goatless 0.1.0

data/bin/goatless

@@ -0,0 +1,28 @@
+#!/usr/bin/env ruby
+
+require 'optparse'
+require 'goatless'
+
+options = {}
+optparse = OptionParser.new do |opts|
+  opts.banner = "Usage: goatless.rb -u username -p password -f path"
+  opts.on("-u", "--username [username]", "LJ Username") do |v|
+    options[:username] = v
+  end
+  opts.on("-p", "--password [password]", "LJ Password") do |v|
+    options[:password] = v
+  end
+  opts.on("-f", "--file [filepath]", "File to dump RSS feed") do |v|
+    options[:filename] = v
+  end
+end
+optparse.parse!
+
+
+if [options[:username], options[:password], options[:filename]].include? nil
+  puts optparse
+  exit(-1)
+end
+
+goatless = Goatless::Goatless.new(options[:username], options[:password])
+File.new(options[:filename], "w+").write(goatless.rss)

data/lib/goatless/friend_fetcher.rb

@@ -0,0 +1,76 @@
+module Goatless
+  class FriendFetcher
+    attr_accessor :friend_url, :username, :password, :items
+
+    def initialize(friend_url, username, password)
+      puts friend_url
+      @friend_url = friend_url
+      @username = username
+      @password = password
+    end
+
+    def fetch
+      @patron_session = Patron::Session.new
+      @patron_session.username = @username
+      @patron_session.password = @password
+      @patron_session.timeout  = 10
+      @patron_session.connect_timeout = 30000
+      @patron_session.auth_type = :digest
+      @items = []
+      xml = @patron_session.get("#{@friend_url}?auth=digest").body
+      parse(xml)
+    end
+
+    def parse(xml)
+      xml_doc = Nokogiri::XML::Reader(xml)
+      @storage = PStore.new('goatless.pstore') # Hope this can decrease memory usage
+
+      max_items = 50 # Move this to the CLI flag or config
+      current_item = nil
+      current_author = nil
+
+      xml_doc.each do |node|
+        next if max_items == 0
+        case node.name
+        when 'lj:journal'
+          current_author = node.inner_xml if current_author.nil?
+        when 'item'
+          if current_item.nil?
+            current_item = RssItem.new
+          else
+            current_item.author = current_author
+            if current_item.pub_date > (DateTime.now - 7) # do not add items older then week
+              @storage.transaction do
+                @storage[:items] ||= Array.new
+                unless @storage[:items].include? current_item
+                  @storage[:items] << current_item
+                  @storage.commit
+                end
+              end
+            end
+            current_item = nil
+            max_items -= 1
+          end
+        when 'link'
+          unless current_item.nil?
+            current_item.permalink = node.inner_xml if current_item.permalink.nil?
+          end
+        when 'pubDate'
+          current_item.pub_date = DateTime.parse(node.inner_xml) if current_item.pub_date.nil?
+        when 'description'
+          unless current_item.nil?
+            if current_item.body.nil?
+              current_item.body = node.inner_xml
+            end
+          end
+        when 'lj:security'
+          current_item.security_type = node.inner_xml if current_item.security_type.nil?
+        when 'title'
+          unless current_item.nil?
+            current_item.title = node.inner_xml if current_item.title.nil?
+          end
+        end
+      end
+    end
+  end
+end

data/lib/goatless/goatless.rb

@@ -0,0 +1,30 @@
+module Goatless
+  class Goatless
+    def initialize(username, password)
+      @username, @password = username, password
+      @threads = []
+      friends_urls = OpmlFetcher.new(@username).fetch
+      friends_urls.each_slice(3) do |rss_urls|
+        @threads << Thread.new {
+          rss_urls.each do |rss_url|
+            friend_fetcher = FriendFetcher.new(rss_url, @username, @password)
+            begin
+              friend_fetcher.fetch
+            rescue Exception => e
+              # Still have a lot's of "connection timed out" here
+              puts e
+            end
+          end
+        }
+      end
+      @threads.each { |t| t.join }
+    end
+
+    def rss
+      @storage = PStore.new('goatless.pstore')
+      @storage.transaction(true) do
+        RssBuilder.new(@storage[:items]).document
+      end
+    end
+  end
+end

data/lib/goatless/opml_fetcher.rb

@@ -0,0 +1,24 @@
+module Goatless
+  class OpmlFetcher
+    attr_accessor :rss_urls
+
+    def initialize(username)
+      @url = "http://www.livejournal.com/tools/opml.bml?user=#{username}"
+      # @url = 'http://sdfgh153.ru/sample.opml'
+    end
+
+    def fetch
+      patron_session = Patron::Session.new
+      patron_session.timeout = 30
+      patron_session.connect_timeout = 30000
+      @rss_urls = []
+
+      opml_doc = Nokogiri::XML(patron_session.get(@url).body)
+      opml_doc.elements.xpath("//outline").each do |outline|
+        @rss_urls << outline['xmlUrl']
+      end
+      puts "You have #{@rss_urls.size} friends!"
+      @rss_urls
+    end
+  end
+end

data/lib/goatless/rss_builder.rb

@@ -0,0 +1,40 @@
+module Goatless
+  class RssBuilder
+
+    def initialize(items)
+      @items = items
+    end
+
+    def document
+      #TODO: Move number of items in feed into settings
+      items = @items.sort { |x, y| x.pub_date <=> y.pub_date }.last(100).reverse
+      
+      builder = Nokogiri::XML::Builder.new(:encoding => 'UTF-8') do |rss|
+        rss.rss(:version => '2.0') {
+          rss.channel {
+            rss.title "Friends"
+            rss.generator "Goatless"
+            rss.description "Goatless friends"
+            rss.link "https://github.com/semka/goatless"
+
+            items.each do |i|
+              rss.item {
+                rss.pubDate i.pub_date.strftime("%a, %d %b %Y %H:%M:%S %z") # to RFC 822
+                rss.author  i.author
+                rss.link    i.permalink
+                if i.public?
+                  rss.title   i.title
+                  rss.description do |d| d << i.body end
+                else
+                  rss.title  "Private post"
+                  rss.description "Private post"
+                end
+              }
+            end
+          }
+        }
+      end
+      builder.to_xml
+    end
+  end 
+end

data/lib/goatless/rss_item.rb

@@ -0,0 +1,14 @@
+module Goatless
+
+  class RssItem
+    attr_accessor :pub_date, :xml_element, :security_type, :pub_date, :title, :body, :permalink, :author
+
+    def public?
+      @security_type == 'public'
+    end
+
+    def ==(an_rss_item)
+      self.permalink == an_rss_item.permalink && !self.permalink.nil?
+    end
+  end 
+end

data/lib/goatless.rb

@@ -0,0 +1,7 @@
+require 'rubygems'
+require 'date'
+require 'patron'
+require 'nokogiri'
+require 'pstore'
+
+Dir.glob(File.join(File.dirname(__FILE__), 'goatless/**.rb')).each { |f| require f }

metadata

@@ -0,0 +1,71 @@
+--- !ruby/object:Gem::Specification 
+name: goatless
+version: !ruby/object:Gem::Version 
+  hash: 27
+  prerelease: 
+  segments: 
+  - 0
+  - 1
+  - 0
+  version: 0.1.0
+platform: ruby
+authors: 
+- Semka Novikov
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2011-08-19 00:00:00 Z
+dependencies: []
+
+description: 
+email: semka.novikov@gmail.com
+executables: 
+- goatless
+extensions: []
+
+extra_rdoc_files: []
+
+files: 
+- bin/goatless
+- lib/goatless/friend_fetcher.rb
+- lib/goatless/goatless.rb
+- lib/goatless/opml_fetcher.rb
+- lib/goatless/rss_builder.rb
+- lib/goatless/rss_item.rb
+- lib/goatless.rb
+homepage: http://yoursite.example.com
+licenses: []
+
+post_install_message: 
+rdoc_options: []
+
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  none: false
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  none: false
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.8.5
+signing_key: 
+specification_version: 3
+summary: What this thing does
+test_files: []
+