feed_me 0.6.2 → 0.7.0

data/lib/feed_me/abstract_parser.rb

@@ -68,7 +68,7 @@ private
     return nil unless xml
     association = self.class.properties[name]
     
-    nodes = xml.xpath("./#{association[:path]}")
+    nodes = xml.xpath(association[:path].to_s)
     parser = FeedMe.const_get(association[:use].to_s)
     
     nodes.map do |node|
@@ -80,7 +80,7 @@ private
     return nil unless xml
     association = self.class.properties[name]
     
-    node = xml.xpath("./#{association[:path]}").first
+    node = xml.xpath(association[:path].to_s).first
     parser = FeedMe.const_get(association[:use].to_s)
     
     parser.new(self, node)
@@ -90,11 +90,11 @@ private
     return nil unless xml
     property = self.class.properties[name]
 
-    values = xml.xpath("./#{property[:path]}").map do |node|
+    values = xml.xpath(property[:path].to_s).map do |node|
       result = extract_result(node, property[:from])
       cast_result(result, property[:as])
     end
-    
+
     case property[:cardinality]
     when :many
       return values
@@ -113,9 +113,9 @@ private
 
   def cast_result(result, as)
     if as == :time
-      Time.parse(result)
+      Time.parse(result).utc
     else
-      result
+      result.to_s.strip
     end
   end
 

data/lib/feed_me/feed_parser.rb

@@ -26,12 +26,15 @@ module FeedMe
       # then returns a parser object
       def parse(feed)
         document = Nokogiri::XML(feed)
-        if root = document.root 
+        if root = document.root
           root.add_namespace_definition('atom', 'http://www.w3.org/2005/Atom')
+          root.add_namespace_definition('rdf', 'http://www.w3.org/1999/02/22-rdf-syntax-ns#')
+          root.add_namespace_definition('rss1', 'http://purl.org/rss/1.0/')
+          root.add_namespace_definition('dc', 'http://purl.org/dc/elements/1.1/')
           parsers.each do |parser|
             node = root.xpath(parser.root_node).first
             if node
-              return parser.new(node, document.encoding)
+              return parser.new(node)
             end
           end
         end
@@ -41,13 +44,10 @@ module FeedMe
 
     end # class << self
 
-    def initialize(xml, encoding)
+    def initialize(xml)
       @xml = xml
-      @encoding = encoding
     end
 
-    attr_reader :encoding
-
   end
 
   class AtomFeedParser < FeedParser
@@ -81,4 +81,21 @@ module FeedMe
 
     has_one :author, :path => 'managingEditor', :use => :Rss2PersonParser
   end
+
+  class Rss1FeedParser < FeedParser
+    root_node "//rdf:RDF"
+
+    property :title, :path => 'rss1:channel/rss1:title'
+    property :description, :path => 'rss1:channel/rss1:description'
+    property :feed_id, :path => :undefined
+    property :updated_at, :path => 'rss1:channel/dc:date', :as => :time
+    property :url, :path => 'rss1:channel/rss1:link'
+    property :href, :path => 'rss1:channel/@rdf:about'
+    property :generator, :path => :undefined
+
+    has_many :entries, :path => 'rss1:channel/rss1:items/rdf:Seq/rdf:li', :use => :Rss1ItemParser
+
+    has_one :author, :path => 'rss1:channel', :use => :Rss1PersonParser
+
+  end
 end

data/lib/feed_me/item_parser.rb

@@ -22,4 +22,19 @@ module FeedMe
 
     has_one :author, :use => :Rss2PersonParser
   end
+
+  class Rss1ItemParser < ItemParser
+    property :title, :path => 'rss1:title'
+    property :updated_at, :path => 'dc:date', :as => :time
+    property :item_id, :path => '@rdf:about'
+    property :url, :path => 'rss1:link'
+    property :content, :path => 'rss1:description'
+
+    has_one :author, :path => '.', :use => :Rss1PersonParser
+
+    def xml
+      resource = super.xpath('@rdf:resource').to_s
+      super.xpath("//rss1:item[@rdf:about='#{resource}']").first
+    end
+  end
 end

data/lib/feed_me/person_parser.rb

@@ -20,4 +20,10 @@ module FeedMe
     end
   end
 
+  class Rss1PersonParser < PersonParser
+    property :email, :path => :undefined
+    property :name, :path => 'dc:creator'
+    property :uri, :path => :undefined
+  end
+
 end

data/spec/feed_parser_spec.rb

@@ -31,6 +31,7 @@ describe FeedMe::FeedParser do
   before :each do
     @atom = FeedMe::FeedParser.parse(open(fixture('welformed.atom')).read)
     @rss2 = FeedMe::FeedParser.parse(open(fixture('welformed.rss2')).read)
+    @rss1 = FeedMe::FeedParser.parse(open(fixture('welformed.rss1')).read)
   end
 
   it "should be an atom parser for an atom feed" do
@@ -55,16 +56,6 @@ describe FeedMe::FeedParser do
     it_should_behave_like "all parsing methods"
   end
 
-  describe '#encoding' do
-    it "should return the encoding for an atom feed" do
-      @atom.encoding.should == 'UTF-8'
-    end
-
-    it "should return the encoding for an atom feed" do
-      @rss2.encoding.should == 'iso-8859-1'
-    end
-  end
-
   describe '#title' do
     it "should be valid for an atom feed" do
       @atom.title.should == "Test feed"
@@ -73,6 +64,10 @@ describe FeedMe::FeedParser do
     it "should be valid for an rss2 feed" do
       @rss2.title.should == "Lift Off News"
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.title.should == "XML.com"
+    end
   end
 
   describe '#description' do
@@ -83,6 +78,10 @@ describe FeedMe::FeedParser do
     it "should be valid for an rss2 feed" do
       @rss2.description.should == "Liftoff to Space Exploration."
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.description.should == "XML.com features a rich mix of information and services"
+    end
   end
 
   describe '#feed_id' do
@@ -93,6 +92,10 @@ describe FeedMe::FeedParser do
     it "should be nil for an rss2 feed" do
       @rss2.feed_id.should be_nil
     end
+
+    it "should be nil for an rss1 feed" do
+      @rss1.feed_id.should be_nil
+    end
   end
 
   describe '#updated_at' do
@@ -103,6 +106,10 @@ describe FeedMe::FeedParser do
     it "should be valid for an rss2 feed" do
       @rss2.updated_at.should == Time.utc(2003, 6, 10, 9, 41, 1)
     end
+
+    it "should be taken from dublin core time for an rss1 feed" do
+      @rss1.updated_at.should == Time.utc(2010, 6, 3, 14, 56, 42)
+    end
   end
 
   describe '#href' do
@@ -110,9 +117,13 @@ describe FeedMe::FeedParser do
       @atom.href.should == "http://imaginary.host/posts.atom"
     end
 
-    it "should be nil for an atom feed" do
+    it "should be nil for an rss2 feed" do
       @rss2.href.should be_nil
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.href.should == "http://www.xml.com/xml/news.rss"
+    end
   end
 
   describe '#url' do
@@ -123,6 +134,10 @@ describe FeedMe::FeedParser do
     it "should be valid for an rss2 feed" do
       @rss2.url.should == "http://liftoff.msfc.nasa.gov/"
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.url.should == "http://xml.com/pub"
+    end
   end
 
   describe '#generator' do
@@ -133,6 +148,10 @@ describe FeedMe::FeedParser do
     it "should be valid for an rss2 feed" do
       @rss2.generator.should == "Weblog Editor 2.0"
     end
+
+    it "should be nil for an rss1 feed" do
+      @rss1.generator.should be_nil
+    end
   end
 
   describe '#author.name' do
@@ -143,6 +162,10 @@ describe FeedMe::FeedParser do
     it "should be valid for an rss2 feed" do
       @rss2.author.name.should == "Mary Jo"
     end
+
+    it "should be taken from dublin core for an rss1 feed" do
+      @rss1.author.name.should == "Foopaq (mailto:support@foopaq.se)"
+    end
   end
 
   describe '#author.email' do
@@ -153,6 +176,10 @@ describe FeedMe::FeedParser do
     it "should be valid for an rss2 feed" do
       @rss2.author.email.should == "editor@example.com"
     end
+
+    it "should be nil for an rss1 feed" do
+      @rss1.author.email.should be_nil
+    end
   end
 
   describe '#author.uri' do
@@ -163,6 +190,10 @@ describe FeedMe::FeedParser do
     it "should be nil for an rss2 feed" do
       @rss2.author.uri.should be_nil
     end
+
+    it "should be nil for an rss1 feed" do
+      @rss1.author.uri.should be_nil
+    end
   end
 
   describe '#entries' do
@@ -193,6 +224,20 @@ describe FeedMe::FeedParser do
       @rss2.entries.first.item_id.should == "http://liftoff.msfc.nasa.gov/2003/06/03.html#item573"
     end
 
+    it "should return an array of entries for an rss1 feed" do
+      @rss1.entries.should be_an_instance_of(Array)
+    end
+
+    it "should have the correct length for an rss1 feed" do
+      @rss1.should have(3).entries
+    end
+
+    it "should return items that are properly parsed for an rss1 feed" do
+      @rss1.entries.first.title.should == "Processing Inclusions with XSLT"
+      @rss1.entries.first.url.should == "http://google.com/foobar"
+      @rss1.entries.first.item_id.should == "http://xml.com/pub/2000/08/09/xslt/xslt.html"
+    end
+
     it "should allow items to be read more than once" do
       item = @rss2.entries.first
       item.item_id.should == "http://liftoff.msfc.nasa.gov/2003/06/03.html#item573"
@@ -285,6 +330,49 @@ describe FeedMe::FeedParser do
       author.email.should == "editor@example.com"
       author.uri.should be_nil
     end
+
+    it "should serialize the title of an rss1 feed" do
+      @rss1.to_hash[:title].should == "XML.com"
+    end
+
+    it "should serialize the description of an rss1 feed" do
+      @rss1.to_hash[:description].should == "XML.com features a rich mix of information and services"
+    end
+
+    it "should serialize the feed_id of an rss1 feed" do
+      @rss1.to_hash[:feed_id].should be_nil
+    end
+
+    it "should serialize the updated_at time of an rss1 feed" do
+      @rss1.to_hash[:updated_at].should == Time.utc(2010, 6, 3, 14, 56, 42)
+    end
+
+    it "should serialize the href of an rss1 feed" do
+      @rss1.to_hash[:href].should == "http://www.xml.com/xml/news.rss"
+    end
+
+    it "should serialize the url of an rss1 feed" do
+      @rss1.to_hash[:url].should == "http://xml.com/pub"
+    end
+
+    it "should serialize the generator of an rss1 feed" do
+      @rss1.to_hash[:generator].should be_nil
+    end
+
+    it "should serialize the entries of an rss1 feed" do
+      @rss1.to_hash[:entries].should be_an_instance_of(Array)
+      @rss1.to_hash[:entries].first.title.should == "Processing Inclusions with XSLT"
+      @rss1.to_hash[:entries].first.url.should == "http://google.com/foobar"
+    end
+
+    it "should serialize the author of an rss1 feed" do
+
+      author = @rss1.to_hash[:author]
+
+      author.name.should == "Foopaq (mailto:support@foopaq.se)"
+      author.email.should be_nil
+      author.uri.should be_nil
+    end
   end
 
 end

data/spec/item_parser_spec.rb

@@ -9,6 +9,8 @@ describe FeedMe::ItemParser do
     @atom = @atom_feed.entries.first
     @rss2_feed = FeedMe::FeedParser.open(fixture('welformed.rss2'))
     @rss2 = @rss2_feed.entries.first
+    @rss1_feed = FeedMe::FeedParser.open(fixture('welformed.rss1'))
+    @rss1 = @rss1_feed.entries.first
   end
 
   describe '#to_hash' do
@@ -25,6 +27,10 @@ describe FeedMe::ItemParser do
     it "should be valid for an rss2 feed" do
       @rss2.title.should == "Star City"
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.title.should == "Processing Inclusions with XSLT"
+    end
   end
 
   describe '#content' do
@@ -35,6 +41,10 @@ describe FeedMe::ItemParser do
     it "should be valid for an rss2 feed" do
       @rss2.content.should == "This is content"
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.content.should == "Processing document inclusions with general XML tools can be problematic."
+    end
   end
 
   describe '#item_id' do
@@ -45,6 +55,10 @@ describe FeedMe::ItemParser do
     it "should be valid for an rss2 feed" do
       @rss2.item_id.should == "http://liftoff.msfc.nasa.gov/2003/06/03.html#item573"
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.item_id.should == "http://xml.com/pub/2000/08/09/xslt/xslt.html"
+    end
   end
 
   describe '#updated_at' do
@@ -55,6 +69,10 @@ describe FeedMe::ItemParser do
     it "should be valid for an rss2 feed" do
       @rss2.updated_at.should == Time.utc(2003, 6, 3, 9, 39, 21)
     end
+
+    it "should be taken from dublin core for an rss1 feed" do
+      @rss1.updated_at.should == Time.utc(2010, 6, 3, 10, 7, 42)
+    end
   end
 
   describe '#url' do
@@ -65,6 +83,10 @@ describe FeedMe::ItemParser do
     it "should be valid for an rss2 feed" do
       @rss2.url.should == "http://liftoff.msfc.nasa.gov/news/2003/news-starcity.asp"
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.url.should == "http://google.com/foobar"
+    end
   end
 
   describe '#categories' do
@@ -81,6 +103,10 @@ describe FeedMe::ItemParser do
     it "should be valid for an rss2 feed" do
       @rss2.author.name.should == "Chuck Norris"
     end
+
+    it "should be taken from dublin core for an rss1 feed" do
+      @rss1.author.name.should == "akeri.se"
+    end
   end
 
   describe '#author.email' do
@@ -91,6 +117,10 @@ describe FeedMe::ItemParser do
     it "should be valid for an rss2 feed" do
       @rss2.author.email.should == "da_man@example.com"
     end
+
+    it "should be nil for an rss1 feed" do
+      @rss1.author.email.should be_nil
+    end
   end
 
   describe '#author.uri' do
@@ -101,6 +131,10 @@ describe FeedMe::ItemParser do
     it "should be nil for an rss2 feed" do
       @rss2.author.uri.should be_nil
     end
+
+    it "should be nil for an rss1 feed" do
+      @rss1.author.uri.should be_nil
+    end
   end
 
   describe '#to_hash' do
@@ -152,6 +186,31 @@ describe FeedMe::ItemParser do
       author.email.should == "da_man@example.com"
       author.uri.should be_nil
     end
+
+    it "should serialize the title for an rss1 feed" do
+      @rss1.to_hash[:title].should == "Processing Inclusions with XSLT"
+    end
+
+    it "should serialize the item_id for an rss1 feed" do
+      @rss1.to_hash[:item_id].should == "http://xml.com/pub/2000/08/09/xslt/xslt.html"
+    end
+
+    it "should serialize updated_at for an rss1 feed" do
+      @rss1.to_hash[:updated_at].should == Time.utc(2010, 6, 3, 10, 7, 42)
+    end
+
+    it "should serialize the url for an rss1 feed" do
+      @rss1.to_hash[:url].should == "http://google.com/foobar"
+    end
+
+    it "should serialize the author of an rss1 feed" do
+      author = @rss1.to_hash[:author]
+
+      author.name.should == "akeri.se"
+      author.email.should be_nil
+      author.uri.should be_nil
+    end
+
   end
 
 end
@@ -163,6 +222,8 @@ describe "Without an author", FeedMe::ItemParser do
     @atom = @atom_feed.entries[1]
     @rss2_feed = FeedMe::FeedParser.open(fixture('welformed.rss2'))
     @rss2 = @rss2_feed.entries[1]
+    @rss1_feed = FeedMe::FeedParser.open(fixture('welformed.rss1'))
+    @rss1 = @rss1_feed.entries[1]
   end
 
   describe '#author.name' do
@@ -173,6 +234,10 @@ describe "Without an author", FeedMe::ItemParser do
     it "should be valid for an rss2 feed" do
       @rss2.author.name.should be_nil
     end
+
+    it "should be valid for an rss1 feed" do
+      @rss1.author.name.should be_nil
+    end
   end
 
 end

metadata

@@ -1,12 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: feed_me
 version: !ruby/object:Gem::Version 
+  hash: 3
   prerelease: false
   segments: 
   - 0
-  - 6
-  - 2
-  version: 0.6.2
+  - 7
+  - 0
+  version: 0.7.0
 platform: ruby
 authors: 
 - Jonas Nicklas
@@ -15,16 +16,18 @@ autorequire: feed_me
 bindir: bin
 cert_chain: []
 
-date: 2010-03-22 00:00:00 +00:00
+date: 2010-06-04 00:00:00 +02:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: nokogiri
   prerelease: false
   requirement: &id001 !ruby/object:Gem::Requirement 
+    none: false
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
+        hash: 3
         segments: 
         - 0
         version: "0"
@@ -50,6 +53,9 @@ files:
 - lib/feed_me/feed_parser.rb
 - lib/feed_me/item_parser.rb
 - lib/feed_me/person_parser.rb
+- spec/feed_parser_spec.rb
+- spec/item_parser_spec.rb
+- spec/spec_helper.rb
 has_rdoc: true
 homepage: http://github.com/jnicklas/feed_me
 licenses: []
@@ -60,23 +66,27 @@ rdoc_options:
 require_paths: 
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement 
+  none: false
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
+      hash: 3
       segments: 
       - 0
       version: "0"
 required_rubygems_version: !ruby/object:Gem::Requirement 
+  none: false
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
+      hash: 3
       segments: 
       - 0
       version: "0"
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.3.6
+rubygems_version: 1.3.7
 signing_key: 
 specification_version: 3
 summary: Nice and simple RSS and atom feed parsing built on hpricot