]
+        #Street.parts
+      end
+      @city = []
+      if !text[:city].nil?
+        @city.push(text[:city])
+        @text = text[:city].to_s
+      else
+        @city.push("")
+      end
+      if !text[:region].nil?
+        # @state = []
+        @state = text[:region]
+        if @state.length > 2
+          # full_state = @state.strip # special case: New York
+          @state = State[@state]
         end
+      elsif !text[:state].nil?
+        @state = text[:state]
+      elsif !text[:country].nil?
+        @state = text[:country]
       end
 
-      return @text, @city, @street, @number, @prenum, @sufnum, @full_state, @state, @zip, @plus4, @country
+      @zip = text[:postal_code]
+      @plus4 = text[:plus4]
+      if !@zip
+        @zip = @plus4 = ""
+      end
     end
-
   end
 end

From 43fd2fdb8b33f16a88d97b2eb6adcf32808d728a Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:30:37 -0400
Subject: [PATCH 18/22] extract city extraction code into method

---
 lib/indirizzo/address_hash_extractor.rb | 20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

diff --git a/lib/indirizzo/address_hash_extractor.rb b/lib/indirizzo/address_hash_extractor.rb
index c571cc9..be863d0 100644
--- a/lib/indirizzo/address_hash_extractor.rb
+++ b/lib/indirizzo/address_hash_extractor.rb
@@ -44,13 +44,9 @@ def handle_hash
         @street = Street.expand(@street) if @options[:expand_streets]
         #Street.parts
       end
-      @city = []
-      if !text[:city].nil?
-        @city.push(text[:city])
-        @text = text[:city].to_s
-      else
-        @city.push("")
-      end
+
+      handle_city
+
       if !text[:region].nil?
         # @state = []
         @state = text[:region]
@@ -70,5 +66,15 @@ def handle_hash
         @zip = @plus4 = ""
       end
     end
+
+    def handle_city
+      @city = []
+      if !@address_hash[:city].nil?
+        @city.push(@address_hash[:city])
+        @text = @address_hash[:city].to_s
+      else
+        @city.push("")
+      end
+    end
   end
 end

From 2c8180b5a14c0acfc4aec2afb2c2a0ffecf1ed17 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:32:21 -0400
Subject: [PATCH 19/22] extract state extraction into method

---
 lib/indirizzo/address_hash_extractor.rb | 29 ++++++++++++++-----------
 1 file changed, 16 insertions(+), 13 deletions(-)

diff --git a/lib/indirizzo/address_hash_extractor.rb b/lib/indirizzo/address_hash_extractor.rb
index be863d0..c7dcfba 100644
--- a/lib/indirizzo/address_hash_extractor.rb
+++ b/lib/indirizzo/address_hash_extractor.rb
@@ -46,19 +46,7 @@ def handle_hash
       end
 
       handle_city
-
-      if !text[:region].nil?
-        # @state = []
-        @state = text[:region]
-        if @state.length > 2
-          # full_state = @state.strip # special case: New York
-          @state = State[@state]
-        end
-      elsif !text[:state].nil?
-        @state = text[:state]
-      elsif !text[:country].nil?
-        @state = text[:country]
-      end
+      handle_state
 
       @zip = text[:postal_code]
       @plus4 = text[:plus4]
@@ -76,5 +64,20 @@ def handle_city
         @city.push("")
       end
     end
+
+    def handle_state
+      if !@address_hash[:region].nil?
+        # @state = []
+        @state = @address_hash[:region]
+        if @state.length > 2
+          # full_state = @state.strip # special case: New York
+          @state = State[@state]
+        end
+      elsif !@address_hash[:state].nil?
+        @state = @address_hash[:state]
+      elsif !@address_hash[:country].nil?
+        @state = @address_hash[:country]
+      end
+    end
   end
 end

From 8c3aa96534030ef28326ee393c01f9540875ec92 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:33:34 -0400
Subject: [PATCH 20/22] extract state and number extraction into method

---
 lib/indirizzo/address_hash_extractor.rb | 38 ++++++++++++++-----------
 1 file changed, 21 insertions(+), 17 deletions(-)

diff --git a/lib/indirizzo/address_hash_extractor.rb b/lib/indirizzo/address_hash_extractor.rb
index c7dcfba..ae937e7 100644
--- a/lib/indirizzo/address_hash_extractor.rb
+++ b/lib/indirizzo/address_hash_extractor.rb
@@ -22,16 +22,22 @@ def extract
 
     private
     def handle_hash
-      text = @address_hash
+      handle_street_and_numbers
+      handle_city
+      handle_state
+      handle_zip
+    end
+
+    def handle_street_and_numbers
       @street = []
-      @prenum = text[:prenum]
-      @sufnum = text[:sufnum]
-      if !text[:street].nil?
-        @street = text[:street].scan(Match[:street])
+      @prenum = @address_hash[:prenum]
+      @sufnum = @address_hash[:sufnum]
+      if !@address_hash[:street].nil?
+        @street = @address_hash[:street].scan(Match[:street])
       end
       @number = ""
       if !@street.nil?
-        if text[:number].nil?
+        if @address_hash[:number].nil?
           @street.map! { |single_street|
             single_street.downcase!
             @number = single_street.scan(Match[:number])[0].reject{|n| n.nil? || n.empty?}.first.to_s
@@ -39,19 +45,9 @@ def handle_hash
             single_street.sub! /^\s*,?\s*/o, ""
           }
         else
-          @number = text[:number].to_s
+          @number = @address_hash[:number].to_s
         end
         @street = Street.expand(@street) if @options[:expand_streets]
-        #Street.parts
-      end
-
-      handle_city
-      handle_state
-
-      @zip = text[:postal_code]
-      @plus4 = text[:plus4]
-      if !@zip
-        @zip = @plus4 = ""
       end
     end
 
@@ -79,5 +75,13 @@ def handle_state
         @state = @address_hash[:country]
       end
     end
+
+    def handle_zip
+      @zip = @address_hash[:postal_code]
+      @plus4 = @address_hash[:plus4]
+      if !@zip
+        @zip = @plus4 = ""
+      end
+    end
   end
 end

From 78d4617f39c552491bbd8d889bf90af729aa03e2 Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:34:21 -0400
Subject: [PATCH 21/22] use attr_accessor correctly

---
 lib/indirizzo/address_hash_extractor.rb | 39 +++++++++++++------------
 1 file changed, 20 insertions(+), 19 deletions(-)

diff --git a/lib/indirizzo/address_hash_extractor.rb b/lib/indirizzo/address_hash_extractor.rb
index ae937e7..b3d0eda 100644
--- a/lib/indirizzo/address_hash_extractor.rb
+++ b/lib/indirizzo/address_hash_extractor.rb
@@ -8,10 +8,11 @@ def initialize(address_hash, options={})
       @address_hash = address_hash
       @options = options
     end
+    attr_accessor :address_hash
 
     def extract
-      if !@address_hash[:address].nil?
-        @text = Helper.clean @address_hash[:address]
+      if !address_hash[:address].nil?
+        @text = Helper.clean address_hash[:address]
         return Parser.new(@text, @options).parse
       else
         handle_hash
@@ -30,14 +31,14 @@ def handle_hash
 
     def handle_street_and_numbers
       @street = []
-      @prenum = @address_hash[:prenum]
-      @sufnum = @address_hash[:sufnum]
-      if !@address_hash[:street].nil?
-        @street = @address_hash[:street].scan(Match[:street])
+      @prenum = address_hash[:prenum]
+      @sufnum = address_hash[:sufnum]
+      if !address_hash[:street].nil?
+        @street = address_hash[:street].scan(Match[:street])
       end
       @number = ""
       if !@street.nil?
-        if @address_hash[:number].nil?
+        if address_hash[:number].nil?
           @street.map! { |single_street|
             single_street.downcase!
             @number = single_street.scan(Match[:number])[0].reject{|n| n.nil? || n.empty?}.first.to_s
@@ -45,7 +46,7 @@ def handle_street_and_numbers
             single_street.sub! /^\s*,?\s*/o, ""
           }
         else
-          @number = @address_hash[:number].to_s
+          @number = address_hash[:number].to_s
         end
         @street = Street.expand(@street) if @options[:expand_streets]
       end
@@ -53,32 +54,32 @@ def handle_street_and_numbers
 
     def handle_city
       @city = []
-      if !@address_hash[:city].nil?
-        @city.push(@address_hash[:city])
-        @text = @address_hash[:city].to_s
+      if !address_hash[:city].nil?
+        @city.push(address_hash[:city])
+        @text = address_hash[:city].to_s
       else
         @city.push("")
       end
     end
 
     def handle_state
-      if !@address_hash[:region].nil?
+      if !address_hash[:region].nil?
         # @state = []
-        @state = @address_hash[:region]
+        @state = address_hash[:region]
         if @state.length > 2
           # full_state = @state.strip # special case: New York
           @state = State[@state]
         end
-      elsif !@address_hash[:state].nil?
-        @state = @address_hash[:state]
-      elsif !@address_hash[:country].nil?
-        @state = @address_hash[:country]
+      elsif !address_hash[:state].nil?
+        @state = address_hash[:state]
+      elsif !address_hash[:country].nil?
+        @state = address_hash[:country]
       end
     end
 
     def handle_zip
-      @zip = @address_hash[:postal_code]
-      @plus4 = @address_hash[:plus4]
+      @zip = address_hash[:postal_code]
+      @plus4 = address_hash[:plus4]
       if !@zip
         @zip = @plus4 = ""
       end

From 2dd0890801d3b9ce7304bca778a044af7fa2198e Mon Sep 17 00:00:00 2001
From: Dave Worth <dave@bignerdranch.com>
Date: Sun, 5 May 2013 00:37:11 -0400
Subject: [PATCH 22/22] tiny cleanup to maybe appease code climate

---
 lib/indirizzo/address_hash_extractor.rb | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/lib/indirizzo/address_hash_extractor.rb b/lib/indirizzo/address_hash_extractor.rb
index b3d0eda..4568fa6 100644
--- a/lib/indirizzo/address_hash_extractor.rb
+++ b/lib/indirizzo/address_hash_extractor.rb
@@ -64,12 +64,9 @@ def handle_city
 
     def handle_state
       if !address_hash[:region].nil?
-        # @state = []
         @state = address_hash[:region]
-        if @state.length > 2
-          # full_state = @state.strip # special case: New York
-          @state = State[@state]
-        end
+        # full_state = @state.strip # special case: New York
+        @state = State[@state] if @state.length > 2
       elsif !address_hash[:state].nil?
         @state = address_hash[:state]
       elsif !address_hash[:country].nil?