geomash 0.3.8 → 0.3.9

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 5b352aa88b7f74448b67dc9aa8bc1e0886483b48
-  data.tar.gz: 6b217b344779f407b08906f1bb250fcd35f086db
+  metadata.gz: 0e61c5deca72356d721f4e23a4a85e0f3c0cb3aa
+  data.tar.gz: dfbf8a130fc2074622ad784a6f9449a6d7f6160c
 SHA512:
-  metadata.gz: 58b2714da28ec388c7b07e82ea2c51c40b1b2d9fbded27f54bd8f0521cc6561cd3e5115cbc4b370275a51692a18020c9a5a7f78fce04c796e69d3899d5b1513c
-  data.tar.gz: 881c889566bce33fdcba5109b9efef976f75409c9349ccdef41bc6fc09b0551d2bb6f33d8aa9bb7dd100a6789e95fdc90ab944e31a6d3cf7e2c7ad16c0ec6bdd
+  metadata.gz: b3ffdd653f2b124ff2c8ceea4c7b59300762f2b2b721673a57d9ccd99eaaffe45f2cdf123b7016882a0a9dc4bea8e4e6d450dfa8aebaa9612aa2bfdb8b8cca20
+  data.tar.gz: 6b041c033c95a04a55683acb016f6416ac54129d18e69888c10c250b0cfc7d621789d08c4f36db54404f2a25db1e00086fb13ae4c3ab2cce971504bc61209eac

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(' ,', ', ').gsub(' >', ',')
+          geo_term = term.gsub('(', ',').gsub(' ,', ', ').gsub(' >', ',').gsub(' >', ',')
           geo_term = geo_term.gsub(')', '')
           #end
 
@@ -96,18 +96,18 @@ module Geomash
       #Replace any four TGN dashes from removing a junk term
       geo_term = geo_term.gsub('----', '--')
 
-      #Replace any semicolons with commas... possible strip them?
-      geo_term = geo_term.gsub(';', ',')
-
       #Replace > with commas
       geo_term = geo_term.gsub('>', ',').gsub('>', ',')
 
+      #Replace any semicolons with commas... possible strip them?
+      geo_term = geo_term.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)
         if geo_term.match(/ \(+.*\)+/)
           #Make this replacement better?
-          geo_term = geo_term.gsub(/ *\((?=[\S ]+\))/,', ')
+          geo_term = geo_term.gsub(/ *\((?=[\S ]*\))/,', ')
           geo_term = geo_term.gsub(')', '')
 
           #Else skip this as data returned likely will be unreliable for now... FIXME when use case occurs.

data/lib/geomash/version.rb

@@ -1,3 +1,3 @@
 module Geomash
-  VERSION = "0.3.8"
+  VERSION = "0.3.9"
 end

data/test/geomash_test.rb

@@ -21,6 +21,17 @@ class GeomashTest < ActiveSupport::TestCase
     assert_equal '4933002', 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('Massachusetts > Hampden (county) > 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 '2049596', 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 '4933002', result[:geonames][:id] if Geomash::Geonames.geonames_username != '<username>'
+    assert_equal false, result[:geonames][:original_string_differs] if Geomash::Geonames.geonames_username != '<username>'
+
     #Slight variation problem with neighborhood: 11. Bezirk (Vienna, Austria)--Biography
     result = Geomash.parse('15. Bezirk (Rudolfsheim-Fünfhaus, Vienna, Austria)--Exhibitions', true)
     assert_equal 'Vienna', result[:city_part]

data/test/standardizer_test.rb

@@ -57,6 +57,20 @@ class ParserTest < ActiveSupport::TestCase
 
     result = Geomash::Standardizer.parse_for_geographic_term('Blah (North Korea)')
     assert_equal 'Blah, North Korea', result
+
+    result = Geomash::Standardizer.parse_for_geographic_term('Massachusetts > Hampden (county) > Chicopee')
+    assert_equal 'Massachusetts, Hampden, county, Chicopee', result
+
+    result = Geomash::Standardizer.parse_for_geographic_term('Massachusetts &gt; Hampden (county) &gt; Chicopee')
+    assert_equal 'Massachusetts, Hampden, county, Chicopee', result
+  end
+
+  def test_geographic_standardizer
+    result = Geomash::Standardizer.standardize_geographic_term('Massachusetts > Hampden (county) > Chicopee')
+    assert_equal 'Massachusetts, Hampden, Chicopee', result
+
+    result = Geomash::Standardizer.standardize_geographic_term('Massachusetts &gt; Hampden (county) &gt; Chicopee')
+    assert_equal 'Massachusetts, Hampden, Chicopee', result
   end
 
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: geomash
 version: !ruby/object:Gem::Version
-  version: 0.3.8
+  version: 0.3.9
 platform: ruby
 authors:
 - Boston Public Library
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-10-25 00:00:00.000000000 Z
+date: 2016-10-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activesupport