hangry 0.0.6 → 0.0.7

data/.gitignore

@@ -15,3 +15,4 @@ spec/reports
 test/tmp
 test/version_tmp
 tmp
+.DS_Store

data/.travis.yml

@@ -2,5 +2,5 @@ language: ruby
 rvm:
   - 2.0.0
   - 1.9.3
-  - jruby-19mode # JRuby in 1.9 mode
+  #- jruby-19mode # JRuby in 1.9 mode - disabled for now due to differences in JRuby Nokogiri results...
   - rbx-19mode

data/README.md

@@ -7,7 +7,9 @@ Currently supported microformats:
 - http://data-vocabulary.org/Recipe
 - http://microformats.org/wiki/hrecipe
 
-## Build Status [![Build Status](https://travis-ci.org/iancanderson/hangry.png?branch=master)](http://travis-ci.org/iancanderson/hangry)
+## Pieces of Flair
+- [![Build Status](https://travis-ci.org/iancanderson/hangry.png?branch=master)](http://travis-ci.org/iancanderson/hangry)
+- [![Code Climate](https://codeclimate.com/github/iancanderson/hangry.png)](https://codeclimate.com/github/iancanderson/hangry)
 
 ## Installation
 
@@ -45,8 +47,6 @@ recipe.yield          # "4 servings"
 # etc..
 ```
 
-Here's a live demo: http://hangry.herokuapp.com/
-
 ## Contributing
 
 1. Fork it

data/lib/hangry/canonical_url_parser.rb

@@ -0,0 +1,31 @@
+module Hangry
+  class CanonicalUrlParser
+
+    attr_accessor :nokogiri_doc
+
+    def initialize(html_or_nokogiri_doc)
+      self.nokogiri_doc = case html_or_nokogiri_doc
+      when String
+        Nokogiri::HTML(html_or_nokogiri_doc)
+      when Nokogiri::HTML::Document
+        html_or_nokogiri_doc
+      else
+        raise ArgumentError
+      end
+    end
+
+    def canonical_domain
+      canonical_url ? URI.parse(canonical_url).host : nil
+    end
+
+    def canonical_url
+      node = nokogiri_doc.css('link[rel="canonical"]').first
+      return node['href'] if node
+
+      # Fall back to open graph URL (see food network example)
+      node = nokogiri_doc.css('meta[property="og:url"]').first
+      node ? node['content'] : nil
+    end
+
+  end
+end

data/lib/hangry/default_recipe_parser.rb

@@ -0,0 +1,14 @@
+module Hangry
+  class DefaultRecipeParser < RecipeParser
+    def self.can_parse?(html)
+      true
+    end
+
+    def initialize(html)
+    end
+
+    def parse
+      Recipe.new
+    end
+  end
+end

data/lib/hangry/parser_class_selecter.rb

@@ -0,0 +1,23 @@
+require 'hangry/recipe_parser'
+require 'hangry/default_recipe_parser'
+require 'hangry/hrecipe_parser'
+require 'hangry/schema_org_recipe_parser'
+require 'hangry/data_vocabulary_recipe_parser'
+require 'hangry/parsers/non_standard/all_recipes_parser'
+
+module Hangry
+  class ParserClassSelecter
+    def initialize(html)
+      @html = html
+    end
+
+    def parser_class
+      # Prefer the more specific parsers
+      parser_classes = [Parsers::NonStandard::AllRecipesParser]
+      parser_classes += [SchemaOrgRecipeParser, HRecipeParser, DataVocabularyRecipeParser]
+      parser_classes << DefaultRecipeParser
+      parser_classes.detect { |p| p.can_parse?(@html) }
+    end
+    
+  end
+end

data/lib/hangry/parsers/non_standard/all_recipes_parser.rb

@@ -0,0 +1,18 @@
+module Hangry
+  module Parsers
+    module NonStandard
+      class AllRecipesParser < SchemaOrgRecipeParser
+
+        def self.can_parse?(html)
+          canonical_url_matches_domain?(html, 'allrecipes.com')
+        end
+
+        def parse_instructions
+          content = recipe_ast.css('.directions ol').first.content
+          clean_string content, preserve_newlines: true
+        end
+
+      end
+    end
+  end
+end

data/lib/hangry/recipe_parser.rb

@@ -1,14 +1,16 @@
+require 'hangry/canonical_url_parser'
+
 module Hangry
   class RecipeParser
     attr_reader :recipe_html
-    attr_accessor :nutrition_ast, :recipe_ast, :recipe
+    attr_accessor :nokogiri_doc, :nutrition_ast, :recipe_ast, :recipe
 
     def initialize(recipe_html)
       @recipe_html = recipe_html
       @recipe = Recipe.new
       initialize_nutrition
-      doc = Nokogiri::HTML(recipe_html)
-      self.recipe_ast = doc.css(self.class.root_selector).first
+      self.nokogiri_doc = Nokogiri::HTML(recipe_html)
+      self.recipe_ast = nokogiri_doc.css(self.class.root_selector).first
       self.nutrition_ast = recipe_ast && recipe_ast.css(self.class.nutrition_selector)
     end
 
@@ -20,8 +22,12 @@ module Hangry
       recipe
     end
 
-    def can_parse?
-      recipe_ast
+    def self.can_parse?(html)
+      new(html).recipe_ast
+    end
+
+    def self.canonical_url_matches_domain?(html, domain)
+      CanonicalUrlParser.new(html).canonical_domain == domain
     end
 
     private
@@ -67,6 +73,10 @@ module Hangry
       end
     end
 
+    def parse_canonical_url
+      CanonicalUrlParser.new(nokogiri_doc).canonical_url
+    end
+
     def parse_duration(iso8601_string)
       duration = ISO8601::Duration.new(iso8601_string)
       duration.hours.to_i * 60 + duration.minutes.to_i

data/lib/hangry/version.rb

@@ -1,3 +1,3 @@
 module Hangry
-  VERSION = "0.0.6"
+  VERSION = "0.0.7"
 end

data/lib/hangry.rb

@@ -1,8 +1,5 @@
 require "hangry/version"
-require 'hangry/recipe_parser'
-require 'hangry/hrecipe_parser'
-require 'hangry/schema_org_recipe_parser'
-require 'hangry/data_vocabulary_recipe_parser'
+require 'hangry/parser_class_selecter'
 require 'active_support/core_ext/object/blank'
 require 'date'
 require 'iso8601'
@@ -11,6 +8,7 @@ require "nokogiri"
 module Hangry
   RECIPE_ATTRIBUTES = [
     :author,
+    :canonical_url,
     :cook_time,
     :description,
     :ingredients,
@@ -40,11 +38,8 @@ module Hangry
   Recipe = Struct.new(*RECIPE_ATTRIBUTES)
 
   def self.parse(html)
-    parser_classes = [SchemaOrgRecipeParser, HRecipeParser, DataVocabularyRecipeParser]
-    parsers = parser_classes.map { |klass| klass.new(html) }
-    parser = parsers.detect { |p| p.can_parse? }
-
-    parser ? parser.parse : Recipe.new
+    parser_class = ParserClassSelecter.new(html).parser_class
+    parser_class.new(html).parse
   end
 
 end

data/spec/real_examples/all_recipes_spec.rb

@@ -3,9 +3,15 @@ require 'hangry'
 describe Hangry do
 
   context "allrecipes.com recipe" do
-    subject { Hangry.parse(File.read("spec/fixtures/allrecipes.html")) }
+    let(:html) { File.read("spec/fixtures/allrecipes.html") }
+    subject { Hangry.parse(html) }
+
+    it "should use the correct parser" do
+      Hangry::ParserClassSelecter.new(html).parser_class.should == Hangry::Parsers::NonStandard::AllRecipesParser
+    end
     
     its(:author) { should == "United Soybean Board" }
+    its(:canonical_url) { should == "http://allrecipes.com/recipe/roasted-vegetable-and-couscous-salad/" }
     its(:cook_time) { should == 15 }
     its(:description) { should == "\"This better-for-you main-dish salad is quick, colorful and full of satisfying texture. To explore a variety of grains, substitute 3 cups cooked regular couscous, brown rice or quinoa.\"" }
     its(:ingredients) {
@@ -36,7 +42,17 @@ describe Hangry do
         unsaturated_fat: nil
       }
     end
-    its(:instructions) { should == nil }
+    its(:instructions) {
+      instructions = <<-EOS
+Preheat oven to 425 degrees F.
+Toss broccoli, peppers and onions with 2 tablespoons soybean oil and 1/4 teaspoon each salt and pepper. Place on foil-lined baking sheet.
+Bake for 15 minutes until vegetables are tender and lightly browned.
+Meanwhile, cook couscous according to package directions.
+Place cooked couscous and roasted vegetables in large bowl. Pour vinegar and remaining soybean oil over salad and sprinkle with remaining salt and pepper; toss lightly until combined.
+You may substitute 3 cups cooked regular couscous, brown rice or quinoa.
+      EOS
+      should == instructions.strip
+    }
     its(:prep_time) { should == 15 }
     its(:published_date) { should == nil }
     its(:total_time) { should == 30 }

data/spec/real_examples/big_oven_spec.rb

@@ -10,6 +10,7 @@ describe Hangry do
       # BigOven puts the author element outside of the hRecipe element...
       should == nil
     end
+    its(:canonical_url) { should == "http://www.bigoven.com/recipe/178920/steves-fish-tacos" }
     its(:cook_time) { should == nil }
     its(:description) { should == "I had never tried fish tacos until my son, fresh out of boot camp, asked me to make them. I found a basic recipe, then adapted it from there, and now it's one of my favorite things to eat!" }
     its(:ingredients) {

data/spec/real_examples/epicurious_spec.rb

@@ -6,6 +6,7 @@ describe Hangry do
     subject { Hangry.parse(File.read("spec/fixtures/epicurious.html")) }
     
     its(:author) { should == "by Janet Taylor McCracken" }
+    its(:canonical_url) { should == "http://www.epicurious.com/articlesguides/bestof/toprecipes/bestburgerrecipes/recipes/food/views/Grilled-Turkey-Burgers-with-Cheddar-and-Smoky-Aioli-354289" }
     its(:cook_time) { should == nil }
     its(:description) { should == "A simple Moroccan-spiced aioli is mixed in with the ground turkey to keep the burgers moist and give them tons of flavor. Smoked paprika is available in the spice aisle of most supermarkets." }
     its(:prep_time) { should == 40 }

data/spec/real_examples/food_network_spec.rb

@@ -5,6 +5,7 @@ describe Hangry do
   context "food network recipe" do
     subject { Hangry.parse(File.read("spec/fixtures/food_network_schema_org.html")) }
     
+    its(:canonical_url) { should == 'http://www.foodnetwork.com/recipes/rachael-ray/spinach-and-mushroom-stuffed-chicken-breasts-recipe/index.html' }
     its(:cook_time) { should == 20 }
     its(:description) { should == nil }
     its(:ingredients) do

data/spec/real_examples/myrecipes_com_spec.rb

@@ -7,6 +7,7 @@ describe Hangry do
     subject { Hangry.parse(File.read("spec/fixtures/myrecipes.com.html")) }
     
     its(:author) { should == "Southern Living" }
+    its(:canonical_url) { should == 'http://www.myrecipes.com/recipe/best-carrot-cake-10000000257583/' }
     its(:cook_time) { should == nil }
     its(:description) { should == "" } # not valid HTML to have a p inside of an h2...
     its(:ingredients) {

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: hangry
 version: !ruby/object:Gem::Version
-  version: 0.0.6
+  version: 0.0.7
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-03-11 00:00:00.000000000Z
+date: 2013-03-25 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &70141407538320 !ruby/object:Gem::Requirement
+  requirement: &70314683800080 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70141407538320
+  version_requirements: *70314683800080
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &70141407537900 !ruby/object:Gem::Requirement
+  requirement: &70314683799600 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,10 +32,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70141407537900
+  version_requirements: *70314683799600
 - !ruby/object:Gem::Dependency
   name: pry
-  requirement: &70141407537460 !ruby/object:Gem::Requirement
+  requirement: &70314683799140 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,10 +43,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70141407537460
+  version_requirements: *70314683799140
 - !ruby/object:Gem::Dependency
   name: activesupport
-  requirement: &70141407536960 !ruby/object:Gem::Requirement
+  requirement: &70314683798640 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -54,10 +54,10 @@ dependencies:
         version: '3.0'
   type: :runtime
   prerelease: false
-  version_requirements: *70141407536960
+  version_requirements: *70314683798640
 - !ruby/object:Gem::Dependency
   name: iso8601
-  requirement: &70141407536420 !ruby/object:Gem::Requirement
+  requirement: &70314683798120 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -65,10 +65,10 @@ dependencies:
         version: 0.4.0
   type: :runtime
   prerelease: false
-  version_requirements: *70141407536420
+  version_requirements: *70314683798120
 - !ruby/object:Gem::Dependency
   name: nokogiri
-  requirement: &70141407535940 !ruby/object:Gem::Requirement
+  requirement: &70314683797640 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -76,7 +76,7 @@ dependencies:
         version: '1.5'
   type: :runtime
   prerelease: false
-  version_requirements: *70141407535940
+  version_requirements: *70314683797640
 description: A recipe microformat parser.
 email:
 - ian@iancanderson.com
@@ -92,8 +92,12 @@ files:
 - Rakefile
 - hangry.gemspec
 - lib/hangry.rb
+- lib/hangry/canonical_url_parser.rb
 - lib/hangry/data_vocabulary_recipe_parser.rb
+- lib/hangry/default_recipe_parser.rb
 - lib/hangry/hrecipe_parser.rb
+- lib/hangry/parser_class_selecter.rb
+- lib/hangry/parsers/non_standard/all_recipes_parser.rb
 - lib/hangry/recipe_parser.rb
 - lib/hangry/schema_org_recipe_parser.rb
 - lib/hangry/version.rb