normalic 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- data/Manifest +6 -0
- data/README.rdoc +15 -0
- data/Rakefile +15 -0
- data/lib/constants.rb +42337 -0
- data/lib/normalic.rb +131 -0
- data/normalic.gemspec +29 -0
- data/spec/normalic_spec.rb +78 -0
- metadata +79 -0
data/lib/normalic.rb
ADDED
@@ -0,0 +1,131 @@
|
|
1
|
+
#only handles U.S addresses
|
2
|
+
require 'constants'
|
3
|
+
|
4
|
+
module Normalic
|
5
|
+
class Address
|
6
|
+
|
7
|
+
attr_accessor :number, :direction, :street, :type, :city, :state, :zipcode
|
8
|
+
|
9
|
+
def initialize(fields={})
|
10
|
+
@number = fields[:number]
|
11
|
+
@direction = fields[:direction]
|
12
|
+
@street = fields[:street]
|
13
|
+
@type = fields[:type]
|
14
|
+
@city = fields[:city]
|
15
|
+
@state = fields[:state]
|
16
|
+
@zipcode = fields[:zipcode]
|
17
|
+
end
|
18
|
+
|
19
|
+
def self.titlize(str)
|
20
|
+
if str
|
21
|
+
str.gsub(/\w+/){|w| w.capitalize}
|
22
|
+
else
|
23
|
+
nil
|
24
|
+
end
|
25
|
+
end
|
26
|
+
|
27
|
+
def [](field_name)
|
28
|
+
begin
|
29
|
+
self.send(field_name.to_s)
|
30
|
+
rescue NoMethodError => e
|
31
|
+
nil
|
32
|
+
end
|
33
|
+
end
|
34
|
+
|
35
|
+
def []=(field_name, value)
|
36
|
+
begin
|
37
|
+
self.send("#{field_name}=", value)
|
38
|
+
rescue NoMethodError => e
|
39
|
+
nil
|
40
|
+
end
|
41
|
+
end
|
42
|
+
|
43
|
+
def to_s
|
44
|
+
#"#{line1},#{" #{city.gsub(/\w+/){|w| w.capitalize}}," if city}#{" #{state.upcase}" if state}#{" " + zipcode if zipcode}".strip
|
45
|
+
"#{line1}#{", #{city}" if city}#{", #{state}" if state}#{" " + zipcode if zipcode}".strip
|
46
|
+
#"#{line1}, #{city}, #{state} #{zipcode}"
|
47
|
+
end
|
48
|
+
|
49
|
+
def line1
|
50
|
+
#"#{number}#{" " + direction.upcase if direction}#{" " + street.gsub(/\w+/){|w| w.capitalize} if street}#{" " + type.capitalize if type}".strip
|
51
|
+
"#{number}#{" " + direction if direction}#{" " + street if street}#{" " + type if type}"
|
52
|
+
end
|
53
|
+
|
54
|
+
#Iteratively take chunks off of the string.
|
55
|
+
def self.parse(address)
|
56
|
+
address.strip!
|
57
|
+
regex = {
|
58
|
+
:unit => /(((\#?\w*)?\W*(su?i?te|p\W*[om]\W*b(?:ox)?|dept|department|ro*m|floor|fl|apt|apartment|unit|box))$)|(\W((su?i?te|p\W*[om]\W*b(?:ox)?|dept|department|ro*m|floor|fl|apt|apartment|unit|box)\W*(\#?\w*)?)\W{0,3}$)/i,
|
59
|
+
:direct => Regexp.new(Directional.keys * '|' + '|' + Directional.values * '\.?|',Regexp::IGNORECASE),
|
60
|
+
:type => Regexp.new('(' + StreetTypes_list * '|' + ')\\W*?$',Regexp::IGNORECASE),
|
61
|
+
:number => /\d+-?\d*/,
|
62
|
+
:fraction => /\d+\/\d+/,
|
63
|
+
:country => /\W+USA$/,
|
64
|
+
:zipcode => /\W+(\d{5}|\d{5}-\d{4})$/,
|
65
|
+
:state => Regexp.new('\W+(' + StateCodes.values * '|' + '|' + StateCodes.keys * '|' + ')$',Regexp::IGNORECASE),
|
66
|
+
}
|
67
|
+
regex[:street] = Regexp.new('((' + regex[:direct].source + ')\\W)?\\W*(.*)\\W*(' + regex[:type].source + ')?', Regexp::IGNORECASE)
|
68
|
+
|
69
|
+
#get rid of USA at the end
|
70
|
+
country_code = address[regex[:country]]
|
71
|
+
address.gsub!(regex[:country], "")
|
72
|
+
zipcode = address[regex[:zipcode]]
|
73
|
+
address.gsub!(regex[:zipcode], "")
|
74
|
+
zipcode.gsub!(/\W/, "") if zipcode
|
75
|
+
|
76
|
+
state = address[regex[:state]]
|
77
|
+
address.gsub!(regex[:state], "")
|
78
|
+
state.gsub!(/(^\W*|\W*$)/, "").downcase! if state
|
79
|
+
state = StateCodes[state] || state
|
80
|
+
|
81
|
+
if ZipCityMap[zipcode]
|
82
|
+
regex[:city] = Regexp.new("\\W+" + ZipCityMap[zipcode] + "$", Regexp::IGNORECASE)
|
83
|
+
regex[:city] = /,.*$/ if !address[regex[:city]]
|
84
|
+
city = ZipCityMap[zipcode]
|
85
|
+
else
|
86
|
+
regex[:city] = /,.*$/
|
87
|
+
city = address[regex[:city]]
|
88
|
+
city.gsub!(/(^\W*|\W*$)/, "").downcase! if city
|
89
|
+
end
|
90
|
+
|
91
|
+
address.gsub!(regex[:city], "")
|
92
|
+
address.gsub!(regex[:unit], "")
|
93
|
+
address.gsub!(Regexp.new('\W(' + regex[:direct].source + ')\\W{0,3}$', Regexp::IGNORECASE), "")
|
94
|
+
|
95
|
+
type = address[regex[:type]]
|
96
|
+
address.gsub!(regex[:type], "")
|
97
|
+
type.gsub!(/(^\W*|\W*$)/, "").downcase! if type
|
98
|
+
type = StreetTypes[type] || type if type
|
99
|
+
|
100
|
+
if address =~ /(\Wand\W|\W\&\W)/
|
101
|
+
#intersections. print as is
|
102
|
+
address.gsub!(/(\Wand\W|\W\&\W)/, " and ")
|
103
|
+
arr = ["", address, "", ""]
|
104
|
+
else
|
105
|
+
regex[:address] = Regexp.new('^\W*(' + regex[:number].source + '\\W)?\W*(?:' + regex[:fraction].source + '\W*)?' + regex[:street].source, Regexp::IGNORECASE)
|
106
|
+
arr = regex[:address].match(address).to_a
|
107
|
+
end
|
108
|
+
|
109
|
+
number = arr[1].strip if arr[1]
|
110
|
+
if arr[2] && (!arr[4] || arr[4].empty?)
|
111
|
+
street = arr[2].strip.downcase
|
112
|
+
else
|
113
|
+
dir = Directional[arr[2].strip.downcase] || arr[2].strip.downcase if arr[2]
|
114
|
+
dir.gsub!(/\W/, "") if dir
|
115
|
+
end
|
116
|
+
street = arr[4].strip.downcase if arr[4] && !street
|
117
|
+
|
118
|
+
self.new(
|
119
|
+
{
|
120
|
+
:number => number,
|
121
|
+
:direction => dir ? dir.upcase : nil,
|
122
|
+
:street => titlize(street),
|
123
|
+
:type => titlize(type),
|
124
|
+
:city => titlize(city),
|
125
|
+
:state => state ? state.upcase : nil,
|
126
|
+
:zipcode => zipcode
|
127
|
+
}
|
128
|
+
)
|
129
|
+
end
|
130
|
+
end
|
131
|
+
end
|
data/normalic.gemspec
ADDED
@@ -0,0 +1,29 @@
|
|
1
|
+
# -*- encoding: utf-8 -*-
|
2
|
+
|
3
|
+
Gem::Specification.new do |s|
|
4
|
+
s.name = %q{normalic}
|
5
|
+
s.version = "0.1.0"
|
6
|
+
|
7
|
+
s.required_rubygems_version = Gem::Requirement.new(">= 1.2") if s.respond_to? :required_rubygems_version=
|
8
|
+
s.authors = [%q{Eric Tang}]
|
9
|
+
s.date = %q{2011-07-31}
|
10
|
+
s.description = %q{Normalize U.S addresses}
|
11
|
+
s.email = %q{eric.x.tang@gmail.com}
|
12
|
+
s.extra_rdoc_files = [%q{README.rdoc}, %q{lib/constants.rb}, %q{lib/normalic.rb}]
|
13
|
+
s.files = [%q{README.rdoc}, %q{Rakefile}, %q{lib/constants.rb}, %q{lib/normalic.rb}, %q{spec/normalic_spec.rb}, %q{Manifest}, %q{normalic.gemspec}]
|
14
|
+
s.homepage = %q{http://github.com/ericxtang/normalic}
|
15
|
+
s.rdoc_options = [%q{--line-numbers}, %q{--inline-source}, %q{--title}, %q{Normalic}, %q{--main}, %q{README.rdoc}]
|
16
|
+
s.require_paths = [%q{lib}]
|
17
|
+
s.rubyforge_project = %q{normalic}
|
18
|
+
s.rubygems_version = %q{1.8.6}
|
19
|
+
s.summary = %q{Normalize U.S addresses}
|
20
|
+
|
21
|
+
if s.respond_to? :specification_version then
|
22
|
+
s.specification_version = 3
|
23
|
+
|
24
|
+
if Gem::Version.new(Gem::VERSION) >= Gem::Version.new('1.2.0') then
|
25
|
+
else
|
26
|
+
end
|
27
|
+
else
|
28
|
+
end
|
29
|
+
end
|
@@ -0,0 +1,78 @@
|
|
1
|
+
require 'lib/normalic'
|
2
|
+
|
3
|
+
describe "Normalic test" do
|
4
|
+
|
5
|
+
it "should parse an address with unit(floor) information" do
|
6
|
+
addr = Normalic::Address.parse("201 Varick St. floor 12th, New York, NY 10014")
|
7
|
+
addr[:number].should == "201"
|
8
|
+
addr[:street].should == "varick"
|
9
|
+
addr[:type].should == "st"
|
10
|
+
addr[:city].should == "new york"
|
11
|
+
addr[:state].should == "ny"
|
12
|
+
addr[:zipcode].should == "10014"
|
13
|
+
end
|
14
|
+
|
15
|
+
it "should parse an address with direction information" do
|
16
|
+
addr = Normalic::Address.parse("167 West 4th Street, New York, NY 10014")
|
17
|
+
addr[:number].should == "167"
|
18
|
+
addr[:street].should == "4th"
|
19
|
+
addr[:direction].should == "w"
|
20
|
+
addr[:type].should == "st"
|
21
|
+
addr[:city].should == "new york"
|
22
|
+
addr[:state].should == "ny"
|
23
|
+
addr[:zipcode].should == "10014"
|
24
|
+
end
|
25
|
+
|
26
|
+
it "should parse an address with incorrect state info" do
|
27
|
+
addr = Normalic::Address.parse("871 Washington Street, New York, NewYork 10014")
|
28
|
+
addr[:number].should == "871"
|
29
|
+
addr[:street].should == "washington"
|
30
|
+
addr[:type].should == "st"
|
31
|
+
addr[:city].should == "new york"
|
32
|
+
addr[:state].should == "ny"
|
33
|
+
addr[:zipcode].should == "10014"
|
34
|
+
end
|
35
|
+
|
36
|
+
it "should parse an address with floor info and without city info" do
|
37
|
+
addr = Normalic::Address.parse("201 Varick St. floor 12th")
|
38
|
+
addr[:number].should == "201"
|
39
|
+
addr[:street].should == "varick"
|
40
|
+
addr[:type].should == "st"
|
41
|
+
end
|
42
|
+
|
43
|
+
it "should parse an address with no city info" do
|
44
|
+
addr = Normalic::Address.parse("871 Washington Street")
|
45
|
+
addr[:number].should == "871"
|
46
|
+
addr[:street].should == "washington"
|
47
|
+
addr[:type].should == "st"
|
48
|
+
end
|
49
|
+
|
50
|
+
|
51
|
+
it "should parse an address with direction info and no city info" do
|
52
|
+
addr = Normalic::Address.parse("871 West Washington Street")
|
53
|
+
addr[:number].should == "871"
|
54
|
+
addr[:street].should == "washington"
|
55
|
+
addr[:direction].should == "w"
|
56
|
+
addr[:type].should == "st"
|
57
|
+
end
|
58
|
+
|
59
|
+
it "should use dot notation" do
|
60
|
+
addr = Normalic::Address.parse("871 West Washington Street")
|
61
|
+
addr.number.should == "871"
|
62
|
+
end
|
63
|
+
|
64
|
+
it "should return nil if a bad field is passed in" do
|
65
|
+
addr = Normalic::Address.parse("871 West Washington Street")
|
66
|
+
addr[:bad_name].should == nil
|
67
|
+
end
|
68
|
+
|
69
|
+
it "should return a line1" do
|
70
|
+
addr = Normalic::Address.parse("871 West Washington Street")
|
71
|
+
addr.line1.should == "871 W Washington St"
|
72
|
+
end
|
73
|
+
|
74
|
+
it "should have a to_s method" do
|
75
|
+
addr = Normalic::Address.parse("167 West 4th Street, New York, NY 10014")
|
76
|
+
addr.to_s.should == "167 W 4th St, New York, NY 10014"
|
77
|
+
end
|
78
|
+
end
|
metadata
ADDED
@@ -0,0 +1,79 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: normalic
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
hash: 27
|
5
|
+
prerelease:
|
6
|
+
segments:
|
7
|
+
- 0
|
8
|
+
- 1
|
9
|
+
- 0
|
10
|
+
version: 0.1.0
|
11
|
+
platform: ruby
|
12
|
+
authors:
|
13
|
+
- Eric Tang
|
14
|
+
autorequire:
|
15
|
+
bindir: bin
|
16
|
+
cert_chain: []
|
17
|
+
|
18
|
+
date: 2011-07-31 00:00:00 Z
|
19
|
+
dependencies: []
|
20
|
+
|
21
|
+
description: Normalize U.S addresses
|
22
|
+
email: eric.x.tang@gmail.com
|
23
|
+
executables: []
|
24
|
+
|
25
|
+
extensions: []
|
26
|
+
|
27
|
+
extra_rdoc_files:
|
28
|
+
- README.rdoc
|
29
|
+
- lib/constants.rb
|
30
|
+
- lib/normalic.rb
|
31
|
+
files:
|
32
|
+
- README.rdoc
|
33
|
+
- Rakefile
|
34
|
+
- lib/constants.rb
|
35
|
+
- lib/normalic.rb
|
36
|
+
- spec/normalic_spec.rb
|
37
|
+
- Manifest
|
38
|
+
- normalic.gemspec
|
39
|
+
homepage: http://github.com/ericxtang/normalic
|
40
|
+
licenses: []
|
41
|
+
|
42
|
+
post_install_message:
|
43
|
+
rdoc_options:
|
44
|
+
- --line-numbers
|
45
|
+
- --inline-source
|
46
|
+
- --title
|
47
|
+
- Normalic
|
48
|
+
- --main
|
49
|
+
- README.rdoc
|
50
|
+
require_paths:
|
51
|
+
- lib
|
52
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
53
|
+
none: false
|
54
|
+
requirements:
|
55
|
+
- - ">="
|
56
|
+
- !ruby/object:Gem::Version
|
57
|
+
hash: 3
|
58
|
+
segments:
|
59
|
+
- 0
|
60
|
+
version: "0"
|
61
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
62
|
+
none: false
|
63
|
+
requirements:
|
64
|
+
- - ">="
|
65
|
+
- !ruby/object:Gem::Version
|
66
|
+
hash: 11
|
67
|
+
segments:
|
68
|
+
- 1
|
69
|
+
- 2
|
70
|
+
version: "1.2"
|
71
|
+
requirements: []
|
72
|
+
|
73
|
+
rubyforge_project: normalic
|
74
|
+
rubygems_version: 1.8.10
|
75
|
+
signing_key:
|
76
|
+
specification_version: 3
|
77
|
+
summary: Normalize U.S addresses
|
78
|
+
test_files: []
|
79
|
+
|