StreetAddress 1.0.3 → 1.0.4

data/README.rdoc

@@ -1,7 +1,3 @@
-StreetAddress
-    by Derrek Long
-    https://github.com/derrek/street-address
-
 == DESCRIPTION:
   
 Parses one line street addresses and returns a normalized address object.
@@ -30,6 +26,20 @@ Parse United States addresses.
   address.to_s(:line1) # 1600 Pennsylvania Ave
   StreetAddress::US.parse("1600 Pennsylvania Ave") # nil not a full address
 
+=== Informal/Loose Parsing
+  address = StreetAddress::US.parse("1600 Pennsylvania Avenue", :informal => true)
+  address.number # 1600
+  address.street # Pennsylvania
+  address.street_type # Ave
+  address.to_s # 1600 Pennsylvania Ave
+
+=== Zip+4
+  address = StreetAddress::US.parse("5904 Richmond Hwy Ste 340 Alexandria VA 22303-1864")
+  address.postal_code_ext # 1846
+  
+  address = StreetAddress::US.parse("5904 Richmond Hwy Ste 340 Alexandria VA 223031864")
+  address.postal_code_ext # 1846
+
 === Gemfile:
 To use from your Gemfile:
-  gem 'StreetAddress', '1.0.2', :require => "street_address"
+  gem 'StreetAddress', '1.0.3', :require => "street_address"

data/lib/street_address.rb

@@ -64,7 +64,7 @@
 
 module StreetAddress
   class US
-    VERSION = '1.0.3'
+    VERSION = '1.0.4'
     DIRECTIONAL = {
       "north" => "N",
       "northeast" => "NE",
@@ -576,13 +576,15 @@ module StreetAddress
         :number_regexp,
         :fraction_regexp,
         :state_regexp,
+        :city_and_state_regexp,
         :direct_regexp, 
         :zip_regexp,
         :corner_regexp,
         :unit_regexp,
         :street_regexp,
         :place_regexp,
-        :address_regexp
+        :address_regexp,
+        :informal_address_regexp
       )
     end
       
@@ -590,6 +592,12 @@ module StreetAddress
     self.number_regexp = '\d+-?\d*'
     self.fraction_regexp = '\d+\/\d+'
     self.state_regexp = STATE_CODES.to_a.join("|").gsub(/ /, "\\s")
+    self.city_and_state_regexp = '
+      (?:
+        ([^\d,]+?)\W+
+        (' + state_regexp + ')
+      )'
+      
     self.direct_regexp = DIRECTIONAL.keys.join("|") + 
       "|" + 
       DIRECTIONAL.values.sort{ |a,b| 
@@ -598,7 +606,7 @@ module StreetAddress
         f = x.gsub(/(\w)/, '\1.')
         [Regexp::quote(f), Regexp::quote(x)] 
       }.join("|")
-    self.zip_regexp = '(\d{5})(?:-(\d{4}))?'
+    self.zip_regexp = '(\d{5})(?:-?(\d{4})?)'
     self.corner_regexp = '(?:\band\b|\bat\b|&|\@)'
     self.unit_regexp = '(?:(su?i?te|p\W*[om]\W*b(?:ox)?|dept|apt|apartment|ro*m|fl|unit|box)\W+|\#\W*)([\w-]+)'
     self.street_regexp = 
@@ -621,10 +629,7 @@ module StreetAddress
           )
         )'
     self.place_regexp = 
-      '(?:
-       ([^\d,]+?)\W+
-       ($' + state_regexp + ')\W*
-       )?
+      '(?:' + city_and_state_regexp + '\W*)?
        (?:' + zip_regexp + ')?'
     
     self.address_regexp =
@@ -636,26 +641,39 @@ module StreetAddress
         place_regexp +
       '\W*\Z'
       
+    self.informal_address_regexp =
+      '\A\s*
+        (' + number_regexp + ')\W*
+        (?:' + fraction_regexp + '\W*)?' +
+        street_regexp + '(?:\W+|\Z)
+        (?:' + unit_regexp + '(?:\W+|\Z))?' +
+        '(?:' + place_regexp + ')?'
+
 =begin rdoc
 
     parses either an address or intersection and returns an instance of
     StreetAddress::US::Address or nil if the location cannot be parsed
+
+    pass the arguement, :informal => true, to make parsing more lenient
     
 ====example
     StreetAddress::US.parse('1600 Pennsylvania Ave Washington, DC 20006')
     or:
     StreetAddress::US.parse('Hollywood & Vine, Los Angeles, CA')
+    or
+    StreetAddress::US.parse("1600 Pennsylvania Ave", :informal => true)
     
 =end
     class << self
-      def parse(location)
+      def parse(location, args = {})
         if Regexp.new(corner_regexp, Regexp::IGNORECASE).match(location)
-          parse_intersection(location);
+          parse_intersection(location)
+        elsif args[:informal]
+          parse_address(location) || parse_informal_address(location)
         else 
           parse_address(location);
         end
       end
-
 =begin rdoc
     
     parses only an intersection and returnsan instance of
@@ -724,6 +742,28 @@ module StreetAddress
            )
         )
       end
+
+      def parse_informal_address(addr)
+         regex = Regexp.new(informal_address_regexp, Regexp::IGNORECASE + Regexp::EXTENDED)
+
+         return unless match = regex.match(addr)
+
+         normalize_address(
+           StreetAddress::US::Address.new(
+           :number => match[1],
+           :street => match[5] || match[10] || match[2],
+           :street_type => match[6] || match[3],
+           :unit => match[14],
+           :unit_prefix => match[13],
+           :suffix => match[7] || match[12],
+           :prefix => match[4],
+           :city => match[15],
+           :state => match[16],
+           :postal_code => match[17],
+           :postal_code_ext => match[18]
+           )
+        )
+      end
       
       private
       def normalize_address(addr)

data/test/test_street_address.rb

@@ -17,6 +17,44 @@ class StreetAddressUsTest < Test::Unit::TestCase
 
   end
 
+  def test_zip_plus_4_with_dash
+    addr = StreetAddress::US.parse("2730 S Veitch St, Arlington, VA 22206-3333")
+    assert_equal "3333", addr.postal_code_ext
+  end
+
+  def test_zip_plus_4_without_dash
+    addr = StreetAddress::US.parse("2730 S Veitch St, Arlington, VA 222064444")
+    assert_equal "4444", addr.postal_code_ext
+  end
+
+  def test_informal_parse_normal_address
+    a = StreetAddress::US.parse("2730 S Veitch St, Arlington, VA 222064444", :informal => true)
+    assert_equal "2730", a.number
+    assert_equal "S", a.prefix
+    assert_equal "Veitch", a.street
+    assert_equal "St", a.street_type
+    assert_equal "Arlington", a.city
+    assert_equal "VA", a.state
+    assert_equal "22206", a.postal_code
+    assert_equal "4444", a.postal_code_ext
+  end
+
+  def test_informal_parse_informal_address
+    a = StreetAddress::US.parse("2730 S Veitch St", :informal => true)
+    assert_equal "2730", a.number
+    assert_equal "S", a.prefix
+    assert_equal "Veitch", a.street
+    assert_equal "St", a.street_type
+  end
+
+  def test_informal_parse_informal_address_trailing_words
+    a = StreetAddress::US.parse("2730 S Veitch St in the south of arlington", :informal => true)
+    assert_equal "2730", a.number
+    assert_equal "S", a.prefix
+    assert_equal "Veitch", a.street
+    assert_equal "St", a.street_type
+  end
+
   def test_parse
     assert_equal StreetAddress::US.parse("&"), nil
     assert_equal StreetAddress::US.parse(" and "), nil

metadata

@@ -1,76 +1,58 @@
---- !ruby/object:Gem::Specification 
+--- !ruby/object:Gem::Specification
 name: StreetAddress
-version: !ruby/object:Gem::Version 
-  hash: 17
+version: !ruby/object:Gem::Version
+  version: 1.0.4
   prerelease: 
-  segments: 
-  - 1
-  - 0
-  - 3
-  version: 1.0.3
 platform: ruby
-authors: 
+authors:
 - Derrek Long
 - Nicholas Schleuter
 autorequire: 
 bindir: bin
 cert_chain: []
-
-date: 2011-09-21 00:00:00 -07:00
-default_executable: 
+date: 2013-01-17 00:00:00.000000000 Z
 dependencies: []
+description: ! 'StreetAddress::US allows you to send any string to parse and if the
+  string is a US address returns an object of the address broken into it''s substituent
+  parts.
+
 
-description: |
-  StreetAddress::US allows you to send any string to parse and if the string is a US address returns an object of the address broken into it's substituent parts.
-  
   A port of Geo::StreetAddress::US by Schuyler D. Erle and Tim Bunce.
 
+'
 email: derreklong@gmail.com
 executables: []
-
 extensions: []
-
 extra_rdoc_files: []
-
-files: 
+files:
 - README.rdoc
 - Rakefile
 - LICENSE
 - lib/street_address.rb
 - test/test_street_address.rb
-has_rdoc: true
 homepage: https://github.com/derrek/street-address
 licenses: []
-
 post_install_message: 
 rdoc_options: []
-
-require_paths: 
+require_paths:
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement 
+required_ruby_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      hash: 3
-      segments: 
-      - 0
-      version: "0"
-required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      hash: 3
-      segments: 
-      - 0
-      version: "0"
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
 requirements: []
-
 rubyforge_project: 
-rubygems_version: 1.4.2
+rubygems_version: 1.8.24
 signing_key: 
 specification_version: 3
 summary: Parse Addresses into substituent parts. This gem includes US only.
-test_files: 
+test_files:
 - test/test_street_address.rb