linkser 0.0.2 → 0.0.3

data/lib/linkser/object.rb

@@ -0,0 +1,10 @@
+module Linkser
+  class Object
+    attr_reader :url, :head
+    
+    def initialize url, head, options={}
+      @url = url
+      @heade = head
+    end
+  end
+end

data/lib/linkser/objects/html.rb

@@ -0,0 +1,121 @@
+require 'nokogiri'
+require 'open-uri'
+require 'net/http'
+require 'image_spec'
+require 'opengraph'
+
+module Linkser
+  module Objects
+    class HTML < Linkser::Object
+      attr_reader :body, :nokogiri
+      attr_reader :title, :description, :images, :ogp
+      def initialize url, head, options={}
+        super url, head, options
+      end
+
+      def title
+        return @title unless @title.nil?
+        if ogp and ogp.title
+           @title = ogp.title
+        else
+          nokogiri.css('title').each do |title|
+            @title = title.text
+          end
+        end          
+        @title
+      end
+
+      def body
+        return @body unless @body.nil?
+        @body = open(url)
+      end
+
+      def description
+        return @description unless @description.nil?
+        if ogp and ogp.description
+           @description = ogp.description
+        else
+          nokogiri.css('meta').each do |meta|
+            if meta.get_attribute("name").eql? "description"
+              @description = meta.get_attribute("content")
+            end
+          end
+        end
+        @description
+      end
+
+      def images
+        return @images unless @images.nil?
+        @images = Array.new
+        if ogp and ogp.image
+          begin
+            img_spec = ImageSpec.new(ogp.image)
+            if valid_img? img_spec.width.to_f, img_spec.height.to_f
+              @images << {:img => ogp.image, :width => img_spec.width, :height => img_spec.height}
+            end
+          rescue
+          end
+        end        
+        nokogiri.css('img').each do |img|
+          img_src = img.get_attribute("src")
+          img_src = complete_url img_src, url
+          img_uri = URI.parse(img_src)
+          img_ext = File.extname(img_uri.path)
+          img_name = File.basename(img_uri.path,img_ext)
+          if [".jpg", ".jpeg", ".png"].include? img_ext
+            begin
+              img_spec = ImageSpec.new(img_src)
+              if valid_img? img_spec.width.to_f, img_spec.height.to_f
+                @images << {:img => img_src, :width => img_spec.width, :height => img_spec.height}
+              end
+            rescue
+            end
+          end
+        end
+        @images
+      end      
+
+      def nokogiri
+        return @nokogiri unless @nokogiri.nil?
+        @nokogiri = Nokogiri::HTML(body)
+      end
+
+      def ogp
+        return @ogp unless @ogp.nil?
+        @ogp = OpenGraph::Object.new
+        nokogiri.css('meta').each do |m|
+          if m.attribute('property') && m.attribute('property').to_s.match(/^og:(.+)$/i)
+            @ogp[$1.gsub('-','_')] = m.attribute('content').to_s
+          end
+        end
+        @ogp = false if @ogp.keys.empty?
+        @ogp = false unless @ogp.valid?  
+        @ogp 
+      end
+
+      private
+
+      def complete_url src, url
+        uri = URI.parse(url)
+        base_url = "http://" + uri.host + (uri.port!=80 ? ":" + uri.port.to_s : "")
+        relative_url = "http://" + uri.host + (uri.port!=80 ? ":" + uri.port.to_s : "") + uri.path
+        if src.index("http://")==0
+          src
+        elsif src.index("/")==0
+          base_url + src
+        else
+          relative_url + src
+        end
+      end
+
+      def valid_img? w, h
+        if w > 199 or w > 199
+          if ((w > 0 and h > 0 and ((w / h) < 3) and ((w / h) > 0.2)) or (w > 0 and h == 0 and w < 700) or (w == 0 and h > 0 and h < 700))
+          return true
+          end
+        end
+        false
+      end
+    end
+  end
+end

data/lib/linkser/parser.rb

@@ -5,23 +5,21 @@ module Linkser
   module Parser
     def self.parse url, options={}
       if !is_valid_url? url
-      raise "Invalid URL"
+        raise "Invalid URL"
       end
       head = get_head url
       case head.content_type
       when "text/html"
-        Linkser::Parser::HTML.new.parse url
+        Linkser::Objects::HTML.new url, head
       else
       raise "I have no idea on how to parse a '" + head.content_type + "'"
       end
     end
 
-    #private
+    private
 
     def self.get_head url, limit = 10
-      if (limit==0)
-      raise 'Too many HTTP redirects. URL was not reacheable within the HTTP redirects limit'
-      end
+      raise 'Too many HTTP redirects. URL was not reacheable within the HTTP redirects limit' if (limit==0)
       uri = URI.parse url
       http = Net::HTTP.start uri.host, uri.port
       response = http.head uri.request_uri
@@ -33,7 +31,7 @@ module Linkser
         warn "Redirecting to #{location}"
         return get_head location, limit - 1
       else
-      raise 'The HTTP responded with an ' + response.code + ' code'
+      raise 'The HTTP request has a ' + response.code + ' code'
       end
     end
 
@@ -41,10 +39,11 @@ module Linkser
       begin
         uri = URI.parse(url)
         if [:scheme, :host].any? { |i| uri.send(i).blank? }
-        raise(URI::InvalidURIError)
+        raise URI::InvalidURIError 
         end
         return true
       rescue URI::InvalidURIError => e
+         warn e.to_s
       return false
       end
     end

data/lib/linkser/version.rb

@@ -1,3 +1,3 @@
 module Linkser
-  VERSION = "0.0.2"
+  VERSION = "0.0.3"
 end

data/lib/linkser.rb

@@ -2,8 +2,9 @@ require 'linkser/version'
 
 module Linkser
   autoload :Parser, 'linkser/parser'
-  module Parser
-    autoload :HTML, 'linkser/parser/html'
+  autoload :Object, 'linkser/object'
+  module Objects
+    autoload :HTML, 'linkser/objects/html'
   end
 end
 

data/linkser.gemspec

@@ -24,6 +24,7 @@ Gem::Specification.new do |s|
   s.add_runtime_dependency('nokogiri', '~> 1.4.2')
   s.add_runtime_dependency('rmagick', '~> 2.13.1')
   s.add_runtime_dependency('ruby-imagespec', "~> 0.2.0")  
+  s.add_runtime_dependency('opengraph', "~> 0.0.4")
   
   # Development Gem dependencies
   #

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: linkser
 version: !ruby/object:Gem::Version 
-  hash: 27
+  hash: 25
   prerelease: 
   segments: 
   - 0
   - 0
-  - 2
-  version: 0.0.2
+  - 3
+  version: 0.0.3
 platform: ruby
 authors: 
 - Eduardo Casanova
@@ -15,8 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-11-18 00:00:00 +01:00
-default_executable: 
+date: 2011-11-21 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: rake
@@ -81,9 +80,25 @@ dependencies:
   type: :runtime
   version_requirements: *id004
 - !ruby/object:Gem::Dependency 
-  name: ruby-debug
+  name: opengraph
   prerelease: false
   requirement: &id005 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ~>
+      - !ruby/object:Gem::Version 
+        hash: 23
+        segments: 
+        - 0
+        - 0
+        - 4
+        version: 0.0.4
+  type: :runtime
+  version_requirements: *id005
+- !ruby/object:Gem::Dependency 
+  name: ruby-debug
+  prerelease: false
+  requirement: &id006 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -95,11 +110,11 @@ dependencies:
         - 3
         version: 0.10.3
   type: :development
-  version_requirements: *id005
+  version_requirements: *id006
 - !ruby/object:Gem::Dependency 
   name: rspec
   prerelease: false
-  requirement: &id006 !ruby/object:Gem::Requirement 
+  requirement: &id007 !ruby/object:Gem::Requirement 
     none: false
     requirements: 
     - - ">="
@@ -111,7 +126,7 @@ dependencies:
         - 0
         version: 2.7.0
   type: :development
-  version_requirements: *id006
+  version_requirements: *id007
 description: Linkser is a link parser for Ruby. It gets an URI, tries to dereference it and returns the relevant information about the resource.
 email: 
 - ecasanovac@gmail.com
@@ -130,13 +145,13 @@ files:
 - README.textile
 - Rakefile
 - lib/linkser.rb
+- lib/linkser/object.rb
+- lib/linkser/objects/html.rb
 - lib/linkser/parser.rb
-- lib/linkser/parser/html.rb
 - lib/linkser/version.rb
 - linkser.gemspec
 - spec/linkser_spec.rb
 - spec/spec_helper.rb
-has_rdoc: true
 homepage: https://github.com/ging/linkser
 licenses: []
 
@@ -166,7 +181,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.6.2
+rubygems_version: 1.8.10
 signing_key: 
 specification_version: 3
 summary: A link parser for Ruby

data/lib/linkser/parser/html.rb

@@ -1,73 +0,0 @@
-require 'nokogiri'
-require 'open-uri'
-require 'net/http'
-require 'image_spec'
-
-module Linkser
-  module Parser
-    class HTML
-      def parse url, options={}
-        parsed_page = Hash.new
-
-        doc = Nokogiri::HTML(open(url))
-
-        doc.css('title').each do |title|
-          parsed_page.update({:title => title.text})
-        end
-
-        doc.css('meta').each do |meta|
-          if meta.get_attribute("name").eql? "description"
-            parsed_page.update({:description => meta.get_attribute("content")})
-          end
-        end
-
-        images = Array.new
-
-        doc.css('img').each do |img|
-          img_src = img.get_attribute("src")
-          img_src = get_complete_url img_src, url
-          img_uri = URI.parse(img_src)
-          img_ext = File.extname(img_uri.path)
-          img_name = File.basename(img_uri.path,img_ext)
-          if [".jpg", ".jpeg", ".png"].include? img_ext
-            begin
-              img_spec = ImageSpec.new(img_src)
-              w = img_spec.width.to_f
-              h = img_spec.height.to_f
-              if w > 199 or w > 199
-                if ((w > 0 and h > 0 and ((w / h) < 3) and ((w / h) > 0.2)) or (w > 0 and h == 0 and w < 700) or (w == 0 and h > 0 and h < 700)) and img_name.index("logo").nil?
-                  image = {:img => img_src, :width => w.to_i, :height => h.to_i}
-                images << image
-                end
-              end
-            rescue
-            end
-          end
-        end
-
-        if images!=[]
-          parsed_page.update({:images => images})
-        end
-
-        return parsed_page
-      end
-
-      private
-
-      def get_complete_url src, url
-        uri = URI.parse(url)
-        base_url = "http://" + uri.host + (uri.port!=80 ? ":" + uri.port.to_s : "")
-        relative_url = "http://" + uri.host + (uri.port!=80 ? ":" + uri.port.to_s : "") + uri.path
-        if src.index("http://")==0
-        src = src
-        #stays the same
-        elsif src.index("/")==0
-        src = base_url + src
-        else
-        src = relative_url + src
-        end
-      end
-    end
-  end
-end
-