recipe_scraper 2.2.1 → 2.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: ad21bfeb82ade7c69ec91efb87659fbf0f43ead3
-  data.tar.gz: f6a3a569c869c41213fca36f213fbb5efc47b48a
+  metadata.gz: e025c453bfd951c2b67abaf3d27a269af4d6dd00
+  data.tar.gz: f32ea1856cdc9caea8caed07153d76450dc94ca3
 SHA512:
-  metadata.gz: b1c6991149659d6003ac2955340ed71563fc8c8e56609bb3442f82bdf56c8758017e668c33411998bd68bf54436eb3bed123cdd43ae77d62ef0873c9dd59971e
-  data.tar.gz: e649b6171fce236dd2829c26aa75d8d445945266fe6c527bd468abce1edb33a198f0b5f35371bf7f13ee9d101f192a2ac317cb38381f0aeadee00e3c4ba991fa
+  metadata.gz: 6d2ab974e447358d9f2f90476caea4f3cee4236c17a5dceef9e3a470ad89653c121297b725fe646f71a7dc3bbc54750a36fe90dd076952d52b4fbf704158aff4
+  data.tar.gz: 2aaff08dbb6c1a794d4e0d9e9ff009832802dca967d104f1b786d42e5ed90e0f6840ecf5d7071e743d8434b14989a10b7340d0069fc8923bb18e50926e250280

data/lib/recipe_scraper.rb

@@ -18,7 +18,7 @@ module RecipeScraper
     CUISINEAZ_HOST = {desktop: 'http://www.cuisineaz.com/'}
 
 
-    
+
     # Instanciate a Recipe object with data crawled from an url
     #
     # @param url [String] representing an url from Marmiton or 750g website
@@ -34,7 +34,7 @@ module RecipeScraper
         fetch_from_cuisineaz url
 
       else
-        raise ArgumentError, "Instantiation cancelled (Host not supported)." 
+        raise ArgumentError, "Instantiation cancelled (Host not supported)."
       end
     end
 
@@ -68,7 +68,7 @@ module RecipeScraper
     # @param text [String] a text to sanitize
     # @return [String] as text corrected formated
     def sanitize text
-      ['  ', '\r\n', "\r\n", "\n", "\r"].each { |text_to_remove| 
+      ['  ', '\r\n', "\r\n", "\n", "\r", "\t", / ^/, / $+/, /^  /, /^ /].each { |text_to_remove|
         text.gsub!(text_to_remove,'')
       }
       return text
@@ -111,29 +111,30 @@ module RecipeScraper
         url.gsub! MARMITON_HOST[:mobile], MARMITON_HOST[:desktop]
 
         page =  Nokogiri::HTML(open(url).read)
-        @title = page.css('h1.m_title span.item span.fn').text
+        @title = page.css('h1').text
 
 
         # get times
-        @preptime = page.css('p.m_content_recette_info span.preptime').text.to_i
-        @cooktime = page.css('p.m_content_recette_info span.cooktime').text.to_i
+        @preptime = page.css('div.recipe-infos__timmings__preparation > span.recipe-infos__timmings__value').text.to_i
+        @cooktime = page.css('div.recipe-infos__timmings__cooking > span.recipe-infos__timmings__value').text.to_i
 
         # get ingredients
-        ingredients_text = page.css('div.m_content_recette_ingredients').text
-        @ingredients = sanitize(ingredients_text).split '- '
-        @ingredients.delete_at(0) # to delete the first `Ingrédients (pour 2 personnes) :`
+        @ingredients = []
+        ingredients_text = page.css('ul.recipe-ingredients__list li.recipe-ingredients__list__item').each do |ingredient_tag|
+          @ingredients << sanitize(ingredient_tag.text)
+        end
 
         # get steps
-        steps_text = page.css('div.m_content_recette_todo').text
-        @steps = sanitize(steps_text).split '. '
-        @steps.delete_at(0) # to delete the first `Ingrédients (pour 2 personnes) :`
+        @steps = []
+        steps_text = page.css('ol.recipe-preparation__list').each do |step_tag|
+          @steps << sanitize(step_tag.text)
+        end
 
         # get image
-        @image = page.css('a.m_content_recette_illu img.m_pinitimage').attr('src').to_s
-        
+        @image = page.css('#af-diapo-desktop-0_img').attr('src').to_s rescue NoMethodError
 
       else
-        raise ArgumentError, "Instantiation cancelled (ulr not from #{MARMITON_HOST})." 
+        raise ArgumentError, "Instantiation cancelled (ulr not from #{MARMITON_HOST})."
       end
     end
 
@@ -147,8 +148,8 @@ module RecipeScraper
         @title = page.css('h1.c-article__title').text
 
         # get times
-        @preptime = page.css('ul.c-recipe-summary li time[itemprop=prepTime]').text.to_i
-        @cooktime = page.css('ul.c-recipe-summary li time[itemprop=cookTime]').text.to_i
+        @preptime = sanitize(page.css('ul.c-recipe-summary > li.c-recipe-summary__rating[title="Temps de préparation"]').text).to_i
+        @cooktime = sanitize(page.css('ul.c-recipe-summary > li.c-recipe-summary__rating[title="Temps de cuisson"]').text).to_i
 
         @steps = []
         css_step = "div[itemprop=recipeInstructions] p"
@@ -166,9 +167,9 @@ module RecipeScraper
           @image = page.css(css_image).attr('src').to_s
         rescue NoMethodError => e
         end
-        
+
       else
-        raise ArgumentError, "Instantiation cancelled (ulr not from #{G750_HOST})." 
+        raise ArgumentError, "Instantiation cancelled (ulr not from #{G750_HOST})."
       end
     end
 
@@ -185,27 +186,26 @@ module RecipeScraper
         @preptime = page.css('#ctl00_ContentPlaceHolder_LblRecetteTempsPrepa').text.to_i
         @cooktime = page.css('#ctl00_ContentPlaceHolder_LblRecetteTempsCuisson').text.to_i
 
-        
         @steps = []
-        page.css("#preparation span p.fs17").each { |step_node|
+        page.css("#preparation p").each { |step_node|
           @steps << sanitize(step_node.text)
         }
 
         @ingredients = []
-        page.css("#ingredients li").each { |ing_node|
+        page.css("section.recipe_ingredients li").each { |ing_node|
           @ingredients << sanitize(ing_node.text)
         }
 
         begin
-          @image = page.css('#shareimg').attr('src').to_s
+          @image = page.css('#ctl00_ContentPlaceHolder_recipeImgLarge').attr('src').to_s
         rescue NoMethodError => e
         end
-        
+
       else
-        raise ArgumentError, "Instantiation cancelled (ulr not from #{G750_HOST})." 
+        raise ArgumentError, "Instantiation cancelled (ulr not from #{G750_HOST})."
       end
     end
-  
+
   end
 
 end

data/lib/recipe_scraper/version.rb

@@ -1,3 +1,3 @@
 module RecipeScraper
-  VERSION = "2.2.1"
+  VERSION = "2.2.2"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: recipe_scraper
 version: !ruby/object:Gem::Version
-  version: 2.2.1
+  version: 2.2.2
 platform: ruby
 authors:
 - madeindjs
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2016-12-07 00:00:00.000000000 Z
+date: 2017-07-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -107,7 +107,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.5.1
+rubygems_version: 2.6.11
 signing_key: 
 specification_version: 4
 summary: A web scrawler to get a Marmiton's, cuisineaz or 750g recipe