gman 4.1.4 → 4.2.0

data/lib/gman.rb

@@ -1,20 +1,31 @@
 require 'naughty_or_nice'
 require 'swot'
 require 'iso_country_codes'
+require 'csv'
 require_relative 'gman/country_codes'
 require_relative 'gman/locality'
+require_relative 'gman/identifier'
 
 class Gman < NaughtyOrNice
   class << self
+
     # returns an instance of our custom public suffix list
     # list behaves like PublicSuffix::List but is limited to our whitelisted domains
     def list
-      @list ||= PublicSuffix::List::parse(File.new(list_path, "r:utf-8"))
+      @list ||= PublicSuffix::List::parse(list_contents)
+    end
+
+    def config_path
+      File.join(File.dirname(__FILE__), "../config")
     end
 
     # Returns the absolute path to the domain list
     def list_path
-      File.join(File.dirname(__FILE__), "../config/domains.txt")
+      File.join(config_path,"domains.txt")
+    end
+
+    def list_contents
+      @list_contents ||= File.new(list_path, "r:utf-8").read
     end
   end
 

data/lib/gman/identifier.rb

@@ -0,0 +1,109 @@
+class Gman < NaughtyOrNice
+
+  def type
+    if state?
+      :state
+    elsif district?
+      :district
+    elsif cog?
+      :cog
+    elsif city?
+      :city
+    elsif federal?
+      :federal
+    elsif county?
+      :county
+    elsif list_category.include?("usagov")
+      :unknown
+    else
+      list_category.to_sym
+    end
+  end
+
+  def state
+    if matches
+      matches[4].upcase
+    elsif dotgov_listing
+      dotgov_listing["State"]
+    elsif list_category
+      matches = list_category.match(/usagov([A-Z]{2})/)
+      matches[1] if matches
+    end
+  end
+
+  def city
+    dotgov_listing["City"] if dotgov_listing
+  end
+
+  def agency
+    dotgov_listing["Agency"] if federal?
+  end
+
+  def dotgov?
+    domain_parts.tld == "gov"
+  end
+
+  def federal?
+    dotgov_listing && dotgov_listing["Domain Type"] == "Federal Agency"
+  end
+
+  def city?
+    if matches
+      %w[ci town vil].include?(matches[3])
+    elsif dotgov_listing
+      dotgov_listing["Domain Type"] == "City"
+    end
+  end
+
+  def county?
+    if matches
+      matches[3] == "co"
+    elsif dotgov_listing
+      dotgov_listing["Domain Type"] == "County"
+    end
+  end
+
+  def state?
+    if matches
+      matches[1] == "state"
+    elsif dotgov_listing
+      dotgov_listing["Domain Type"] == "State/Local Govt"
+    end
+  end
+
+  def district?
+    !!matches && matches[1] == "dst"
+  end
+
+  def cog?
+    !!matches && matches[1] == "cog"
+  end
+
+  private
+
+  def list_category
+    @list_category ||= begin
+      if match = Gman.list.find(domain)
+        regex = Regexp.new "\/\/ ([^\\n]+)\\n?[^\/\/]*\\n#{Regexp.escape(match.name)}\\n", "im"
+        matches = Gman.list_contents.match(regex)
+        matches[1] if matches
+      end
+    end
+  end
+
+  def matches
+    @matches ||= domain.match(LOCALITY_REGEX)
+  end
+
+  def self.dotgov_list_path
+    File.join Gman.config_path, "vendor/dotgovs.csv"
+  end
+
+  def self.dotgov_list
+    @dotgov_list ||= CSV.read(dotgov_list_path, :headers => true)
+  end
+
+  def dotgov_listing
+    @dotgov_listing ||= Gman.dotgov_list.find { |d| d["Domain Name"].downcase == "#{domain_parts.sld}.gov" } if dotgov?
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: gman
 version: !ruby/object:Gem::Version
-  version: 4.1.4
+  version: 4.2.0
 platform: ruby
 authors:
 - Ben Balter
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-12-23 00:00:00.000000000 Z
+date: 2014-12-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: swot
@@ -147,8 +147,10 @@ files:
 - LICENSE
 - bin/gman_filter
 - config/domains.txt
+- config/vendor/dotgovs.csv
 - lib/gman.rb
 - lib/gman/country_codes.rb
+- lib/gman/identifier.rb
 - lib/gman/locality.rb
 homepage: https://github.com/benbalter/gman
 licenses: