elisehuard-media_feed 0.0.1

data/lib/media_feed.rb

@@ -0,0 +1,189 @@
+=begin rdoc
+The media_feed library parses a media rss stream and stores the information in its attributes.
+== Examples:
+Fetch a whole feed
+    media_feed = MediaFeed::Feed.new('http://www.nasa.gov/rss/NASAcast_vodcast.rss ')
+    media_feed.fetch
+Fetch starting at a certain date (date must be in RFC2822 format - use Time.rfc2822)
+    media_feed = MediaFeed::Feed.new('http://www.nasa.gov/rss/NASAcast_vodcast.rss ')
+    media_feed.fetch_since("Wed, 10 Sep 2008 10:29:18 +0200")
+Use the retrieved information
+    thumbnail = media_feed.thumbnail
+    media_feed.items.each do |item|
+      ...
+    end
+=end
+
+require 'rubygems'
+require 'open-uri'
+require 'libxml'
+
+module MediaFeed
+# Exception for an invalid URL format
+class InvalidUrl < Exception
+end
+
+# Exception for an invalid Item
+class InvalidMediaItem < Exception
+end
+
+# Exception for if feed url returns an empty result or a HTTP error
+class FeedNotFound < Exception
+end
+
+# Exception for invalid xml
+class InvalidXML < Exception
+end
+
+# Exception when receiving an invalid date (non-RFC2822) for input
+class InvalidDate < Exception
+end
+
+# Mediafeeds handles the retrieval of feed information.  The feed attributes are stored
+# in feed attributes, and the individual items are stored in an array of Item objects.
+class Feed
+  include(LibXML)
+  
+  attr_reader :title, :description, :thumbnail, :pubDate, :items
+
+  # Feed is initialized with a valid url for the feed itself.
+  def initialize(url)
+    raise ArgumentException if url.nil?
+    raise InvalidUrl if url !~ /(^(http|https):\/\/[a-z0-9]+([-.]{1}[a-z0-9]*)+. [a-z]{2,5}(([0-9]{1,5})?\/.*)?$)/ix
+    @url = url
+    @entries = nil
+  end
+  
+  # fetch the whole media feed
+  def fetch
+    doc = get_feed
+    return nil if doc.nil?
+    parse_feed(doc)
+  end
+  
+  # fetch all entries of the media feed since a certain date (RFC2822 format).
+  def fetch_since(last_date)
+    @last_date = parse_date(last_date)
+    fetch
+  end
+  
+  def to_s
+    "title #{@title}, description #{@description}" +
+    @items.inject('') do |str,item|
+      str + entry.to_s
+    end
+  end
+  
+private
+  
+  def get_feed
+    response = ''
+    open(@url){ |f|
+      # Save the response body
+      response = f.read
+    }    
+    raise FeedNotFound if response.nil? || response.empty?
+    response
+  rescue OpenURI::HTTPError
+    raise FeedNotFound
+  end
+  
+  # for some strange reason, the xml parsing by libxml stumbles over & symbols.  cute.
+  def preprocess(doc)
+    doc.gsub!(/&/,'&amp;')
+    doc
+  end
+  
+  def parse_feed(doc)
+    all_nodes = parse_document(preprocess(doc))
+    channel = all_nodes[0]
+    @title = content(channel,'title')
+    @description = content(channel,'description')
+    @thumbnail = content(channel,'image/url')
+    @pubDate = content(channel,'pubDate')
+    if @pubDate && @last_date
+      feed_pub_date = parse_date(@pubDate)
+      return if feed_pub_date == @last_date
+    end
+    items = channel.find('//item')
+
+    @items = items.inject([]) do |result,node|
+      item = parse_item(node)
+      if !item.valid? # invalid items shouldn't be added to feed but shouldn't stop the rest of the feed
+        puts "Faulty item in the feed - not loaded : "
+        puts item.to_s
+      else
+        result << item if item.is_after?(@last_date)
+      end
+      result
+    end
+  end
+
+  def parse_document(doc)
+    parser = LibXML::XML::Parser.new
+    parser.string = doc
+    xml = parser.parse
+    root = xml.root
+    root.find('*')
+  rescue
+    raise InvalidXML
+  end
+  
+  def parse_item(node)
+    item = Item.new
+    item.title = content(node,'title')
+    item.description = content(node,'description')
+    item.link = content(node,'link')
+    item.enclosure = url(node,'enclosure') || url(node,'media:content') || content(node,'guid')
+    date = content(node,'pubDate')
+    item.pubDate = parse_date(date) if date && !date.empty?
+    item.thumbnail = content(node,'image/url')
+    
+    # pubDate if not filled in by feed -> first (valid) one is probably most recent
+    @pubDate = item.pubDate.rfc2822 if @pubDate.nil? && item.pubDate
+    item
+  end
+
+  def content(parent_node,xpath)
+    parent_node.find(xpath)[0].content if parent_node.find(xpath) && parent_node.find(xpath)[0]
+  end
+  
+  def url(parent_node,xpath)
+    # ,['media:http://search.yahoo.com/mrss']
+    parent_node.find(xpath)[0]['url'] if parent_node.find(xpath) && parent_node.find(xpath)[0]
+  end
+  
+  # date string should be in rfc2822 format ex. Sat, 06 Sep 2008 11:59:15 +0000
+  def parse_date(date_string)
+    time = Time.rfc2822(date_string)
+  rescue => e
+    raise InvalidDate
+  end
+  
+end
+
+class Item
+  attr_accessor :title, :description, :link, :pubDate, :enclosure, :thumbnail
+  
+  def valid?
+    if title.nil? || link.nil? || enclosure.nil?
+      false
+    else
+      true
+    end
+  end
+  
+  def is_after?(last_date)
+    (last_date.nil? || pubDate.nil? || (pubDate > last_date))
+  end
+  
+  def to_s
+    "title #{title}\n" + 
+    "description #{description}\n" +
+    "link #{link}\n" +
+    "enclosure #{enclosure}\n" +
+    "pubDate #{pubDate}\n"
+  end
+  
+end
+end