geomash 0.3.6 → 0.3.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 83306d315da162bf4ed9e305effda2ec568aaeb9
-  data.tar.gz: 68d5865eca8797f33c1ef634c3abf34defa1c461
+  metadata.gz: f51ecd1752538ca69b4f80d4069519842e9f78c4
+  data.tar.gz: f3f8a263990c5441b76532fc3d39f282b25c91c5
 SHA512:
-  metadata.gz: 709129ca2ac3efa45e0a20c3df051d8459d87d2e95a7248f41e329f26023c906a17ef7fc17915924953e1087f582062bf145806a0d45334aef7e921d1ab0e03a
-  data.tar.gz: 57fbd79e91d2d3bded69d4801e298151697069c85e977a5c9bcfc81859c2c9a32df890d478e78c47141f8c7760a1648d95b43af68f469bea215557042625fbcd
+  metadata.gz: f98af5d1b09972e6b16cc5b5beb0360cf965c2eb58494aaa075c565672ee294e909799bf22c46c0cfa24580e0838b75fae973c4c9845f655992e77e2e60e5969
+  data.tar.gz: 439939cef31ad2058efeb8546de0f20fbdfc2457484d23dda80c150ca42a7d8e36b835e61e917e54ce3276d44114dc345b1fcd396952a46f5eb2b6eb333fc4d6

data/config/geomash.yml

@@ -5,7 +5,7 @@ development:
   mapquest_key: <mapquest_key>
   #NOTE: Bing also fairly unreliable but mostly works...
   bing_key: <bing_key>
-  timeout: 7
+  timeout: 15
   parser_cache_enabled: false #See Readme before enabling this
 test: &TEST_
   tgn_enabled: true
@@ -14,7 +14,7 @@ test: &TEST_
   mapquest_key: <mapquest_key>
   #NOTE: Bing also fairly unreliable but mostly works...
   bing_key: <bing_key>
-  timeout: 7
+  timeout: 15
   parser_cache_enabled: false #See Readme before enabling this
 production:
   tgn_enabled: true
@@ -23,5 +23,5 @@ production:
   mapquest_key: <mapquest_key>
   #NOTE: Bing also fairly unreliable but mostly works...
   bing_key: <bing_key>
-  timeout: 7
+  timeout: 15
   parser_cache_enabled: false #See Readme before enabling this

data/lib/geomash/constants.rb

@@ -84,6 +84,7 @@ module Geomash
         '(Republic)',
         'Directories',
         'Biography',
+        'county',
         #Some date removal
         /[\d]+th [cC]entury,/
     ]

data/lib/geomash/parser.rb

@@ -282,16 +282,18 @@ module Geomash
             return_hash[:state_part] = result['long_name'].to_ascii.gsub('-city', '')
           elsif (result['types'] & ['locality']).present?
             return_hash[:city_part] = result['long_name']
-          elsif (result['types'] & ['sublocality', 'political']).length == 2 || result['types'].include?('neighborhood')
-              return_hash[:neighborhood_part] = result['long_name']
+          elsif (result['types'] & ['sublocality', 'political']).length == 2
+              return_hash[:neighborhood_part] ||= result['long_name'] #See term 'Roxbury (Boston, Mass.)' for why neighborhood should take precedence
+          elsif result['types'].include?('neighborhood')
+            return_hash[:neighborhood_part] = result['long_name']
           end
         end
 
         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
 
-      #This changed in Google... need a better way to handle this
-      if return_hash[:state_part] == 'Nord-Pas-de-Calais Picardie'
+      #This changed in Google... twice now actually... need a better way to handle this
+      if return_hash[:state_part] == 'Nord-Pas-de-Calais Picardie' || return_hash[:state_part] == 'Nord-Pas-de-Calais-Picardie'
         return_hash[:state_part] = 'Picardy'
       end
 

data/lib/geomash/standardizer.rb

@@ -11,7 +11,7 @@ module Geomash
         return ''
       end
 
-      term_split_list = term.split(/[,\-\(\(]/).reject{ |e| e.empty? }
+      term_split_list = term.split(/[,\-\(\(]|>/).reject{ |e| e.empty? }
       term_split_list.each{ |e| e.gsub!(/[^\w\s]/, "") } #Remove punctuation
       term_split_list.each{ |e| e.strip! } #Remove any extra remaining whitespace
       term_split_list.reject{ |e| e.empty? }
@@ -72,7 +72,7 @@ module Geomash
           end
         else
           #if term_split_list.length > 1
-          geo_term = term.gsub('(', ',').gsub(' ,', ', ')
+          geo_term = term.gsub('(', ',').gsub(' ,', ', ').gsub(' >', ',')
           geo_term = geo_term.gsub(')', '')
           #end
 
@@ -99,6 +99,9 @@ module Geomash
       #Replace any semicolons with commas... possible strip them?
       geo_term = geo_term.gsub(';', ',')
 
+      #Replace > with commas
+      geo_term = geo_term.gsub('>', ',').gsub('>', ',')
+
       #Terms in paranthesis will cause some geographic parsers to freak out. Switch to commas instead.
       if geo_term.match(/[\(\)]+/)
         #Attempt to fix address if something like (word)

data/lib/geomash/tgn.rb

@@ -375,12 +375,8 @@ EXAMPLE SPARQL:
             #Fix cases like http://vocab.getty.edu/aat/300132316 which are bays (bodies of water)
             aat_term = aat_term.gsub(/ \(.+\)$/, '')
 
-            if (aat_term =~ /ies$/).present?
-              aat_term = aat_term.gsub(/ies$/, 'y')
-            elsif (aat_term =~ /es$/).present?
-              aat_term = aat_term.gsub(/es$/, '')
-            elsif (aat_term =~ /s$/).present?
-              aat_term = aat_term.gsub(/s$/, '')
+            if (aat_term =~ /ies$/).present? || (aat_term =~ /es$/).present? || (aat_term =~ /s$/).present?
+              aat_term = aat_term.singularize
             end
 
             #Fix cases like "Boston Harbor" as "Boston Harbor (harbor)" isn't that helpful

data/lib/geomash/version.rb

@@ -1,3 +1,3 @@
 module Geomash
-  VERSION = "0.3.6"
+  VERSION = "0.3.7"
 end

data/test/geomash_test.rb

@@ -10,15 +10,16 @@ require 'test_helper'
 class GeomashTest < ActiveSupport::TestCase
 
   def test_parse_with_flag
-    result = Geomash.parse('Abbeville (France)--History--20th century.', true)
-    assert_equal 'Abbeville', result[:city_part]
-    assert_equal 'Picardy', result[:state_part]
-    assert_equal 'France', result[:country_part]
+    result = Geomash.parse('Massachusetts &gt; Hampden (county) &gt; Chicopee', true)
+    assert_equal 'Chicopee', result[:city_part]
+    assert_equal 'Massachusetts', result[:state_part]
+    assert_equal 'United States', result[:country_part]
+    assert_equal nil, result[:neighborhood_part]
     assert_equal nil, result[:street_part]
-    assert_equal '7010587', 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 '2987374', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
-    #assert_equal true, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
+    assert_equal '2049596', result[:tgn][:id] if Geomash::TGN.tgn_enabled == true
+    assert_equal true, result[:tgn][:original_string_differs] if Geomash::TGN.tgn_enabled == true #This should be false?
+    assert_equal '4933002', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
+    assert_equal true, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>' #This should be false?
 
     #Slight variation problem with neighborhood: 11. Bezirk (Vienna, Austria)--Biography
     result = Geomash.parse('15. Bezirk (Rudolfsheim-Fünfhaus, Vienna, Austria)--Exhibitions', true)
@@ -65,16 +66,15 @@ class GeomashTest < ActiveSupport::TestCase
     assert_equal '6252001', 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('Lettering--United States--History--19th century', true)
-    assert_equal nil, result[:city_part]
-    assert_equal nil, result[:state_part]
-    assert_equal 'United States', result[:country_part]
-    assert_equal nil, result[:neighborhood_part]
+    result = Geomash.parse('Abbeville (France)--History--20th century.', true)
+    assert_equal 'Abbeville', result[:city_part]
+    assert_equal 'Picardy', result[:state_part]
+    assert_equal 'France', result[:country_part]
     assert_equal nil, result[:street_part]
-    assert_equal '7012149', result[:tgn][:id] if Geomash::TGN.tgn_enabled == true
+    assert_equal '7010587', 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 '6252001', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
-    assert_equal false, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
+      #assert_equal '2987374', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
+      #assert_equal true, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
 
 
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: geomash
 version: !ruby/object:Gem::Version
-  version: 0.3.6
+  version: 0.3.7
 platform: ruby
 authors:
 - Boston Public Library
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-06-16 00:00:00.000000000 Z
+date: 2016-10-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport