geo-spider 0.1.3 → 0.2.0

data/lib/geo-spider/extractors/postcode.rb

@@ -11,14 +11,14 @@ module GeoSpider
       REGEXP = /(GIR 0AA|[A-PR-UWYZ]([0-9]{1,2}|([A-HK-Y][0-9]|[A-HK-Y][0-9]([0-9]|[ABEHMNPRV-Y]))|[0-9][A-HJKS-UW])(\s*)[0-9][ABD-HJLNP-UW-Z]{2})/i
       
       def locations
-        results = @element.inner_html.scan(REGEXP)
+        results = @element.inner_text.scan(REGEXP)
         results = results.map(&:first)
         
         found_locations = []
         
         results.each do |result|
           begin
-            p = geocoder.locate(result)
+            p = geocoder.locate(:zip => result, :country => "GB")
             found_locations << Location.new(:latitude => p.latitude, :longitude => p.longitude, :title => result)
           rescue Graticule::Error => e
             next
@@ -27,7 +27,7 @@ module GeoSpider
         return found_locations
       end
       
-      # You need to set a valid Yahoo API key before the UK postcode geocoding will work. Yahoo have vastly better UK postcode accuracy than the other large mapping providers, apart from perhaps Multimap.
+      # You need to set a valid Multimap API key before the UK postcode geocoding will work. Multimap have the most accurate UK postcode geocoding I've discovered.
       
       def self.api_key=(api_key)
         @@api_key = api_key
@@ -36,8 +36,8 @@ module GeoSpider
       private
       
       def geocoder
-        raise "No Yahoo API key set" unless @@api_key
-        Graticule.service(:yahoo).new @@api_key
+        raise "No Multimap API key set" unless @@api_key
+        Graticule.service(:multimap).new @@api_key
       end
       
     end

data/lib/geo-spider/version.rb

@@ -1,8 +1,8 @@
 module GeoSpider
   module VERSION #:nodoc:
     MAJOR = 0
-    MINOR = 1
-    TINY  = 3
+    MINOR = 2
+    TINY  = 0
 
     STRING = [MAJOR, MINOR, TINY].join('.')
   end

data/spec/geo-spider/page_spec.rb

@@ -25,9 +25,9 @@ describe Page, "with a single postcode which is being parsed" do
     OpenURI.should_receive(:open_uri).and_return(page_as_string('single_postcode.html'))
     @page = Page.new("http://www.example.com")
     GeoSpider::Extractors::Postcode.api_key = "waffles"
-    mock_geocoder_result = OpenStruct.new( {:location => [51.000000, -1.000000]} )
+    mock_geocoder = OpenStruct.new( { :locate => OpenStruct.new({ :longitude => -1.000000, :latitude => 51.000000 })})
     Graticule.stub!(:service)
-    Graticule.service.should_receive(:new).and_return(mock_geocoder_result)
+    Graticule.service.should_receive(:new).and_return(mock_geocoder)
   end
   
   it "should find one location" do
@@ -49,9 +49,9 @@ describe Page, "with multiple microformats and postcodes being parsed" do
     OpenURI.should_receive(:open_uri).and_return(page_as_string('multiple_postcodes_and_microformats.html'))
     @page = Page.new("http://www.example.com")
     
-    mock_geocoder_result = OpenStruct.new( {:location => [51.000000, -1.000000]} )
+    mock_geocoder = OpenStruct.new( { :locate => OpenStruct.new({ :longitude => -1.000000, :latitude => 51.000000 })})
     Graticule.stub!(:service)
-    Graticule.service.should_receive(:new).twice.and_return(mock_geocoder_result)
+    Graticule.service.should_receive(:new).twice.and_return(mock_geocoder)
   end
   
   it "should find four locations" do
@@ -122,4 +122,17 @@ describe Page, "which is finding the title" do
       @page.title.should == "Heading 1"
     end
   end
+end
+
+describe Page, "which is parsing a page with a string in a URL that happens to match a postcode" do
+  
+  before(:each) do
+    OpenURI.should_receive(:open_uri).and_return(page_as_string('postcode_in_url.html'))
+    @page = Page.new("http://www.example.com")
+  end
+  
+  it "should not find any locations" do
+    @page.locations.should be_empty
+  end
+  
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: geo-spider
 version: !ruby/object:Gem::Version 
-  version: 0.1.3
+  version: 0.2.0
 platform: ruby
 authors: 
 - Tom Taylor
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2008-10-01 00:00:00 +01:00
+date: 2008-10-03 00:00:00 +01:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency