Indirizzo 0.1.0 → 0.1.1

data/Gemfile

@@ -1,3 +1,12 @@
 source :rubygems
 
-gemspec
+gem 'rake'
+
+group :test, :development do
+  gem 'cover_me'
+  gem 'awesome_print'
+end
+
+group :test do
+  gem 'test-unit'
+end

data/Indirizzo.gemspec

@@ -1,6 +1,6 @@
 Gem::Specification.new do |s|
   s.name = %q{Indirizzo}
-  s.version = "0.1.0"
+  s.version = "0.1.1"
 
   s.authors = [%q{Dave Worth}]
   s.date = %q{2011-12-14}

data/Rakefile

@@ -26,6 +26,6 @@ namespace :cover_me do
   end
 end
 
-task :test do
-  Rake::Task['cover_me:report'].invoke
-end
+#task :test do
+  #Rake::Task['cover_me:report'].invoke
+#end

data/lib/indirizzo.rb

@@ -1 +1 @@
-require 'Indirizzo/Address'
+require 'indirizzo/address'

data/lib/indirizzo/address.rb

@@ -1,4 +1,5 @@
 require 'indirizzo/constants'
+require 'awesome_print'
 
 module Indirizzo
   # Defines the matching of parsed address tokens.
@@ -6,9 +7,9 @@ module Indirizzo
     # 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'-]+\s*)+/io,
-    :state    => Regexp.new(State.regexp.source + "\s*$", Regexp::IGNORECASE),
-    :zip      => /(\d{5})(?:-\d{4})?\s*$/o,
+    :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/
   }
@@ -23,6 +24,7 @@ module Indirizzo
     attr_accessor :city
     attr_accessor :state
     attr_accessor :zip, :plus4
+    attr_accessor :country
 
     # Takes an address or place name string as its sole argument.
     def initialize (text)
@@ -122,55 +124,51 @@ module Indirizzo
       strings
     end
 
-    def parse_zip(regex_match, text)
-      idx = text.rindex(regex_match)
-      text[idx...idx+regex_match.length] = ""
-      text.sub! /\s*,?\s*$/o, ""
-      @zip, @plus4 = @zip.map {|s|s.strip}
-      text
-    end
-
     def parse_state(regex_match, text)
       idx = text.rindex(regex_match)
-      text[idx...idx+regex_match.length] = ""
-      text.sub! /\s*,?\s*$/o, ""
       @full_state = @state[0].strip # special case: New York
       @state = State[@full_state]
-      text
-    end
-
-    def parse_number(regex_match, text)
-      # FIXME: What if this string appears twice?
-      idx = text.index(regex_match)
-      text[idx...idx+regex_match.length] = ""
-      text.sub! /^\s*,?\s*/o, ""
-      @prenum, @number, @sufnum = @number.map {|s| s and s.strip}
+      @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])[-1]
+      @zip = text.scan(Match[:zip]).last
       if @zip
-        text = parse_zip($&, text)
+        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
 
-      @state = text.scan(Match[:state])[-1]
+      @country = @text[zip_end_index+1..-1].sub(/^\s*,\s*/, '').strip
+      @country = nil if @country == text
+
+      @state = text.scan(Match[:state]).last
       if @state
-        text = parse_state($&, text)
+        last_match = $&
+        state_index = text.rindex(last_match)
+        text = parse_state(last_match, text)
       else
         @full_state = ""
         @state = ""
       end
 
-      @number = text.scan(Match[:number])[0]
+      @number = text.scan(Match[:number]).first
       # FIXME: 230 Fish And Game Rd, Hudson NY 12534
       if @number # and not intersection?
-        text = parse_number($&, text)
+        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
 
@@ -179,24 +177,32 @@ module Indirizzo
       # Sault Ste. Marie
 
       # FIXME: PO Box should geocode to ZIP
-      @street = text.scan(Match[:street])
+      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)
       # SPECIAL CASE: 1600 Pennsylvania 20050
       @street << @full_state if @street.empty? and @state.downcase != @full_state.downcase
 
-      @city = text.scan(Match[:city])
-      if !@city.empty?
-        @city = [@city[-1].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 = []
+      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
 
-      # 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
 
     def expand_streets(street)

data/lib/indirizzo/constants.rb

@@ -20,10 +20,10 @@ module Indirizzo
     # whitespace-delimited prefixes to keys and values in the two-way Map.
     def build_partial
       @partial = Set.new()
-      [keys, values].flatten.each {|item|
+      [keys, values].flatten.each do |item|
         @partial << item.downcase
         item.downcase.split.each {|token| @partial << token}
-      }
+      end
     end
     def build_match
       @regexp = Regexp.new(
@@ -612,6 +612,8 @@ module Indirizzo
     "Connecticut"	=> "CT",
     "Delaware"		=> "DE",
     "District of Columbia" => "DC",
+    "Washington DC" => "DC",
+    "Washington D.C." => "DC",
     "Federated States of Micronesia" => "FM",
     "Florida"		=> "FL",
     "Georgia"		=> "GA",

data/test/test_address.rb

@@ -9,18 +9,6 @@ class TestAddress < Test::Unit::TestCase
     addr = Address.new("1600 Pennsylvania Av., Washington DC")
     assert_equal "1600 Pennsylvania Av, Washington DC", addr.text
   end
-  def test_clean
-    fixtures = [
-      [ "cleaned text", "cleaned: text!" ],
-      [ "cleaned-text 2", "cleaned-text: #2?" ],
-      [ "it's working 1/2", "~it's working 1/2~" ],
-      [ "it's working, yes", "it's working, yes...?" ],
-      [ "it's working & well", "it's working & well?" ]
-    ]
-    fixtures.each {|output, given|
-      assert_equal output, Address.new(given).text
-    }
-  end
   def test_expand_numbers
     num_list = ["5", "fifth", "five"]
     num_list.each {|n|
@@ -28,146 +16,12 @@ class TestAddress < Test::Unit::TestCase
       assert_equal num_list, addr.expand_numbers(n).to_a.sort
     }
   end
-  def test_city_parse
-    places = [
-      [ "New York, NY",     "New York", "NY", "" ],
-      [ "NY",               "", "NY",   "" ],
-      [ "New York",         "New York", "NY",   "" ],
-      [ "Philadelphia",     "Philadelphia", "", "" ],
-      [ "Philadelphia PA",  "Philadelphia", "PA", "" ],
-      [ "Philadelphia, PA", "Philadelphia", "PA", "" ],
-      [ "Philadelphia, Pennsylvania", "Philadelphia", "PA", "" ],
-      [ "Philadelphia, Pennsylvania 19131", "Philadelphia", "PA", "19131" ],
-      [ "Philadelphia 19131", "Philadelphia", "", "19131" ],
-      [ "Pennsylvania 19131", "Pennsylvania", "PA", "19131" ], # kind of a misfeature
-      [ "19131", "", "", "19131" ],
-      [ "19131-9999", "", "", "19131" ],
-    ]
-    for fixture in places
-      addr  = Address.new fixture[0]
-      [:city, :state, :zip].zip(fixture[1..3]).each {|key,val|
-        result = addr.send key
-        result = [result.downcase] unless result.kind_of? Array
-        if result.empty?
-          assert_equal val, "", key.to_s + " test no result " + fixture.join("/")
-        else
-          assert result.member?(val.downcase), key.to_s + " test " + result.inspect + fixture.join("/")
-        end
-      }
-    end
-  end
 
   def test_po_box
     addr_po = Address.new "PO Box 1111 Herndon VA 20171"
     assert addr_po.po_box?
   end
 
-  def test_parse
-    addrs = [
-      {:text   => "1600 Pennsylvania Av., Washington DC 20050",
-       :number => "1600",
-       :street => "Pennsylvania Ave",
-       :city   => "Washington",
-       :state  => "DC",
-       :zip    => "20050"},
-
-      {:text   => "1600 Pennsylvania, Washington DC",
-       :number => "1600",
-       :street => "Pennsylvania",
-       :city   => "Washington",
-       :state  => "DC"},
-
-      {:text   => "1600 Pennsylvania Washington DC",
-       :number => "1600",
-       :street => "Pennsylvania Washington",
-       :city   => "Pennsylvania Washington", # FIXME
-       :state  => "DC"},
-
-      {:text   => "1600 Pennsylvania Washington",
-       :number => "1600",
-       :street => "Pennsylvania",
-       :city   => "Washington",
-       :state  => "WA"}, # FIXME
-
-      {:text   => "1600 Pennsylvania 20050",
-       :number => "1600",
-       :street => "Pennsylvania", # FIXME
-       :zip    => "20050"},
-
-      {:text   => "1600 Pennsylvania Av, 20050-9999",
-       :number => "1600",
-       :street => "Pennsylvania Ave",
-       :zip    => "20050"},
-
-      {:text   => "1005 Gravenstein Highway North, Sebastopol CA",
-       :number => "1005",
-       :street => "Gravenstein Hwy N",
-       :city   => "Sebastopol",
-       :state  => "CA"},
-
-      {:text   => "100 N 7th St, Brooklyn",
-       :number => "100",
-       :street => "N 7 St",
-       :city   => "Brooklyn"},
-
-      {:text   => "100 N Seventh St, Brooklyn",
-       :number => "100",
-       :street => "N 7 St",
-       :city   => "Brooklyn"},
-
-      {:text   => "100 Central Park West, New York, NY",
-       :number => "100",
-       :street => "Central Park W",
-       :city   => "New York",
-       :state  => "NY"},
-
-      {:text   => "100 Central Park West, 10010",
-       :number => "100",
-       :street => "Central Park W",
-       :zip    => "10010"},
-
-      {:text   => "1400 Avenue of the Americas, New York, NY 10019",
-       :number => "1400",
-       :street => "Ave of the Americas",
-       :city   => "New York",
-       :state  => "NY"},
-
-      {:text   => "1400 Avenue of the Americas, New York",
-       :number => "1400",
-       :street => "Ave of the Americas",
-       :city   => "New York"},
-
-      {:text   => "1400 Ave of the Americas, New York",
-       :number => "1400",
-       :street => "Ave of the Americas",
-       :city   => "New York"},
-
-      {:text   => "1400 Av of the Americas, New York",
-       :number => "1400",
-       :street => "Ave of the Americas",
-       :city   => "New York"},
-
-      {:text   => "1400 Av of the Americas New York",
-       :number => "1400",
-       :street => "Ave of the Americas",
-       :city   => "New York"},
-
-    ]
-    for fixture in addrs
-      text = fixture.delete(:text)
-      addr = Address.new(text)
-      for key, val in fixture
-        result = addr.send key
-        if result.kind_of? Array
-          result.map! {|str| str.downcase}
-          assert result.member?(val.downcase), "#{text} (#{key}) = #{result.inspect}"
-        else
-          assert_equal val, result, "#{text} (#{key}) = #{result.inspect}"
-        end
-      end
-    end
-  end
-
   def test_skip_parse
     addresses = [
       {:street => "1233 Main St", :city => "Springfield", :region => "VA", :postal_code => "12345", :final_number => "1233", :parsed_street => "main st"},
@@ -225,4 +79,201 @@ class TestAddress < Test::Unit::TestCase
         assert_equal preparsed_address[:country],address_for_geocode.state
       end
   end
+
+  # test cleaning code
+  [
+    [ "cleaned text", "cleaned: text!" ],
+    [ "cleaned-text 2", "cleaned-text: #2?" ],
+    [ "it's working 1/2", "~it's working 1/2~" ],
+    [ "it's working, yes", "it's working, yes...?" ],
+    [ "it's working & well", "it's working & well?" ]
+  ].each do |output, given|
+    define_method "test_clean_#{output.tr('-/\'&', '').gsub(/\s+/, '_')}" do
+      assert_equal output, Address.new(given).text
+    end
+  end
+
+  # test the city parsing code
+  [
+    [ "New York, NY",     "New York", "NY", "" ],
+    [ "NY",               "", "NY",   "" ],
+    [ "New York",         "New York", "NY",   "" ],
+    [ "Philadelphia",     "Philadelphia", "", "" ],
+    [ "Philadelphia PA",  "Philadelphia", "PA", "" ],
+    [ "Philadelphia, PA", "Philadelphia", "PA", "" ],
+    [ "Philadelphia, Pennsylvania", "Philadelphia", "PA", "" ],
+    [ "Philadelphia, Pennsylvania 19131", "Philadelphia", "PA", "19131" ],
+    [ "Philadelphia 19131", "Philadelphia", "", "19131" ],
+    [ "Pennsylvania 19131", "Pennsylvania", "PA", "19131" ], # kind of a misfeature
+    [ "19131", "", "", "19131" ],
+    [ "19131-9999", "", "", "19131" ],
+  ].each do |fixture|
+    define_method "test_city_parse_#{fixture[0].gsub(/(?:\s+|[,])/,'_')}" do
+      check_city(fixture)
+    end
+  end
+
+  def check_city(fixture)
+      addr  = Address.new(fixture[0])
+      [:city, :state, :zip].zip(fixture[1..3]).each do |key,val|
+        result = addr.send(key)
+        result = [result.downcase] unless result.kind_of?(Array)
+        if result.empty?
+          assert_equal val, "", key.to_s + " test no result " + fixture.join("/")
+        else
+          assert result.member?(val.downcase), key.to_s + " test " + result.inspect + fixture.join("/")
+        end
+      end
+  end
+
+  # test address parsing code
+  [
+    {:text   => "1600 Pennsylvania Av., Washington DC 20050",
+     :number => "1600",
+     :street => "Pennsylvania Ave",
+     :city   => "Washington",
+     :state  => "DC",
+     :zip    => "20050"},
+
+    {:text   => "1600 Pennsylvania, Washington DC",
+     :number => "1600",
+     :street => "Pennsylvania",
+     :city   => "Washington",
+     :state  => "DC"},
+
+    {:text   => "1600 Pennsylvania Washington DC",
+     :number => "1600",
+     :city   => "Washington",
+     :street => "Pennsylvania",
+     :state  => "DC"},
+
+    {:text   => "1600 Pennsylvania Washington",
+     :pending => true,
+     :number => "1600",
+     :street => "Pennsylvania",
+     :city   => "Washington",
+     :state  => "DC"},
+
+    {:text   => "1600 Pennsylvania 20050",
+     :number => "1600",
+     :state  => "PA",
+     :zip    => "20050"},
+
+    {:text   => "1600 Pennsylvania Av, Washington DC 20050-9999",
+     :number => "1600",
+     :state  => "DC",
+     :street => "Pennsylvania Ave",
+     :plus4  => "9999",
+     :zip    => "20050"},
+
+    {:text   => "1600 Pennsylvania Av, 20050-9999",
+     :pending => true,
+     :number => "1600",
+     #:state  => "PA",
+     :street => "Pennsylvania Ave",
+     :plus4  => "9999",
+     :zip    => "20050"},
+
+    {:text   => "1005 Gravenstein Highway North, Sebastopol CA",
+     :number => "1005",
+     :street => "Gravenstein Hwy N",
+     :city   => "Sebastopol",
+     :state  => "CA"},
+
+    {:text   => "100 N 7th St, Brooklyn",
+     :number => "100",
+     :street => "N 7 St",
+     :city   => "Brooklyn"},
+
+    {:text   => "100 N Seventh St, Brooklyn",
+     :number => "100",
+     :street => "N 7 St",
+     :city   => "Brooklyn"},
+
+    {:text   => "100 Central Park West, New York, NY",
+     :number => "100",
+     :street => "Central Park W",
+     :city   => "New York",
+     :state  => "NY"},
+
+    {:text   => "100 Central Park West, 10010",
+     :number => "100",
+     :street => "Central Park W",
+     :zip    => "10010"},
+
+    {:text   => "1400 Avenue of the Americas, New York, NY 10019, US",
+     :number => "1400",
+     :street => "Ave of the Americas",
+     :city   => "New York",
+     :state  => "NY",
+     :country => "US"},
+
+    {:text   => "1400 Avenue of the Americas, New York, NY 10019, USA",
+     :number => "1400",
+     :street => "Ave of the Americas",
+     :city   => "New York",
+     :state  => "NY",
+     :country => "USA"},
+
+    {:text   => "1400 Avenue of the Americas, New York, NY 10019,   United States of America  ",
+     :number => "1400",
+     :street => "Ave of the Americas",
+     :city   => "New York",
+     :state  => "NY",
+     :country => "United States of America"},
+
+    {:text   => "1400 Avenue of the Americas, New York",
+     :number => "1400",
+     :street => "Ave of the Americas",
+     :city   => "New York"},
+
+    {:text   => "1400 Ave of the Americas, New York",
+     :number => "1400",
+     :street => "Ave of the Americas",
+     :city   => "New York"},
+
+    {:text   => "1400 Av of the Americas, New York",
+     :number => "1400",
+     :street => "Ave of the Americas",
+     :city   => "New York"},
+
+    {:text   => "1400 Av of the Americas New York",
+     :number => "1400",
+     :street => "Ave of the Americas",
+     :city   => "New York"},
+
+    {:text   => "23 Home St,    Hometown  PA,  12345  US",
+     :number => "23",
+     :state  => "PA",
+     :street => "Home St",
+     :city   => "Hometown",
+     :zip    => "12345"},
+
+    {:text   => "23 Home St, Apt. A, Hometown  PA,  12345  US",
+     :number => "23",
+     :state  => "PA",
+     :street => "Home St",
+     :city   => "Hometown",
+     :zip    => "12345",
+     :country => "US"}
+  ].each do |fixture|
+    define_method "test_parse_#{fixture[:text].gsub(/(?:\s+|[.,])/,'_')}" do
+      pend if fixture[:pending]
+      check_addr(fixture)
+    end
+  end
+
+  def check_addr(fixture)
+    text = fixture[:text]
+    addr = Address.new(text)
+    fixture.reject{|k,v| k == :text}.each do |key, val|
+      result = addr.send key
+      if result.kind_of? Array
+        result.map! {|str| str.downcase}
+        assert result.member?(val.downcase), "#{text} (#{key}) = #{result.inspect}"
+      else
+        assert_equal val, result, "#{text} (#{key}) = #{result.inspect}"
+      end
+    end
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: Indirizzo
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
   prerelease: 
 platform: ruby
 authors:
@@ -13,7 +13,7 @@ date: 2011-12-14 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &70182937832940 !ruby/object:Gem::Requirement
+  requirement: &70302219305600 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70182937832940
+  version_requirements: *70302219305600
 - !ruby/object:Gem::Dependency
   name: cover_me
-  requirement: &70182937832500 !ruby/object:Gem::Requirement
+  requirement: &70302219305160 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,10 +32,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70182937832500
+  version_requirements: *70302219305160
 - !ruby/object:Gem::Dependency
   name: awesome_print
-  requirement: &70182937832080 !ruby/object:Gem::Requirement
+  requirement: &70302219304740 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,7 +43,7 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70182937832080
+  version_requirements: *70302219304740
 description: Indirizzo is simply an extraction of the US Street Address parsing code
   from Geocoder::US
 email: dave@highgroove.com