plumnailer 0.0.4 → 0.1.0

data/VERSION

@@ -1 +1 @@
-0.0.4
+0.1.0

data/lib/plumnailer/caching_fetcher.rb

@@ -10,7 +10,7 @@ module Plumnailer
 
     def initialize(cache_dir)
       @cache_dir = cache_dir
-      FileUtils.mkdir_p(cache_dir)
+      FileUtils.mkdir_p cache_dir
     end
 
     # Fetch the contents of a url and cache result on filesystem.

data/lib/plumnailer/chooser.rb

@@ -6,7 +6,7 @@ module Plumnailer
     def initialize
       @fetcher = Plumnailer::Fetcher.new
       @doc_parser = Plumnailer::DocParser.new
-      @img_url_filters = [Plumnailer::ImgHostnameFilter.new]
+      @img_url_filters = [Plumnailer::ImgUrlFilter.new]
       @img_parser = Plumnailer::ImgParser.new(fetcher)
       @img_comparator = Plumnailer::ImgComparator
     end
@@ -28,7 +28,7 @@ module Plumnailer
         imgs.each do |img|
           # set source document on image so it can be used in comparator
           img.doc = doc
-          img.extend(@img_comparator)
+          img.extend @img_comparator
         end
         imgs.sort.first
       end

data/lib/plumnailer/doc.rb

@@ -5,6 +5,12 @@ module Plumnailer
   # Nokogiri::HTML:Document mixin.
   module Doc
 
+    # Get the href attribute of the base tag from the head of the document.
+    def doc_base_href
+      base = at('//head/base')
+      base['href']  if base
+    end
+
     # Return a list of the src attributes of all img tags.
     def img_srcs
       search('//img').map { |x| x['src'] }.compact
@@ -20,7 +26,12 @@ module Plumnailer
         rescue URI::InvalidURIError
           next
         end
-        result.push(u.is_a?(URI::HTTP) ? u : URI.join(base_url || source_url, i))
+
+        result <<  if u.is_a?(URI::HTTP)
+          u
+        else
+          URI.join(base_url || doc_base_href || source_url, i)
+        end
       end
 
       result

data/lib/plumnailer/img_parser.rb

@@ -14,7 +14,7 @@ module Plumnailer
 
     # Parse image data from one or more urls.
     def parse(img_urls)
-      if img_urls.respond_to? :inject
+      if img_urls.respond_to?(:inject)
         cache = {}
         img_urls.inject([]) do |memo,u|
           # nil values should be cached
@@ -29,7 +29,7 @@ module Plumnailer
     # additional fields.
     def parse_one(img_url)
       img_data = fetcher.fetch(img_url)
-      unless !img_data or img_data.empty?
+      unless not img_data or img_data.empty?
         img = Magick::ImageList.new.from_blob(img_data).extend(
           Plumnailer::WebImage)
         img.source_url = img_url

data/lib/plumnailer/img_url_filter.rb

@@ -0,0 +1,23 @@
+module Plumnailer
+
+  # Decide whether to process images based on their url.
+  class ImgUrlFilter
+
+    # Return true if this image url should not be considered.
+    def reject?(img_url)
+      ImgUrlPatterns.each do |re|
+        return true  if img_url and img_url.to_s[re]
+      end
+      false
+    end
+
+    ImgUrlPatterns = [
+      %r{^http://ad\.doubleclick\.net/},
+      %r{^http://b\.scorecardresearch\.com/},
+      %r{^http://pixel\.quantserve\.com/},
+      %r{^http://s7\.addthis\.com/},
+      ]
+
+  end
+
+end

data/lib/plumnailer.rb

@@ -3,8 +3,8 @@ require 'plumnailer/doc_parser'
 require 'plumnailer/doc'
 require 'plumnailer/fetcher'
 require 'plumnailer/img_comparator'
-require 'plumnailer/img_hostname_filter'
 require 'plumnailer/img_parser'
+require 'plumnailer/img_url_filter'
 require 'plumnailer/web_image'
 
 require 'plumnailer/caching_fetcher'

data/plumnailer.gemspec

@@ -1,15 +1,15 @@
 # Generated by jeweler
 # DO NOT EDIT THIS FILE DIRECTLY
-# Instead, edit Jeweler::Tasks in Rakefile, and run the gemspec command
+# Instead, edit Jeweler::Tasks in Rakefile, and run 'rake gemspec'
 # -*- encoding: utf-8 -*-
 
 Gem::Specification.new do |s|
   s.name = %q{plumnailer}
-  s.version = "0.0.4"
+  s.version = "0.1.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Matthew M. Boedicker"]
-  s.date = %q{2010-11-11}
+  s.date = %q{2010-11-21}
   s.description = %q{Choose the most representative image on an HTML page for use as a thumbnail}
   s.email = %q{matthewm@boedicker.org}
   s.extra_rdoc_files = [
@@ -17,24 +17,23 @@ Gem::Specification.new do |s|
   ]
   s.files = [
     "COPYING",
-     "README.textile",
-     "Rakefile",
-     "VERSION",
-     "lib/plumnailer.rb",
-     "lib/plumnailer/caching_fetcher.rb",
-     "lib/plumnailer/chooser.rb",
-     "lib/plumnailer/doc.rb",
-     "lib/plumnailer/doc_parser.rb",
-     "lib/plumnailer/fetcher.rb",
-     "lib/plumnailer/img_comparator.rb",
-     "lib/plumnailer/img_hostname_filter.rb",
-     "lib/plumnailer/img_parser.rb",
-     "lib/plumnailer/web_image.rb",
-     "plumnailer.gemspec",
-     "test.rb"
+    "README.textile",
+    "Rakefile",
+    "VERSION",
+    "lib/plumnailer.rb",
+    "lib/plumnailer/caching_fetcher.rb",
+    "lib/plumnailer/chooser.rb",
+    "lib/plumnailer/doc.rb",
+    "lib/plumnailer/doc_parser.rb",
+    "lib/plumnailer/fetcher.rb",
+    "lib/plumnailer/img_comparator.rb",
+    "lib/plumnailer/img_parser.rb",
+    "lib/plumnailer/img_url_filter.rb",
+    "lib/plumnailer/web_image.rb",
+    "plumnailer.gemspec",
+    "test.rb"
   ]
   s.homepage = %q{http://github.com/mmb/plumnailer}
-  s.rdoc_options = ["--charset=UTF-8"]
   s.require_paths = ["lib"]
   s.rubygems_version = %q{1.3.7}
   s.summary = %q{Choose the most representative image on an HTML page}

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: plumnailer
 version: !ruby/object:Gem::Version 
-  hash: 23
+  hash: 27
   prerelease: false
   segments: 
   - 0
+  - 1
   - 0
-  - 4
-  version: 0.0.4
+  version: 0.1.0
 platform: ruby
 authors: 
 - Matthew M. Boedicker
@@ -15,7 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-11-11 00:00:00 -05:00
+date: 2010-11-21 00:00:00 -05:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -71,8 +71,8 @@ files:
 - lib/plumnailer/doc_parser.rb
 - lib/plumnailer/fetcher.rb
 - lib/plumnailer/img_comparator.rb
-- lib/plumnailer/img_hostname_filter.rb
 - lib/plumnailer/img_parser.rb
+- lib/plumnailer/img_url_filter.rb
 - lib/plumnailer/web_image.rb
 - plumnailer.gemspec
 - test.rb
@@ -81,8 +81,8 @@ homepage: http://github.com/mmb/plumnailer
 licenses: []
 
 post_install_message: 
-rdoc_options: 
-- --charset=UTF-8
+rdoc_options: []
+
 require_paths: 
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement 

data/lib/plumnailer/img_hostname_filter.rb

@@ -1,22 +0,0 @@
-module Plumnailer
-
-  # Decide whether to process images based on their url hostname.
-  class ImgHostnameFilter
-
-    # Return true if this image url should not be considered.
-    def reject?(img_url)
-      HostnameRejectPatterns.each do |re|
-        return true if img_url.host and img_url.host[re]
-      end
-      false
-    end
-
-    HostnameRejectPatterns = [
-      %r{^ad\.doubleclick\.net$},
-      %r{^b\.scorecardresearch\.com$},
-      %r{^pixel\.quantserve\.com$},
-      ]
-
-  end
-
-end