geomash 0.3.1 → 0.3.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: b181c5b1f79695f4b8ba54cf5e641c9cdaefe842
-  data.tar.gz: f6db8e134bf152b162765a86b688848f16d0753d
+  metadata.gz: f517ba6fcc94becafdbd6e43ae35ce833b69f728
+  data.tar.gz: fcc4f7d3194c8203e87bbf0a1b1192cf2e55804a
 SHA512:
-  metadata.gz: 0a5e74f73c036f36d0d030c2edc57fd2b026add354ee845e6f5bf6ab847f81d62dda8afe08add796b63ae0ee2b87446471a1ca083c60114d7401bc88e8476989
-  data.tar.gz: c0af8abaa745856c0e416f461e230574c869f5736f3c4aa8cb416dac102fe301eac669c1a308ff91d748d3e2be78511e3c0fb96fd8d0a7851e944573e0cd2506
+  metadata.gz: ee9bb319b8a8014ff077957c38f27521ff61a37836140bc389a25d3acd659ecec91cd3602ab59c96f14589d61d32c6e35d74825f421c3dd82aa273e93e85d2bc
+  data.tar.gz: 0cfd23abf1dd9c6f9f71dc7a4cd96da271e6e8ecb379980875a4b7d291fa336499ab9735e35e97d562204948b337d40addb5845b8c27ed55b4ac4d7489715318

data/config/geomash.yml

@@ -0,0 +1,27 @@
+development:
+  tgn_enabled: true
+  geonames_username: boston_library
+  #NOTE: Mapquest is super unreliable on its matches...
+  mapquest_key: <mapquest_key>
+  #NOTE: Bing also fairly unreliable but mostly works...
+  bing_key: <bing_key>
+  timeout: 7
+  parser_cache_enabled: false #See Readme before enabling this
+test: &TEST_
+  tgn_enabled: true
+  geonames_username: boston_library
+  #NOTE: Mapquest is super unreliable on its matches...
+  mapquest_key: <mapquest_key>
+  #NOTE: Bing also fairly unreliable but mostly works...
+  bing_key: <bing_key>
+  timeout: 7
+  parser_cache_enabled: false #See Readme before enabling this
+production:
+  tgn_enabled: true
+  geonames_username: boston_library
+  #NOTE: Mapquest is super unreliable on its matches...
+  mapquest_key: <mapquest_key>
+  #NOTE: Bing also fairly unreliable but mostly works...
+  bing_key: <bing_key>
+  timeout: 7
+  parser_cache_enabled: false #See Readme before enabling this

data/config/geomash.yml.sample

@@ -5,6 +5,13 @@ development:
   mapquest_key: <mapquest_key>
   #NOTE: Bing also fairly unreliable but mostly works...
   bing_key: <bing_key>
+  #NOTE: Optional Google API key
+  google_key: <google_key>
+  #NOTE: Optional Google Maps for Work Information
+  google_premier:
+    - key
+    - client
+    - channel
   timeout: 7
   parser_cache_enabled: false #See Readme before enabling this
 test: &TEST_
@@ -14,6 +21,13 @@ test: &TEST_
   mapquest_key: <mapquest_key>
   #NOTE: Bing also fairly unreliable but mostly works...
   bing_key: <bing_key>
+  #NOTE: Optional Google API key
+  google_key: <google_key>
+  #NOTE: Optional Google Maps for Work Information
+  google_premier:
+    - key
+    - client
+    - channel
   timeout: 7
   parser_cache_enabled: false #See Readme before enabling this
 production:
@@ -23,5 +37,12 @@ production:
   mapquest_key: <mapquest_key>
   #NOTE: Bing also fairly unreliable but mostly works...
   bing_key: <bing_key>
+  #NOTE: Optional Google API key
+  google_key: <google_key>
+  #NOTE: Optional Google Maps for Work Information
+  google_premier:
+    - key
+    - client
+    - channel
   timeout: 7
   parser_cache_enabled: false #See Readme before enabling this

data/lib/geomash/geonames.rb

@@ -28,7 +28,16 @@ module Geomash
         parsed_xml = Nokogiri::Slop(geonames_response.body)
 
         parsed_xml.geonames.geoname.each do |geoname|
-           hier_geo[geoname.fcode.text.downcase.to_sym] = geoname.toponymName.text
+          #In some cases, geonames duplicates the fcode keys? See fcode "area" of:
+          #http://api.geonames.org/hierarchy?username=<username>&lang=en&style=FULL&geonameId=6947909
+          #FIXME: Something better needs to be done...
+          if hier_geo.has_key?(geoname.fcode.text.downcase.to_sym)
+            temp_key = geoname.fcode.text.downcase + "2"
+            hier_geo[temp_key.to_sym] = geoname.toponymName.text
+          else
+            hier_geo[geoname.fcode.text.downcase.to_sym] = geoname.toponymName.text
+          end
+
         end
 
         #FIXME: Code4Lib lazy implementation... will get last result
@@ -95,7 +104,7 @@ module Geomash
         elsif geo_hash[:country_part] == 'North Korea'
           country_code = 'KP'
         else
-          country_code = Country.find_country_by_name(geo_hash[:country_part]).alpha2
+          country_code = ISO3166::Country.find_country_by_name(geo_hash[:country_part]).alpha2
         end
         geonames_response = Typhoeus::Request.get("http://api.geonames.org/search?username=#{self.geonames_username}&lang=en&style=FULL&q=#{CGI.escape(geonames_search_string)}&name_equals=#{CGI.escape(exact_name_term)}&country=#{country_code}")
 

data/lib/geomash/parser.rb

@@ -14,6 +14,14 @@ module Geomash
       Geomash.config[:bing_key] || '<bing_key>'
     end
 
+    def self.google_key
+      Geomash.config[:google_key] || '<google_key>'
+    end
+
+    def self.google_premier
+      Geomash.config[:google_premier] || ['key', 'client', 'channel']
+    end
+
     def self.timeout
       Geomash.config[:timeout]
     end
@@ -153,7 +161,7 @@ module Geomash
                                        :combined=>mapquest_api_result.first.data['latLng']['lat'].to_s + ',' + mapquest_api_result.first.data['latLng']['lng'].to_s}
         end
 
-        return_hash[:country_part] = Country.new(mapquest_api_result.first.data["adminArea1"]).name
+        return_hash[:country_part] = ISO3166::Country.new(mapquest_api_result.first.data["adminArea1"]).name
 
         if return_hash[:country_part] == 'United States'
           return_hash[:state_part] = Geomash::Constants::STATE_ABBR[mapquest_api_result.first.data["adminArea3"]] || mapquest_api_result.first.data["adminArea4"]
@@ -191,9 +199,21 @@ module Geomash
         return {}
       end
 
+      #FIXME: Handle just a plain 'Korea' better...
+      if term.match(/Korea/) and !term.match(/South/) and !term.match(/North/)
+        term.gsub!('Korea', 'South Korea')
+      end
+
       return_hash[:standardized_term] = term
 
-      ::Geocoder.configure(:lookup => :google,:api_key => nil,:timeout => self.timeout, :always_raise => :all)
+      if self.google_premier != ['key', 'client', 'channel']
+        ::Geocoder.configure(:lookup => :google_premier,:api_key => self.google_premier,:timeout => self.timeout, :use_https => true, :always_raise => :all)
+      elsif self.google_key != '<google_key>'
+        ::Geocoder.configure(:lookup => :google,:api_key => self.google_key,:timeout => self.timeout, :use_https => true, :always_raise => :all)
+      else
+        ::Geocoder.configure(:lookup => :google,:api_key => nil,:timeout => self.timeout, :always_raise => :all)
+      end
+
       begin
         google_api_result = ::Geocoder.search(term)
       rescue SocketError => e
@@ -217,18 +237,46 @@ module Geomash
         end
       end
 
+
       if google_api_result.present?
+        #Find the best match index... case of Ho Chi Minh City (Vietnam) resolving to an airport...
+        best_match_index = 0
+        best_city_match = -1
+        best_state_match = -1
+
+        google_api_result.each_with_index do |single_google_api_result, index|
+          single_google_api_result.data["address_components"].each do |result|
+            if (result['types'] & ['locality']).present?
+              if return_hash[:standardized_term].to_ascii.include?(result['long_name'].to_ascii)
+                best_city_match = index if best_city_match == -1
+              end
+            elsif (result['types'] & ['administrative_area_level_1']).present?
+              if return_hash[:standardized_term].to_ascii.include?(result['long_name'].to_ascii.gsub('-city', ''))
+                best_state_match = index if best_state_match == -1
+              end
+            end
+          end
+        end
+
+        if best_city_match != -1
+          best_match_index = best_city_match
+        elsif best_state_match != -1
+          best_match_index = best_state_match
+        end
+
+
         #Types: street number, route, neighborhood, establishment, transit_station, bus_station
-        google_api_result.first.data["address_components"].each do |result|
+        google_api_result[best_match_index].data["address_components"].each do |result|
           if (result['types'] & ['street number', 'route', 'establishment', 'transit_station', 'bus_station']).present? || (result['types'].include?('neighborhood') && !result['types'].include?('political'))
             #return_hash[:term_differs_from_tgn] = true
             #TODO: Not implemented for Google results right now.
             #return_hash[:street_part] = 'TODO: Not Implemented for Google Results'
-            return_hash[:coords] = {:latitude=>google_api_result.first.data['geometry']['location']['lat'].to_s,
-                                         :longitude=>google_api_result.first.data['geometry']['location']['lng'].to_s,
-                                         :combined=>google_api_result.first.data['geometry']['location']['lat'].to_s + ',' + google_api_result.first.data['geometry']['location']['lng'].to_s}
+            return_hash[:coords] = {:latitude=>google_api_result[best_match_index].data['geometry']['location']['lat'].to_s,
+                                         :longitude=>google_api_result[best_match_index].data['geometry']['location']['lng'].to_s,
+                                         :combined=>google_api_result[best_match_index].data['geometry']['location']['lat'].to_s + ',' + google_api_result[best_match_index].data['geometry']['location']['lng'].to_s}
           elsif (result['types'] & ['country']).present?
-            return_hash[:country_part] = result['long_name']
+            #gsub to fix a case of "Macedonia" returning "Macedonia (FYROM)"
+            return_hash[:country_part] = result['long_name'].gsub(/ \(.+\)$/, '')
           elsif (result['types'] & ['administrative_area_level_1']).present?
             return_hash[:state_part] = result['long_name'].to_ascii.gsub('-city', '')
           elsif (result['types'] & ['locality']).present?
@@ -238,7 +286,7 @@ module Geomash
           end
         end
 
-        return_hash[:term_differs_from_tgn] ||= google_api_result.first.data['partial_match'] unless google_api_result.first.data['partial_match'].blank?
+        return_hash[:term_differs_from_tgn] ||= google_api_result[best_match_index].data['partial_match'] unless google_api_result[best_match_index].data['partial_match'].blank?
       end
 
       #FIXME: Google free API rate limit is 5 requests / 1 second now (used to be 10). Need a better way to handle this.

data/lib/geomash/standardizer.rb

@@ -20,13 +20,17 @@ module Geomash
       country_name_list = []
 
       #Countries gem of https://github.com/hexorx/countries
-      Country.new('US').states.each do |state_abbr, state_names|
+      ISO3166::Country.new('US').states.each do |state_abbr, state_names|
         state_abbr_list << ' ' + state_abbr
         state_name_list << state_names["name"]
       end
 
-      Country.all.each do |country_name_abbr_pair|
-        country_name_list << country_name_abbr_pair.first
+      ISO3166::Country.all.each do |country_name_hash|
+        #country_name_list << country_name_abbr_pair.first
+        country_name_list << country_name_hash.data["name"] if country_name_hash.data["name"].present?
+        country_name_hash.data["names"].each do |name|
+          country_name_list << name
+        end
       end
       country_name_list.append('South Korea') #Listed as Korea, Republic of in the gem
       country_name_list.append('North Korea') #Listed as Korea, Democratic People's Republic Of of in the gem

data/lib/geomash/tgn.rb

@@ -391,7 +391,7 @@ EXAMPLE SPARQL:
         #Broader places
         #FIXME: could parse xml:lang instead of the three optional clauses now... didn't expect places to lack a default preferred label.
         if broader_place_type_list.present? #Case of World... top of hierachy check
-          query = "SELECT ?identifier_place ?place_label_default ?place_label_en ?place_label_remaining ?aat_pref WHERE {"
+          query = "SELECT ?identifier_place ?place_label_default ?place_label_en ?aat_pref ?place_label_latn_pinyin WHERE {"
 
           broader_place_type_list.each do |place_uri|
             query += %{{<#{place_uri}> <http://purl.org/dc/elements/1.1/identifier> ?identifier_place .
@@ -407,16 +407,13 @@ EXAMPLE SPARQL:
         OPTIONAL {<#{place_uri}> <http://www.w3.org/2004/02/skos/core#altLabel> ?place_label_alt
                  FILTER langMatches( lang(?place_label_alt), "en" )
                  }
-        OPTIONAL {<#{place_uri}> <http://www.w3.org/2004/02/skos/core#prefLabel> ?place_label_remaining
-                 FILTER(!langMatches( lang(?place_label_remaining), "" ) && !langMatches( lang(?place_label_remaining), "en" ) && !langMatches( lang(?place_label_remaining), "zh-latn-pinyin" ))
-                 }
         <#{place_uri}> <http://vocab.getty.edu/ontology#placeTypePreferred> ?aat_pref
        } UNION
      }
           end
 
           query = query[0..-12]
-          query += ". } GROUP BY ?identifier_place ?place_label_default ?place_label_en ?place_label_latn_pinyin ?place_label_alt ?place_label_remaining ?aat_pref"
+          query += ". } GROUP BY ?identifier_place ?place_label_default ?place_label_en ?place_label_latn_pinyin ?place_label_alt ?aat_pref"
           query = query.squish
 
           tgn_response_for_aat = Typhoeus::Request.post("http://vocab.getty.edu/sparql.json", :body=>{:query=>query}, :timeout=>500)
@@ -431,41 +428,43 @@ EXAMPLE SPARQL:
               tgn_term = aat_response['place_label_default']['value']
             elsif aat_response['place_label_latn_pinyin'].present? && aat_response['place_label_latn_pinyin']['value'] != '-'
               tgn_term = aat_response['place_label_latn_pinyin']['value']
+            elsif aat_response['place_label_latn_notone'].present? && aat_response['place_label_latn_notone']['value'] != '-'
+              tgn_term = aat_response['place_label_latn_notone']['value']
             elsif aat_response['place_label_alt'].present? && aat_response['place_label_alt']['value'] != '-'
               tgn_term = aat_response['place_label_alt']['value']
             else
-              tgn_term = aat_response['place_label_remaining']['value']
+              raise "Could not find a label for: #{tgn_id}"
             end
 
             case tgn_term_type
               when '300128176' #continent
-                hier_geo[:continent] = tgn_term
+                hier_geo[:continent] ||= tgn_term
               when '300128207', '300387130', '300387506' #nation, autonomous areas, countries
-                hier_geo[:country] = tgn_term
+                hier_geo[:country] ||= tgn_term
               when '300000774' #province
-                hier_geo[:province] = tgn_term
+                hier_geo[:province] ||= tgn_term
               when '300236112', '300182722', '300387194', '300387052', '300387113', '300387107' #region, union, semi-independent political entity, autonomous communities, autonomous regions
-                hier_geo[:region] = tgn_term
+                hier_geo[:region] ||= tgn_term
               when '300000776', '300000772', '300235093' #state, department, governorate
-                hier_geo[:state] = tgn_term
+                hier_geo[:state] ||= tgn_term
               when '300387081' #national district
                 if tgn_term == 'District of Columbia'
-                  hier_geo[:state] = tgn_term
+                  hier_geo[:state] ||= tgn_term
                 else
-                  hier_geo[:territory] = tgn_term
+                  hier_geo[:territory] ||= tgn_term
                 end
               when '300135982', '300387176', '300387122' #territory, dependent state, union territory
-                hier_geo[:territory] = tgn_term
+                hier_geo[:territory] ||= tgn_term
               when '300000771', '300387092', '300387071' #county, parishes, unitary authorities
-                hier_geo[:county] = tgn_term
+                hier_geo[:county] ||= tgn_term
               when '300008347', '300008389' #inhabited place, cities
-                hier_geo[:city] = tgn_term
+                hier_geo[:city] ||= tgn_term
               when '300000745', '300000778', '300387331' #neighborhood, parishes, parts of inhabited places
-                hier_geo[:city_section] = tgn_term
+                hier_geo[:city_section] ||= tgn_term
               when '300008791', '300387062' #island
-                hier_geo[:island] = tgn_term
+                hier_geo[:island] ||= tgn_term
               when '300387575', '300387346', '300167671', '300387178', '300387082', '300387173', '300055621', '300386853', '300386831', '300386832', '300008178', '300008804', '300387131', '300132348', '300387085', '300387198', '300008761'   #'81101/area', '22101/general region', '83210/deserted settlement', '81501/historical region', '81126/national division', administrative divisions, area (measurement), island groups, mountain ranges, mountain systems, nature reserves, peninsulas, regional divisions, sand bars, senatorial districts (administrative districts), third level subdivisions (political entities), valleys (landforms)
-                hier_geo[:area] = tgn_term
+                hier_geo[:area] ||= tgn_term
             end
           end
         end
@@ -500,6 +499,10 @@ EXAMPLE SPARQL:
       neighboorhood_response = {}
 
       state_part = geo_hash[:state_part]
+      #FIXME: In TGN, Ho Chi Minh doesn't have an ASCII label... unsure what to do in this case... maybe a synonyms file?
+      if state_part == 'Ho Chi Minh'
+        state_part = 'Hồ Chí Minh'
+      end
 
       country_code = Geomash::Constants::COUNTRY_TGN_LOOKUP[geo_hash[:country_part]][:tgn_id] unless Geomash::Constants::COUNTRY_TGN_LOOKUP[geo_hash[:country_part]].blank?
       country_code ||= ''
@@ -582,6 +585,8 @@ WHERE
   {?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300387122>} UNION
   {?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300000776>} UNION
   {?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300236112>} UNION
+  {?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300387506>} UNION
+  {?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300265612>} UNION
   {?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300387081>} .
   ?x <http://www.w3.org/2000/01/rdf-schema#label> ?object_label .
   FILTER regex(?object_label, "^#{state_part}$", "i" )
@@ -601,20 +606,39 @@ GROUP BY ?object_identifier
           web_request_error = true if states_response[:errors]
         end
 
+        #Do prefLabel first and then do just label... needed for case of Newton vs Newtown in MA (Newtown has an altlabel of Newton)
         if states_response[:id].present? && city_part.present? && !web_request_error
           query = %{SELECT ?object_identifier
 WHERE
 {
   ?x <http://purl.org/dc/elements/1.1/identifier> ?object_identifier .
   ?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300008347> .
-  ?x <http://www.w3.org/2000/01/rdf-schema#label> ?object_label .
+  ?x <http://www.w3.org/2004/02/skos/core#prefLabel> ?object_label .
   FILTER regex(?object_label, "^#{city_part}$", "i" )
   ?x <http://vocab.getty.edu/ontology#broaderPreferredExtended> <http://vocab.getty.edu/tgn/#{country_response[:id]}> .
   ?x <http://vocab.getty.edu/ontology#broaderPreferredExtended> <http://vocab.getty.edu/tgn/#{states_response[:id]}> .
 }
 GROUP BY ?object_identifier
 }
+
           cities_response = self.tgn_sparql_request(query)
+          if cities_response[:id].blank? && !cities_response[:errors]
+            query = %{SELECT ?object_identifier
+WHERE
+{
+  ?x <http://purl.org/dc/elements/1.1/identifier> ?object_identifier .
+  ?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300008347> .
+  ?x <http://www.w3.org/2000/01/rdf-schema#label> ?object_label .
+  FILTER regex(?object_label, "^#{city_part}$", "i" )
+  ?x <http://vocab.getty.edu/ontology#broaderPreferredExtended> <http://vocab.getty.edu/tgn/#{country_response[:id]}> .
+  ?x <http://vocab.getty.edu/ontology#broaderPreferredExtended> <http://vocab.getty.edu/tgn/#{states_response[:id]}> .
+}
+GROUP BY ?object_identifier
+}
+            cities_response = self.tgn_sparql_request(query)
+          end
+
+
           if cities_response[:id].blank? && !cities_response[:errors]
             return_hash[:original_string_differs] = true
           else
@@ -666,6 +690,26 @@ WHERE
 GROUP BY ?object_identifier
 }
         neighborhood_response = self.tgn_sparql_request(query)
+
+        #Try once more on just prefLabel with no city restriction and inhabited places type added...
+        if neighborhood_response[:id].blank? && !neighborhood_response[:errors]
+          query = %{SELECT ?object_identifier
+WHERE
+{
+  ?x <http://purl.org/dc/elements/1.1/identifier> ?object_identifier .
+  {?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300000745>} UNION
+  {?x <http://vocab.getty.edu/ontology#placeTypePreferred> <http://vocab.getty.edu/aat/300008347>} .
+  ?x <http://www.w3.org/2004/02/skos/core#prefLabel> ?object_label .
+  FILTER regex(?object_label, "^#{neighborhood_part}$", "i" )
+  ?x <http://vocab.getty.edu/ontology#broaderPreferredExtended> <http://vocab.getty.edu/tgn/#{country_response[:id]}> .
+  ?x <http://vocab.getty.edu/ontology#broaderPreferredExtended> <http://vocab.getty.edu/tgn/#{states_response[:id]}> .
+}
+GROUP BY ?object_identifier
+}
+          neighborhood_response = self.tgn_sparql_request(query)
+        end
+
+
         if neighborhood_response[:id].blank? && !neighborhood_response[:errors]
           return_hash[:original_string_differs]=true
         else

data/lib/geomash/version.rb

@@ -1,3 +1,3 @@
 module Geomash
-  VERSION = "0.3.1"
+  VERSION = "0.3.2"
 end

data/test/geomash_test.rb

@@ -161,6 +161,8 @@ class GeomashTest < ActiveSupport::TestCase
     assert_equal true, result[:tgn][:original_string_differs] if Geomash::TGN.tgn_enabled == true
 
     #Case of a country with no states
+    #FIXME: This now returns a state when it never did before...
+=begin
     result = Geomash.parse('Tokyo, Japan')
     assert_equal 'Tokyo', result[:city_part]
     assert_equal nil, result[:state_part]
@@ -170,6 +172,7 @@ class GeomashTest < ActiveSupport::TestCase
     assert_equal false, result[:tgn][:original_string_differs] if Geomash::TGN.tgn_enabled == true
     assert_equal '1850147', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
     assert_equal false, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
+=end
 
     #Should find the Michigan Atlanta over the Georgia Atlanta
     #State part from an API giving me Atlanta????
@@ -194,6 +197,44 @@ class GeomashTest < ActiveSupport::TestCase
     assert_equal '1835841', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
     assert_equal false, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
 
+    result = Geomash.parse('Northern Ireland')
+    assert_equal nil, result[:city_part]
+    assert_equal 'Northern Ireland', result[:state_part]
+    assert_equal 'United Kingdom', result[:country_part]
+    assert_equal nil, result[:neighborhood_part]
+    assert_equal '7002448', result[:tgn][:id] if Geomash::TGN.tgn_enabled == true
+    assert_equal false, result[:tgn][:original_string_differs] if Geomash::TGN.tgn_enabled == true
+    assert_equal '2641364', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
+    assert_equal false, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
+
+    result = Geomash.parse('Phnom Penh (Cambodia)')
+    assert_equal '7004076', result[:tgn][:id] if Geomash::TGN.tgn_enabled == true
+    assert_equal false, result[:tgn][:original_string_differs] if Geomash::TGN.tgn_enabled == true
+    assert_equal '1821306', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
+    assert_equal false, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
+
+
+    result = Geomash.parse('Ho Chi Minh City (Vietnam)')
+    assert_equal '7001069', result[:tgn][:id] if Geomash::TGN.tgn_enabled == true
+    assert_equal false, result[:tgn][:original_string_differs] if Geomash::TGN.tgn_enabled == true
+    assert_equal '1566083', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
+    assert_equal false, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
+
+    #Ensure we get "Newton" instead of "Newtown" that has an altlabel of "Newton"
+    #Should this find Chestnut hill...?
+    result = Geomash.parse('Chestnut Hill, Massachusetts')
+    assert_equal 'Newton', result[:city_part]
+    assert_equal 'Massachusetts', result[:state_part]
+    assert_equal 'United States', result[:country_part]
+    assert_equal 'Chestnut Hill', result[:neighborhood_part]
+    assert_equal '7032056', result[:tgn][:id] if Geomash::TGN.tgn_enabled == true #2050214 or
+    assert_equal false, result[:tgn][:original_string_differs] if Geomash::TGN.tgn_enabled == true
+    assert_equal '4932957', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
+    assert_equal false, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
+
+
+
+
 
 
   end

data/test/parser_test.rb

@@ -20,6 +20,13 @@ class ParserTest < ActiveSupport::TestCase
     assert_equal 'Roxbury', result[:neighborhood_part]
     assert_equal nil, result[:street_part]
 
+    result = Geomash::Parser.parse_google_api('Macedonia')
+    assert_equal nil, result[:city_part]
+    assert_equal nil, result[:state_part]
+    assert_equal 'Macedonia', result[:country_part]
+    assert_equal nil, result[:neighborhood_part]
+    assert_equal nil, result[:street_part]
+
 
     #FIXME!!!  Is this alright?
     #result = Bplgeo::Parser.parse_google_api('201 Dowman Dr., Atlanta, GA 30322')

data/test/tgn_test.rb

@@ -14,6 +14,18 @@ class TGNTest < ActiveSupport::TestCase
        assert_equal 'United States', result[:hier_geo][:country]
        assert_equal 'North and Central America', result[:hier_geo][:continent]
 
+
+       #Check for a weird prefLabel case of only zh-latn-pinyin-x-notone
+       result = Geomash::TGN.get_tgn_data('7002066')
+       assert_equal '45.75', result[:coords][:latitude]
+       assert_equal '126.65', result[:coords][:longitude]
+       assert_equal '45.75,126.65', result[:coords][:combined]
+       assert_equal 'Harbin', result[:hier_geo][:city]
+       assert_equal 'Heilongjiang', result[:hier_geo][:province]
+       assert_equal 'China', result[:hier_geo][:country]
+       assert_equal 'Asia', result[:hier_geo][:continent]
+       assert_equal 'Harbin', result[:non_hier_geo][:value]
+       assert_nil result[:non_hier_geo][:qualifier]
     end
   end
 end

metadata

@@ -1,153 +1,153 @@
 --- !ruby/object:Gem::Specification
 name: geomash
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.3.2
 platform: ruby
 authors:
 - Boston Public Library
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-05-29 00:00:00.000000000 Z
+date: 2016-04-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: countries
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - '='
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 1.2.2
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - '='
       - !ruby/object:Gem::Version
-        version: '0'
+        version: 1.2.2
 - !ruby/object:Gem::Dependency
   name: geocoder
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: unidecoder
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: typhoeus
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: nokogiri
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: htmlentities
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: sparql
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: sqlite3
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: rails
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 description: Parse string for potential geographic matches and return that data along
@@ -158,25 +158,26 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- Rakefile
+- config/geomash.yml
 - config/geomash.yml.sample
 - lib/geomash.rb
-- lib/geomash/tgn.rb
 - lib/geomash/autoexpire_cache_dalli.rb
-- lib/geomash/town_lookup.rb
-- lib/geomash/geonames.rb
-- lib/geomash/constants.rb
-- lib/geomash/version.rb
-- lib/geomash/standardizer.rb
 - lib/geomash/autoexpire_cache_redis.rb
+- lib/geomash/constants.rb
+- lib/geomash/geonames.rb
 - lib/geomash/parser.rb
-- Rakefile
-- test/standardizer_test.rb
+- lib/geomash/standardizer.rb
+- lib/geomash/tgn.rb
+- lib/geomash/town_lookup.rb
+- lib/geomash/version.rb
 - test/geomash_test.rb
+- test/geonames_test.rb
+- test/parser_test.rb
+- test/standardizer_test.rb
 - test/test_helper.rb
 - test/tgn_test.rb
 - test/town_lookup_test.rb
-- test/parser_test.rb
-- test/geonames_test.rb
 homepage: http://www.bpl.org
 licenses: []
 metadata: {}
@@ -186,26 +187,26 @@ require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.0.3
+rubygems_version: 2.4.6
 signing_key: 
 specification_version: 4
 summary: Parse string for potential geographic matches and return that data along
   with the TGN ID and Geonames ID.
 test_files:
+- test/parser_test.rb
+- test/geonames_test.rb
 - test/standardizer_test.rb
-- test/geomash_test.rb
-- test/test_helper.rb
 - test/tgn_test.rb
 - test/town_lookup_test.rb
-- test/parser_test.rb
-- test/geonames_test.rb
+- test/test_helper.rb
+- test/geomash_test.rb