preadly-bulbasaur 0.4.0 → 0.5.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: e2ce49896cc75739b7b348c8562bd27b782ae050
-  data.tar.gz: 8e085b9b873fe13483f5885bf4c5a26d7f4171b8
+  metadata.gz: 014f0ead8b757383f291bc31f15134e1c325032d
+  data.tar.gz: e3299a0a0b906620a99240065456189bebc08aee
 SHA512:
-  metadata.gz: 36409b5eda6f8928b710698fbb7787fa387f59276f08a724e448a333e8b259096efd30cff1b8632b0e77401a3194c7b7f2a42c2b3fd64a44e20c649b3dc1be1a
-  data.tar.gz: bedb92be982705d050f8ea472410a34af96f4df36a528d370a593cf6087246717a4f2bc3e08f2eac34473cd71d0433d7920482d42a96bd639630dcbef75940ea
+  metadata.gz: 7b18fbb448369cafdce3789bd9a4c61ec87a198118908fddfdef685c828c65331da6972701d3b713a86abcffde31ce5af0b446ee3abef285aa3df41bfe88affe
+  data.tar.gz: 37ae03db04175312915da5cd8e9511f3db4f9c56dcebf87d3c3385fb2573b90dd832a73f1cb1a43dec3d3a9b25397a7ffebf4f1bc06898be209fc8a30e14caad

data/lib/bulbasaur/extracts/extract_images_from_html.rb

@@ -3,7 +3,7 @@ module Bulbasaur
   class ExtractImagesFromHTML
     
     CSS_IMPORT_URL_REGEX = /(?<=url\()['"]?.+?['"]?.+?(?=\))/
-    IMG_CANDIDATE_URL_REGEX = /https?:\/\/\S*\.(?:png|jpg|jpeg)(?!\.\S)/i
+    IMG_CANDIDATE_URL_REGEX = /https?:\/\/\S*\.(?:png|jpg|jpeg|gif)(?!\.\S)/i
 
     def initialize(html)
       @html = html
@@ -13,6 +13,7 @@ module Bulbasaur
       images = Array.new
       images = images + extract_images_by_tag_image
       images = images + extract_images_by_tag_style
+      images = images + extract_images_by_link
       images
     end
 
@@ -36,6 +37,15 @@ module Bulbasaur
       images
     end
 
+    def extract_images_by_link
+      images = Array.new
+      Nokogiri::HTML(@html).xpath('//a').each do |link|
+        url = link.xpath('@href').text
+        images << create_struct(url) if url =~ IMG_CANDIDATE_URL_REGEX
+      end
+      images
+    end
+
     def create_struct(url, alt=nil)
       {url: url, alt: alt } 
     end

data/lib/bulbasaur/version.rb

@@ -2,7 +2,7 @@ module Bulbasaur
 
   module Version
     MAJOR = 0
-    MINOR = 4
+    MINOR = 5
     PATCH = 0
     STRING = "#{MAJOR}.#{MINOR}.#{PATCH}"
   end

data/spec/bulbasaur/extracts/extract_images_from_html_spec.rb

@@ -60,6 +60,44 @@ RSpec.describe Bulbasaur::ExtractImagesFromHTML do
         expect(subject.first[:alt]).to be_nil
       end
     end
+
+    context 'When sending HTML with a link pointing to an image' do
+      let(:html) do
+        '<p>
+          <a href="http://somewhere.to/get/the_original_image.jpg">
+            Click here to see the original image.
+          </a>
+          <a href="http://somewhere.to/get/the_original_image.jpg?width=400&height=400">
+            Click here to see the original image.
+          </a>
+          <a href="http://somewhere.to/go/to/another_page.html">
+            Click here to go to another page.
+          </a>
+        </p>'
+      end
+
+      it 'Does return an image array with 1 item' do
+        expect(subject.size).to eq 2
+      end
+
+      it 'Does return the image URL with parameters' do
+        expect(subject.last[:url]).to eq 'http://somewhere.to/get/the_original_image.jpg?width=400&height=400'
+      end
+
+      it 'Does return the image URL without parameters' do
+        expect(subject.first[:url]).to eq 'http://somewhere.to/get/the_original_image.jpg'
+      end
+
+      it 'Does return the image alt' do
+        expect(subject.first[:alt]).to be_nil
+      end
+
+      it 'Does not include links other than for images' do
+        expect(subject).to include Hash(url: 'http://somewhere.to/get/the_original_image.jpg', alt: nil)
+        expect(subject).to include Hash(url: 'http://somewhere.to/get/the_original_image.jpg?width=400&height=400', alt: nil)
+        expect(subject).not_to include Hash(url: 'http://somewhere.to/go/to/another_page.html', alt: nil)
+      end
+    end
     
     context "When send html with many images" do
       

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: preadly-bulbasaur
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 0.5.0
 platform: ruby
 authors:
 - Magno Costa
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-07-16 00:00:00.000000000 Z
+date: 2015-08-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -126,7 +126,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.4.6
+rubygems_version: 2.2.2
 signing_key: 
 specification_version: 4
 summary: Bulbasaur is a helper for crawler operations used in Pread.ly
@@ -140,4 +140,3 @@ test_files:
 - spec/bulbasaur/utils/normalize_url_spec.rb
 - spec/bulbasaur_spec.rb
 - spec/spec_helper.rb
-has_rdoc: