open_graph_reader 0.1.0 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: fb82d8a9867ad4c1cf0258ef4e97f993cccdc6aa
-  data.tar.gz: ef67382716657161c2098380bf00c6abc48e6a4e
+  metadata.gz: bea1c38c6567749ce9bd8b369867d5e7c0b7a2b3
+  data.tar.gz: b0805bdd3e5e1efddc7f73c59dd40ad7315b99c3
 SHA512:
-  metadata.gz: a60f2d67610ab9d1cd0587c58b92452f07fff77da489eec68c7906af650247bd42e26506a5d323f9d6ebe4b3bb247b7b26aba2dac5389634c660d80801b6f32b
-  data.tar.gz: 423fb2fe90278c64bd089579c5165c9ee468663c3d9eeb0661c1dfa0a5adc74f12a8772f7f5bdeff0d15ecb52493887d822e13a68dae31de98f9056636e21666
+  metadata.gz: 6d3d1edbaa5b0ac800300c79631d5494e6eebf7cc362b618621ca6006530d303c7de679fa761ae5d6edd3b826fa069f1dde3ff6f76dbd295b3781e68fa4eb29e
+  data.tar.gz: f9ba7d6080e164c75491abbd297994627f2812e907c230c5edb5c4f42681657da27c7e730b83b42df19fc96ddcb9fd9a3cd44b3121f933ac45d8db1635b040c6

data/lib/open_graph_reader.rb

@@ -1,18 +1,14 @@
 require 'uri'
 
-begin
-  require 'faraday_middleware/response/follow_redirects'
-rescue LoadError; end
-
 require 'open_graph_reader/base'
 require 'open_graph_reader/builder'
+require 'open_graph_reader/configuration'
 require 'open_graph_reader/definitions'
 require 'open_graph_reader/fetcher'
 require 'open_graph_reader/object'
 require 'open_graph_reader/parser'
 require 'open_graph_reader/version'
 
-# @todo  quirks mode where invalid attributes don't raise?
 # @todo 1.1 compatibility mode?
 # This module provides the main entry to the library. Please see the
 # {file:README.md} for usage examples.
@@ -72,6 +68,24 @@ module OpenGraphReader
   rescue NoOpenGraphDataError, InvalidObjectError
   end
 
+  # Configure the library, see {Configuration} for the list of available
+  # options and their defaults. Changing configuration at runtime is not
+  # thread safe.
+  #
+  # @yieldparam [Configuration] the configuration object
+  # @see Configuration
+  def self.configure
+    yield config
+  end
+
+  # Get the current {Configuration} instance
+  #
+  # @api private
+  # @return [Configuration]
+  def self.config
+    Configuration.instance
+  end
+
   # The target couldn't be fetched, didn't contain any HTML or
   # any OpenGraph tags.
   class NoOpenGraphDataError < StandardError
@@ -80,4 +94,10 @@ module OpenGraphReader
   # The target did contain OpenGraph tags, but they're not valid.
   class InvalidObjectError < StandardError
   end
+
+  # The target defines a namespace we have no definition for
+  #
+  # @api private
+  class UnknownNamespaceError < StandardError
+  end
 end

data/lib/open_graph_reader/base.rb

@@ -19,7 +19,13 @@ module OpenGraphReader
     #   @param [String] name The name of the root namespace.
     #   @param [Object] object The corresponding root object.
     #   @api private
-    def_delegators :@bases, :[], :[]=
+    # @!method each
+    #   Traverse the available objects
+    #
+    #   @yield [Object]
+    #   @api private
+    def_delegators :@bases, :[], :[]=, :each_value
+    alias_method :each, :each_value
 
     # If available, contains the source location of the document the
     # available objects were parsed from.

data/lib/open_graph_reader/builder.rb

@@ -1,27 +1,13 @@
 module OpenGraphReader
   # Convert a {Parser::Graph} into the right hierarchy of {Object}s attached
-  # to a {Base}.
+  # to a {Base}, then validate it.
   #
-  # @todo validate required, verticals
   # @api private
   class Builder
     # Well-known types from
     #
     # @see http://ogp.me
-    KNOWN_TYPES = %w(
-      website
-      music.song
-      music.album
-      music.playlist
-      music.radio_station
-      video.movie
-      video.episode
-      video.tv_show
-      video.other
-      article
-      book
-      profile
-    ).freeze
+    KNOWN_TYPES = %w(website article book profile).freeze
 
     # Create a new builder.
     #
@@ -41,45 +27,53 @@ module OpenGraphReader
     def base
       base = Base.new
 
-      type = @graph.fetch 'og:type', 'website'
+      type = @graph.fetch('og:type', 'website').downcase
 
       validate_type type
 
       @graph.each do |property|
-        root, *path, name = property.path
-        base[root] ||= Object::Registry[root].new
-        object = resolve base[root], root, path
-
-        if object.respond_to? "#{name}s" # Collection # TODO
-          collection = object.public_send "#{name}s" #TODO
-          if Object::Registry.registered? property.fullname # of subobjects
-            object = Object::Registry[property.fullname].new
-            collection << object
-            object.content = property.content
-          else # of type
-            collection << property.content
-          end
-        elsif Object::Registry.registered? property.fullname # Subobject
-          object[name] ||= Object::Registry[property.fullname].new
-          object[name].content = property.content
-        else # Direct attribute
-          object[name] = property.content
-        end
+        build_property base, property
       end
 
+      validate base
+
       base
     end
 
     private
 
+    def build_property base, property
+      root, *path, name = property.path
+      base[root] ||= Object::Registry[root].new
+      object = resolve base[root], root, path
+
+      if object.has_property?(name) && object.respond_to?("#{name}s") # Collection
+        collection = object.public_send "#{name}s"
+        if Object::Registry.registered? property.fullname # of subobjects
+          object = Object::Registry[property.fullname].new
+          collection << object
+          object.content = property.content
+        else # of type
+          collection << property.content
+        end
+      elsif Object::Registry.registered? property.fullname # Subobject
+        object[name] ||= Object::Registry[property.fullname].new
+        object[name].content = property.content
+      else # Direct attribute
+        object[name] = property.content
+      end
+    rescue UnknownNamespaceError => e
+      raise InvalidObjectError, e.message if OpenGraphReader.config.strict
+    end
+
     def resolve object, last_namespace, path
       return object if path.empty?
 
       next_name = path.shift
-      if object.respond_to? "#{next_name}s" # collection # TODO: do not respond_to? with user data
-        collection = object.public_send("#{next_name}s") # TODO: do not public_send with user data
+      if object.has_property?(next_name) && object.respond_to?("#{next_name}s") # collection
+        collection = object.public_send("#{next_name}s")
         next_object = collection.last
-        if next_object.nil? #|| path.empty? # Final namespace or missing previous declaration, create a new collection item
+        if next_object.nil? # Final namespace or missing previous declaration, create a new collection item
           next_object = Object::Registry[[*last_namespace, next_name].join(':')].new
           collection << next_object
         end
@@ -92,9 +86,43 @@ module OpenGraphReader
     end
 
     def validate_type type
-      unless KNOWN_TYPES.include?(type) || @additional_namespaces.include?(type)
+      return unless OpenGraphReader.config.strict
+
+      unless KNOWN_TYPES.include?(type) ||
+             @additional_namespaces.include?(type) ||
+             Object::Registry.verticals.include?(type)
         raise InvalidObjectError, "Undefined type #{type}"
       end
     end
+
+    def validate base
+      base.each do |object|
+        validate_required object if OpenGraphReader.config.validate_required
+        validate_verticals object, base.og.type
+      end
+    end
+
+    def validate_required object
+      object.class.required_properties.each do |property|
+        if object[property].nil?
+          raise InvalidObjectError, "Missing required property #{property} on #{object.inspect}"
+        end
+      end
+    end
+
+    def validate_verticals object, type
+      return unless type.include? '.'
+      verticals = object.class.verticals
+      if verticals.has_key? type
+        valid_properties = verticals[type]
+        set_properties = object.class.available_properties.select {|property| object[property] }
+        extra_properties = set_properties-valid_properties
+
+        unless extra_properties.empty?
+          raise InvalidObjectError, "Set invalid property #{extra_properties.first} for #{type} " \
+            "in #{object.inspect}, valid properties are #{valid_properties.inspect}"
+        end
+      end
+    end
   end
 end

data/lib/open_graph_reader/configuration.rb

@@ -0,0 +1,51 @@
+require 'singleton'
+
+module OpenGraphReader
+  # The behavior of this library can be tweaked with some parameters.
+  #
+  # @example
+  #   OpenGraphReader.configure do |config|
+  #     config.strict = true
+  #   end
+  class Configuration
+    include Singleton
+
+    # Strict mode (default: <tt>false</tt>)
+    #
+    # In strict mode, if the fetched site defines an unknown type
+    # or property, {InvalidObjectError} is thrown instead of just ignoring
+    # those.
+    #
+    # @return [Bool]
+    attr_accessor :strict
+
+    # Validate required (default: <tt>true</tt>)
+    #
+    # Validate that required properties exist. If this is enabled and
+    # they do not,  {InvalidObjectError} is thrown.
+    #
+    # @return [Bool]
+    attr_accessor :validate_required
+
+    # Validate references (default: <tt>true</tt>)
+    #
+    # If an object should be a reference to another object,
+    # validate that it contains an URL. Be careful in turning this off,
+    # an attacker could place things like <tt>javascript:</tt> links there.
+    #
+    # @return [Bool]
+    attr_accessor :validate_references
+
+    # @private
+    def initialize
+      reset_to_defaults!
+    end
+
+    # Reset configuration to their defaults
+    def reset_to_defaults!
+      @strict = false
+      @validate_required = true
+      @validate_references = true
+    end
+  end
+end

data/lib/open_graph_reader/definitions.rb

@@ -9,7 +9,7 @@ module OpenGraphReader
 
     # @!macro property
     # @return [String]
-    string :type,  required: true, default: 'website'
+    string :type,  required: true, downcase: true, default: 'website'
 
     # @!macro property
     # @return [String]

data/lib/open_graph_reader/fetcher.rb

@@ -1,5 +1,13 @@
 require 'faraday'
 
+begin
+  require 'faraday_middleware/response/follow_redirects'
+rescue LoadError; end
+
+begin
+  require 'faraday/cookie_jar'
+rescue LoadError; end
+
 module OpenGraphReader
   # Fetch an URI to retrieve its HTML body, if available.
   #
@@ -14,9 +22,11 @@ module OpenGraphReader
       @connection = Faraday.default_connection.dup
 
       if defined? FaradayMiddleware
-        unless @connection.builder.handlers.include? FaradayMiddleware::FollowRedirects
-          @connection.builder.insert(0, FaradayMiddleware::FollowRedirects)
-        end
+        prepend_middleware FaradayMiddleware::FollowRedirects
+      end
+
+      if defined? Faraday::CookieJar
+        prepend_middleware Faraday::CookieJar
       end
     end
 
@@ -78,5 +88,13 @@ module OpenGraphReader
     def fetched_headers?
       !@get_response.nil? || !@head_response.nil?
     end
+
+    private
+
+    def prepend_middleware middleware
+      unless @connection.builder.handlers.include? middleware
+          @connection.builder.insert(0, middleware)
+      end
+    end
   end
 end

data/lib/open_graph_reader/object.rb

@@ -71,7 +71,7 @@ module OpenGraphReader
     # @return [String, Object]
     def [] name
       raise InvalidObjectError, "Undefined property #{name} on #{inspect}" unless has_property? name
-      properties[name.to_s]
+      public_send name.to_s #properties[name.to_s]
     end
 
     # Set the property to the given value.
@@ -81,8 +81,11 @@ module OpenGraphReader
     # @param [String, Object] value
     # @raise [InvalidObjectError] If the requested property is undefined.
     def []= name, value
-      raise InvalidObjectError, "Undefined property #{name} on #{inspect}" unless has_property? name
-      public_send "#{name}=", value
+      if has_property?(name)
+        public_send "#{name}=", value
+      elsif OpenGraphReader.config.strict
+        raise InvalidObjectError, "Undefined property #{name} on #{inspect}"
+      end
     end
 
     # Returns {#content} if available.

data/lib/open_graph_reader/object/dsl.rb

@@ -14,6 +14,7 @@ module OpenGraphReader
       #   @option options [Class] :to This property maps to the given object (optional).
       #     belongs to the given verticals of the object (optional).
       #   @option options [Array<String>] :verticials This property
+      #   @option options [Bool] :downcase (false) Normalize the contents case to lowercase.
       #
       # @!macro property
       #   @!attribute [rw] $1
@@ -32,15 +33,18 @@ module OpenGraphReader
         processors[name] = processor
 
         define_method(name) do |name, *args|
-          available_properties << name.to_s
           options = args.pop if args.last.is_a? Hash
           options ||= {}
 
+          available_properties << name.to_s
+          required_properties << name.to_s if options[:required]
           Registry.register [@namespace, name].join(':'), options[:to] if options[:to]
 
           if options[:verticals]
             options[:verticals].each do |vertical|
-              verticals[[@namespace, vertical].join('.')] << name
+              vertical = [@namespace, vertical].join('.')
+              verticals[vertical] << name.to_s
+              Registry.verticals <<  vertical
             end
           end
 
@@ -50,21 +54,22 @@ module OpenGraphReader
             end
 
             define_method(name) do
-              # TODO raise if required
               value = children[name.to_s].first
-              # TODO: figure out a sane way to distinguish subobject properties
+              # @todo figure out a sane way to distinguish subobject properties
               value.content if value && value.is_a?(Object)
               value || options[:default]
             end
           else
             define_method(name) do
-              # TODO raise if required
               properties[name.to_s] || options[:default]
             end
 
             define_method("#{name}=") do |value|
-              # TODO: figure out a sane way to distinguish subobject properties
-              value = processor.call(value, *args, options) unless value.is_a? Object
+              # @todo figure out a sane way to distinguish subobject properties
+              unless value.is_a? Object
+                value.downcase! if options[:downcase]
+                value = processor.call(value, *args, options)
+              end
               properties[name.to_s] = value
             end
           end
@@ -88,11 +93,23 @@ module OpenGraphReader
         Registry.register @namespace, self
       end
 
-      # Set the type for the content attribute
+      # @overload content type, *args, options={}
       #
-      # @param [Symbol] type one of the registered types.
-      def content type
-        @content_processor = DSL.processors[type]
+      #   Set the type for the content attribute
+      #
+      #   @param [Symbol] type one of the registered types.
+      #   @param [Array<Object>] args Additional parameters for the type
+      #   @param [Hash] options
+      #   @option options [Bool] :downcase (false) Normalize the contents case to lowercase.
+      def content type, *args
+        options = args.pop if args.last.is_a? Hash
+        options ||= {}
+
+        @content_processor = proc {|value|
+          value.downcase! if options[:downcase]
+          options[:to] ||= self
+          DSL.processors[type].call(value, *args, options)
+        }
       end
 
       # The list of defined properties on this object.
@@ -102,6 +119,13 @@ module OpenGraphReader
         @available_properties ||= []
       end
 
+      # The list of required properties on this object.
+      #
+      # @return [Array<String]
+      def required_properties
+        @required_properties ||= []
+      end
+
       # A map from type names to processing blocks.
       #
       # @api private
@@ -115,7 +139,7 @@ module OpenGraphReader
       # @api private
       # @return [Proc]
       def content_processor
-        @content_processor || proc {|value| value }
+        @content_processor
       end
 
       # A map from vertical names to attributes that belong to them.