rdfa_parser 0.1.2 → 0.1.3

data/.gitmodules

@@ -0,0 +1,3 @@
+[submodule "spec/rdfa-test-suite"]
+	path = spec/rdfa-test-suite
+	url = git://github.com/msporny/rdfa-test-suite.git

data/History.txt

@@ -1,12 +1,20 @@
-=== 0.1.2 2009-11-04
-* First gem release, all XHTML1 approved test cases pass at http://rdfa.digitalbazaar.com/test-suite/
+=== 0.1.3
+* Added rdfa-test-suite as sub-module and refactored RDFa tests to use them.
+* Fix but in white-space separated attributes
+* Downcase XHTML reserved words
+* Check complete xmlns prefix for validity
 
-=== 0.1.2 2009-10-27
+=== 0.1.2
 * Move most test-case logic into rdfa_helper.
 * Separate approved and unreviewed tests.
 * Allow for soft failure of unreviewed tests by catching Spec::Expectations::ExpectationNotMetError.
 * Integrate rdfa-test-suite and create test cases in the same manner. (Causing SPARQL problems with XMLLiterals)
 
+=== 0.1.1 2009-10-25
+* Added VERSION constant by reading VERSION file from gem root.
+* Fix [] parsing to not return a resource (TC 0121 and 0122).
+* Improve test-case manifest file parsing to extract expectedResults and use in matchers.
+
 === 0.1.0 2009-10-17
 * First gem release, all XHTML1 approved test cases pass at http://rdfa.digitalbazaar.com/test-suite/
 

data/VERSION

@@ -1 +1 @@
-0.1.2
+0.1.3

data/bin/rdfa_parser

@@ -1,5 +1,4 @@
 #!/usr/bin/env ruby -s
-# For some reason, options preceeding files are ignored
 require 'rubygems'
 require File.expand_path(File.dirname(__FILE__) + "/../lib/rdfa_parser")
 require 'getoptlong'
@@ -20,7 +19,6 @@ class Parse
     puts parser.debug.join("\n\t") if $verbose
   rescue RdfaParser::ParserException => e
     puts "Parse failure: #{e.message}"
-    #puts e.backtrace
     puts parser.debug if $verbose && parser
     #raise
   rescue Exception => e
@@ -35,7 +33,6 @@ base_uri  = "http://example.com"
 
 opts = GetoptLong.new(
   ["--verbose", GetoptLong::NO_ARGUMENT],
-  ["--version", GetoptLong::NO_ARGUMENT],
   ["--quiet", GetoptLong::NO_ARGUMENT],
   ["--debug", GetoptLong::NO_ARGUMENT],
   ["--format", GetoptLong::REQUIRED_ARGUMENT],
@@ -44,9 +41,6 @@ opts = GetoptLong.new(
 opts.each do |opt, arg|
   case opt
   when '--verbose' then $verbose = true
-  when '--version'
-    puts "version: #{RdfaParser::VERSION}"
-    exit(0)
   when '--quiet' then $quiet = true
   when '--debug' then $DEBUG = true
   when '--format' then $format = arg

data/lib/rdfa_parser/literal.rb

@@ -243,16 +243,27 @@ module RdfaParser
 
       # Map namespaces from context to each top-level element found within snippet
       def encode_contents(contents, options)
-        ns_hash = options[:namespaces].values.inject({}) {|h, ns| h.merge(ns.xmlns_hash)}
-        ns_strs = []
-        ns_hash.each_pair {|a, u| ns_strs << "#{a}=\"#{u}\""}
-        
-        # Add inherited namespaces to created root element so that they're inherited to sub-elements
-        contents = Nokogiri::XML::Document.parse("<foo #{ns_strs.join(" ")}>#{contents}</foo>").root
-
-        @contents = contents.children.map do |c|
+        if contents.is_a?(String)
+          ns_hash = options[:namespaces].values.inject({}) {|h, ns| h.merge(ns.xmlns_hash)}
+          ns_strs = []
+          ns_hash.each_pair {|a, u| ns_strs << "#{a}=\"#{u}\""}
+
+          # Add inherited namespaces to created root element so that they're inherited to sub-elements
+          contents = Nokogiri::XML::Document.parse("<foo #{ns_strs.join(" ")}>#{contents}</foo>").root.children
+        end
+
+        # Add already mapped namespaces and language
+        @contents = contents.map do |c|
           if c.is_a?(Nokogiri::XML::Element)
-            ns_hash.each_pair { |a, u| c[a] = u unless c.namespaces[a]}
+            # Gather namespaces from self and decendant nodes
+            c.traverse do |n|
+              ns = n.namespace
+              next unless ns
+              prefix = ns.prefix ? "xmlns:#{ns.prefix}" : "xmlns"
+              c[prefix] = ns.href unless c.namespaces[prefix]
+            end
+            
+            # Add lanuage
             if options[:language] && c["lang"].to_s.empty?
               c["xml:lang"] = options[:language]
             end

data/lib/rdfa_parser/namespace.rb

@@ -28,7 +28,7 @@ module RdfaParser
       if shortname_valid?(short)
         @short = short
       else
-        raise ParserException, "Invalid shortname '#{short}"
+        raise ParserException, "Invalid shortname '#{short}'"
       end
     end
 
@@ -81,7 +81,7 @@ module RdfaParser
     
     private
     def shortname_valid?(shortname)
-      if shortname =~ /[a-zA-Z_][a-zA-Z0-9_]+/ || shortname.nil?
+      if shortname =~ /\A[a-zA-Z_][a-zA-Z0-9_]+\Z/ || shortname.nil?
         return true
       else
         return false

data/lib/rdfa_parser/rdfa_parser.rb

@@ -70,7 +70,8 @@ module RdfaParser
     # HTML4::
     # HTML5::
     # Options:
-    # profile:: One of _xhtml1_, _html4_, or _html5_ to override intuition
+    # _profile_:: One of _xhtml1_, _html4_, or _html5_ to override intuition
+    # _strict_:: Fail when error detected, otherwise just continue
     # 
     # Raises RdfaParser::RdfaException or subclass
     def parse(stream, base, options = {}, &block) # :yields: triple
@@ -79,8 +80,10 @@ module RdfaParser
       when Nokogiri::XML::Document then stream
       else   Nokogiri::XML.parse(stream, base)
       end
+      
+      @strict = options[:strict]
       @base = base.to_s
-      raise ParserException, "Empty document" if @doc.nil?
+      raise ParserException, "Empty document" if @doc.nil? && @strict
       @callback = block
 
       # If the doc has a default, use that as "html"
@@ -106,6 +109,9 @@ module RdfaParser
         @graph << triple
       end
       triple
+    rescue RdfException => e
+      add_debug(node, "add_triple raised #{e.class}: #{e.message}")
+      raise if @strict
     end
   
     # Parsing an RDFa document (this is *not* the recursive method)
@@ -131,8 +137,13 @@ module RdfaParser
     
       # look for xmlns
       element.namespaces.each do |attr_name,attr_value|
-        abbr, suffix = attr_name.split(":")
-        mappings[suffix] = @graph.namespace(attr_value, suffix) if abbr == "xmlns"
+        begin
+          abbr, suffix = attr_name.split(":")
+          mappings[suffix] = @graph.namespace(attr_value, suffix) if abbr == "xmlns"
+        rescue RdfException => e
+          add_debug(element, "extract_mappings raised #{e.class}: #{e.message}")
+          raise if @strict
+        end
       end
 
       add_debug(element, "mappings: #{mappings.keys.join(", ")}")
@@ -150,7 +161,11 @@ module RdfaParser
 
     # The recursive helper function
     def traverse(element, evaluation_context)
-      raise ParserException, "Can't parse nil element" if element.nil?
+      if element.nil?
+        add_debug(element, "traverse nil element")
+        raise ParserException, "Can't parse nil element" if @strict
+        return nil
+      end
       
       # local variables [5.5 Step 1]
       recurse = true
@@ -313,7 +328,7 @@ module RdfaParser
         elsif children_node_types != [Nokogiri::XML::Text] and (type == nil or type_resource.to_s == XML_LITERAL.to_s)
           # XML Literal
           add_debug(element, "XML Literal: #{element.inner_html}")
-          current_object_literal = Literal.typed(element.inner_html, XML_LITERAL, :language => language, :namespaces => uri_mappings)
+          current_object_literal = Literal.typed(element.children, XML_LITERAL, :language => language, :namespaces => uri_mappings)
           recurse = false
         end
       
@@ -374,7 +389,9 @@ module RdfaParser
         if curie.include?(":")
           resource_array << curie_to_resource_or_bnode(curie, uri_mappings, base)
         elsif with_link_types
-          link_type_curie = curie_to_resource_or_bnode(":#{value}", XH_MAPPING, base) if LINK_TYPES.include?(value.to_s)
+          # Reserved words are all mapped to lower case
+          curie = curie.to_s.downcase
+          link_type_curie = curie_to_resource_or_bnode(":#{curie}", XH_MAPPING, base) if LINK_TYPES.include?(curie)
           resource_array << link_type_curie if link_type_curie
         end
       end
@@ -397,7 +414,11 @@ module RdfaParser
         nil
       else
         ns = uri_mappings[prefix.to_s]
-        raise ParserException, "No namespace mapping for #{prefix}" unless ns
+        unless ns
+          add_debug(nil, "curie_to_resource_or_bnode No namespace mapping for #{prefix}")
+          raise ParserException, "No namespace mapping for #{prefix}" if @strict
+          return nil
+        end
         ns + suffix
       end
     end

data/lib/rdfa_parser/triple.rb

@@ -84,14 +84,12 @@ module RdfaParser
       case object
       when Addressable::URI
         URIRef.new(object.to_s)
-      when String
+      when String, Integer, Float
         if object.to_s =~ /\S+\/\/\S+/ # does it smell like a URI?
           URIRef.new(object.to_s)
         else
-          Literal.untyped(object)
+          Literal.new(object, nil, nil)
         end
-      when Integer, Float
-        Literal.build_from(object)
       when URIRef, BNode, Literal
         object
       else

data/spec/graph_spec.rb

@@ -74,7 +74,7 @@ describe "Graphs" do
       xhtml = Namespace.new("http://www.w3.org/1999/xhtml", "")
       g = Graph.new
       g << Triple.new(
-        URIRef.new("http://www.w3.org/2006/07/SWD/RDFa/testsuite/xhtml1-testcases/0011.xhtml"),
+        URIRef.new("http://rdfa.digitalbazaar.com/test-suite/test-cases/xhtml1/0011.xhtml"),
         URIRef.new("http://purl.org/dc/elements/1.1/title"),
         Literal.typed("E = mc<sup xmlns=\"http://www.w3.org/1999/xhtml\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\">2</sup>: The Most Urgent Problem of Our Time",
                       "http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral",
@@ -86,7 +86,7 @@ describe "Graphs" do
     }
     
     it "should output NTriple" do
-      nt = '<http://www.w3.org/2006/07/SWD/RDFa/testsuite/xhtml1-testcases/0011.xhtml> <http://purl.org/dc/elements/1.1/title> "E = mc<sup xmlns=\"http://www.w3.org/1999/xhtml\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\">2</sup>: The Most Urgent Problem of Our Time"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> .' + "\n"
+      nt = '<http://rdfa.digitalbazaar.com/test-suite/test-cases/xhtml1/0011.xhtml> <http://purl.org/dc/elements/1.1/title> "E = mc<sup xmlns=\"http://www.w3.org/1999/xhtml\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\">2</sup>: The Most Urgent Problem of Our Time"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> .' + "\n"
       subject.to_ntriples.should == nt
     end
 
@@ -94,7 +94,7 @@ describe "Graphs" do
       rdfxml = <<-HERE
 <?xml version="1.0" encoding="UTF-8"?>
 <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:xml=\"http://www.w3.org/XML/1998/namespace\" xmlns:rdfs=\"http://www.w3.org/2000/01/rdf-schema#\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xmlns=\"http://www.w3.org/1999/xhtml\" xmlns:xhv=\"http://www.w3.org/1999/xhtml/vocab#\">
-  <rdf:Description rdf:about="http://www.w3.org/2006/07/SWD/RDFa/testsuite/xhtml1-testcases/0011.xhtml">
+  <rdf:Description rdf:about="http://rdfa.digitalbazaar.com/test-suite/test-cases/xhtml1/0011.xhtml">
     <dc:title rdf:parseType="Literal">E = mc<sup xmlns="http://www.w3.org/1999/xhtml" xmlns:dc="http://purl.org/dc/elements/1.1/">2>/sup>: The Most Urgent Problem of Our Time</dc:title>
   </rdf:Description>
 </rdf:RDF>

data/spec/literal_spec.rb

@@ -24,8 +24,8 @@ describe "Literals: " do
       subject { Literal.untyped("松本 后子") }
       
       describe "encodings" do
-        it "should return n3" do subject.to_n3.should == '"\u677E\u672C \u540E\u5B50"' end
-        it "should return ntriples" do subject.to_ntriples.should == '"\u677E\u672C \u540E\u5B50"' end
+        it "should return n3" do subject.to_n3.should == "\"\\\u677E\\\u672C \\\u540E\\\u5B50\"" end
+        it "should return ntriples" do subject.to_ntriples.should == "\"\\\u677E\\\u672C \\\u540E\\\u5B50\"" end
         it "should return xml_args" do subject.xml_args.should == ["松本 后子", {}] end
         it "should return TriX" do subject.to_trix.should == "<plainLiteral>" + "松本 后子" + "</plainLiteral>" end
       end
@@ -151,10 +151,10 @@ describe "Literals: " do
       }
     
       describe "encodings" do
-        it "should return n3" do subject.to_n3.should == "\"foo <sup xmlns:dc=\\\"http://purl.org/dc/elements/1.1/\\\">bar</sup> baz!\"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral>" end
+        it "should return n3" do subject.to_n3.should == "\"foo <sup>bar</sup> baz!\"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral>" end
         it "should return ntriples" do subject.to_ntriples.should == subject.to_n3 end
-        it "should return xml_args" do subject.xml_args.should == ["foo <sup xmlns:dc=\"http://purl.org/dc/elements/1.1/\">bar</sup> baz!", {"rdf:parseType" => "Literal"}] end
-        it "should return TriX" do subject.to_trix.should == "<typedLiteral datatype=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral\">foo <sup xmlns:dc=\"http://purl.org/dc/elements/1.1/\">bar</sup> baz!</typedLiteral>" end
+        it "should return xml_args" do subject.xml_args.should == ["foo <sup>bar</sup> baz!", {"rdf:parseType" => "Literal"}] end
+        it "should return TriX" do subject.to_trix.should == "<typedLiteral datatype=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral\">foo <sup>bar</sup> baz!</typedLiteral>" end
       end
       
       describe "and language" do
@@ -165,10 +165,10 @@ describe "Literals: " do
         }
 
         describe "encodings" do
-          it "should return n3" do subject.to_n3.should == "\"foo <sup xmlns:dc=\\\"http://purl.org/dc/elements/1.1/\\\" xml:lang=\\\"fr\\\">bar</sup> baz!\"\^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral>" end
+          it "should return n3" do subject.to_n3.should == "\"foo <sup xml:lang=\\\"fr\\\">bar</sup> baz!\"\^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral>" end
           it "should return ntriples" do subject.to_ntriples.should == subject.to_n3 end
-          it "should return xml_args" do subject.xml_args.should == ["foo <sup xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xml:lang=\"fr\">bar</sup> baz!", {"rdf:parseType" => "Literal"}] end
-          it "should return TriX" do subject.to_trix.should == "<typedLiteral datatype=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral\">foo <sup xmlns:dc=\"http://purl.org/dc/elements/1.1/\" xml:lang=\"fr\">bar</sup> baz!</typedLiteral>" end
+          it "should return xml_args" do subject.xml_args.should == ["foo <sup xml:lang=\"fr\">bar</sup> baz!", {"rdf:parseType" => "Literal"}] end
+          it "should return TriX" do subject.to_trix.should == "<typedLiteral datatype=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral\">foo <sup xml:lang=\"fr\">bar</sup> baz!</typedLiteral>" end
         end
       end
       
@@ -210,7 +210,10 @@ describe "Literals: " do
           content = root.css("h2").children
           Literal.typed(content,
                         "http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral",
-                        :namespaces => {"svg" => Namespace.new("http://www.w3.org/2000/svg", "svg")})
+                        :namespaces => {
+                          "svg" => Namespace.new("http://www.w3.org/2000/svg", "svg"),
+                          "dc" => Namespace.new("http://purl.org/dc/elements/1.1/", "dc")
+                        })
         }
         it "should return xml_args" do subject.xml_args.should == ["<svg:svg xmlns:svg=\"http://www.w3.org/2000/svg\"></svg:svg>", {"rdf:parseType" => "Literal"}] end
       end

data/spec/matchers.rb

@@ -3,6 +3,7 @@ module Matchers
     def initialize(expected, info)
       @expected = expected
       @info = info
+      @diffable = true
     end
     def matches?(actual)
       @actual = actual
@@ -28,7 +29,8 @@ module Matchers
       end +
       "\n\n#{info + "\n" unless info.empty?}" +
       "Unsorted Expected:\n#{@expected.to_ntriples}" +
-      "Unsorted Results:\n#{@actual.to_ntriples}"
+      "Unsorted Results:\n#{@actual.to_ntriples}" +
+      "\nDebug:\n#{@info.trace}"
     end
   end
   
@@ -48,7 +50,7 @@ module Matchers
       @expected_results = @info.respond_to?(:expectedResults) ? @info.expectedResults : true
       model = Redland::Model.new
       ntriples_parser = Redland::Parser.ntriples
-      ntriples_parser.parse_string_into_model(model, actual.to_ntriples, "http://www.w3.org/2006/07/SWD/RDFa/testsuite/xhtml1-testcases/")
+      ntriples_parser.parse_string_into_model(model, actual.to_ntriples, "http://rdfa.digitalbazaar.com/test-suite/test-cases/xhtml1/")
 
       @results = @query.execute(model)
       if @expected_results
@@ -70,7 +72,8 @@ module Matchers
         "Query returned true (expected false)"
       end +
       "\n#{@expected}" +
-      "\nResults:\n#{@actual.to_ntriples}"
+      "\nResults:\n#{@actual.to_ntriples}" +
+      "\nDebug:\n#{@info.trace}"
     end
   end
 

data/spec/rdfa-triples/0001.nt

@@ -0,0 +1 @@
+<$TCPATH/photo1.jpg> <http://purl.org/dc/elements/1.1/creator> "Mark Birbeck" .

data/spec/rdfa-triples/0006.nt

@@ -0,0 +1,2 @@
+<http://www.blogger.com/profile/1109404> <http://xmlns.com/foaf/0.1/img> <$TCPATH/photo1.jpg> .
+<$TCPATH/photo1.jpg> <http://purl.org/dc/elements/1.1/creator> <http://www.blogger.com/profile/1109404> .

data/spec/rdfa-triples/0007.nt

@@ -0,0 +1,3 @@
+<http://www.blogger.com/profile/1109404> <http://xmlns.com/foaf/0.1/img> <$TCPATH/photo1.jpg> .
+<$TCPATH/photo1.jpg> <http://purl.org/dc/elements/1.1/title> "Portrait of Mark" .
+<$TCPATH/photo1.jpg> <http://purl.org/dc/elements/1.1/creator> <http://www.blogger.com/profile/1109404> .

data/spec/rdfa-triples/0008.nt

@@ -0,0 +1 @@
+<$TCPATH/0008.xhtml> <http://creativecommons.org/ns#license> <http://creativecommons.org/licenses/by-nc-nd/2.5/> .

data/spec/rdfa-triples/0011.nt

@@ -0,0 +1,3 @@
+<$TCPATH/0011.xhtml> <http://purl.org/dc/elements/1.1/creator> "Albert Einstein" .
+# This is one of the allowed result XML Literals
+<$TCPATH/0011.xhtml> <http://purl.org/dc/elements/1.1/title> "E = mc<sup xmlns=\"http://www.w3.org/1999/xhtml\" xmlns:dc=\"http://purl.org/dc/elements/1.1/\">2</sup>: The Most Urgent Problem of Our Time"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral> .

data/spec/rdfa-triples/0015.nt

@@ -0,0 +1,2 @@
+<$TCPATH/0015.xhtml> <http://purl.org/dc/elements/1.1/creator> "Fyodor Dostoevsky" .
+<$TCPATH/0015.xhtml> <http://purl.org/dc/elements/1.1/source> <urn:isbn:0140449132> .

data/spec/rdfa-triples/0018.nt

@@ -0,0 +1 @@
+<$TCPATH/photo1.jpg> <http://purl.org/dc/elements/1.1/creator> <http://www.blogger.com/profile/1109404> .

data/spec/rdfa-triples/0020.nt

@@ -0,0 +1 @@
+<$TCPATH/photo1.jpg> <http://purl.org/dc/elements/1.1/creator> "Mark Birbeck" .

data/spec/rdfa-triples/0021.nt

@@ -0,0 +1 @@
+<$TCPATH/0021.xhtml> <http://purl.org/dc/elements/1.1/creator> "Mark Birbeck" .

data/spec/rdfa-triples/0023.nt

@@ -0,0 +1 @@
+<$TCPATH/0023.xhtml> <http://purl.org/dc/elements/1.1/creator> "Mark Birbeck" .

data/spec/rdfa-triples/0025.nt

@@ -0,0 +1,2 @@
+<$TCPATH/0025.xhtml> <http://purl.org/dc/elements/1.1/creator> <$TCPATH/0025.xhtml#me> .
+<$TCPATH/0025.xhtml#me> <http://xmlns.com/foaf/0.1/name> "Ben Adida" .

data/spec/rdfa-triples/0030.nt

@@ -0,0 +1 @@
+<$TCPATH/0030.xhtml> <http://creativecommons.org/ns#license> <http://creativecommons.org/licenses/by-nc-nd/2.5/> .

data/spec/rdfa-triples/0031.nt

@@ -0,0 +1 @@
+<$TCPATH/0031.xhtml#wtw> <http://purl.org/dc/elements/1.1/identifier> <urn:ISBN:0752820907> .

data/spec/rdfa-triples/0032.nt

@@ -0,0 +1 @@
+<$TCPATH/0032.xhtml#wtw> <http://purl.org/dc/elements/1.1/identifier> <urn:ISBN:0752820907> .

data/spec/rdfa-triples/0033.nt

@@ -0,0 +1,2 @@
+<$TCPATH/0033.xhtml> <http://purl.org/dc/elements/1.1/creator> _:a .
+_:a <http://xmlns.com/foaf/0.1/name> "Ben Adida" .

data/spec/rdfa-triples/0051.nt

@@ -0,0 +1,2 @@
+<$TCPATH/0051.xhtml> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://xmlns.com/foaf/0.1/Document> .
+<$TCPATH/0051.xhtml> <http://xmlns.com/foaf/0.1/topic> "John Doe" .